Title:
Dynamic genomic deletion expansion and formulation of molecular marker panels for integrated molecular pathology diagnosis and characterization of tissue, cellular fluid, and pure fluid specimens
Kind Code:
A1


Abstract:
Provided are materials, methods, platforms, and kits for diagnosing, prognosing, and/or determining the biological aggressiveness of a tumor (malignant and non-malignant) based in part on the temporal profile of genomic deletion expansion of the tumor. Also provided are methods of determining marker panels for different diseases and/or tissues and markers identified by these methods.



Inventors:
Finkelstein, Sydney David (Pittsburgh, PA, US)
Swalsky, Patricia (Pittsburgh, PA, US)
Application Number:
11/256152
Publication Date:
04/27/2006
Filing Date:
10/24/2005
Primary Class:
Other Classes:
435/91.2
International Classes:
C12Q1/68; C12P19/34
View Patent Images:



Primary Examiner:
WOOLWINE, SAMUEL C
Attorney, Agent or Firm:
FAEGRE DRINKER BIDDLE & REATH LLP (DC) (WASHINGTON, DC, US)
Claims:
I claim:

1. A method of determining tumor aggressiveness in a patient comprising the steps of: (a) amplifying DNA from a microdissection section of a biological sample from the patient; (b) analyzing two or more genes for the presence of a nucleotide deletion wherein a deletion is the acquisition of genetic mutation; (c) analyzing each gene with an array of markers to determine the extent of nucleotide deletion; (d) determining the order of acquisition of nucleotide deletions in the patient; and (e) collating the data from steps (a) to (d) to determine tumor aggressiveness.

2. The method of claim 1, wherein the biological sample is pretreated with a proteinase in lysis buffer, which contains a nonionic detergent.

3. The method of claim 2, wherein the proteinase is selected from the group consisting of proteinase K, pronase, subtilisin, thermolysin, papain, or a combination of proteinases.

4. The method of claim 3, wherein the proteinase or combination of proteinases is present in the amount of about 0.5% to about 2.0% final volume.

5. The method of claim 4, wherein the proteinase is present in the amount of about 1.0% and is proteinase K.

6. The method of claim 2, wherein the nonionic detergent is Nonidet P40, Tween, Triton X, or Nikkol.

7. The method of claim 6, wherein the nonionic detergent is present in the amount of about 0.5% to about 2.0%.

8. The method of claim 1, wherein the amplifying step is performed in the presence of about 5 to about 10 mM magnesium chloride, and about 5 to about 20 g/100 mL sucrose.

9. The method of claim 8, wherein magnesium chloride is present in the amount of about 8 mM, and sucrose is present in the amount of about 12 g/100 mL.

10. The method of claim 1, wherein the biological sample is from a tissue sample from two different organ sites in said patient, wherein a first organ site is the first diagnosed cancer and a second organ site is a putative metastatic lesion or a second primary tumor.

11. The method of claim 1, wherein the biological sample is a cell-free fluid sample, a blood sample, a cytology sample, resected tissue, or a combination thereof.

12. The method of claim 11, wherein the cell-free fluid sample contains non-nuclear DNA.

13. The method of claim 11, wherein the resected tissue is frozen, fresh, stained, fixative-treated, or stained and fixative-treated.

14. The method of claim 1, wherein step (c) is repeated to obtain replicate data.

15. The method of claim 1, further comprising the step of identifying a treatment plan to treat the tumor of said patient which best treats the tumor based on the determination of tumor aggression.

16. The method of claim 1, wherein the biological sample is microdissected tissue, and the steps of (b) and/or (c) are performed on DNA obtained from two or more microdissected sections of the tissue sample.

17. The method of claim 16, wherein step (c) is repeated to obtain replicate data.

18. The method of claim 1, further comprising determining whether genetic mutation acquisition is due to familial inheritance, an environmental factor, or a spontaneous mutation based on the order of the mutation acquisition.

19. The method of claim 16, wherein a patient fluid sample undergoes analysis comprising: (a) amplifying DNA from the fluid sample of the patient; (b) analyzing two or more genes for the presence of nucleotide deletion from the amplified DNA, wherein a deletion is the acquisition of genetic mutation; (c) analyzing each nucleotide deletion with an array of markers to determine the extent of nucleotide deletion; (d) determining the order of acquisition of nucleotide deletion in the patient; and (e) validating the collated data of claim 16 with the data obtained from the fluid sample.

20. A kit for determining tumor aggressiveness comprising: (i) a device for amplifying DNA and analyzing the presence of a nucleotide deletion and order of acquisition of said nucleotide deletion; and (ii) sets of cancer specific markers for assessing nucleotide deletion and extent of nucleotide deletion.

21. The kit of claim 20, wherein the device also has a data storage component for storing patient information regarding sex, age, weight, medical history, family medical history, prior cancer history and genetic analysis on a prior cancer, and genomic deletion acquisition data in a relational database.

22. The kit of claim 20, wherein the markers are markers for environmentally-induced mutations, germ line mutations, and spontaneous mutations.

23. The kit of claim 22, wherein the markers for environmentally-induced mutations are trichloroethylene markers.

24. The kit of claim 20, wherein the kit further comprises reagents for amplifying DNA, wherein the reagents when admixed contain about 5 to about 10 mM magnesium chloride; about 5 to about 20 g/100 mL sucrose; about 0.5% to about 2.0% nonionic detergent; and about 0.5% to about 2.0% proteinase or a combination of proteinases.

25. The kit of claim 24, wherein the magnesium chloride when admixed is present in the amount of about 8 mM; the sucrose is present in the amount of 12 g/100 mL; the nonionic detergent is nonidet P-40 and is present in the amount of about 1.0%; and the proteinase is proteinase K and is present in the amount of 2 mg/mL.

26. A method of creating a panel of molecular markers for detecting a condition in a patient comprising the steps of: (a) determining gene targets for detection of a mutation to include in a molecular marker analysis for a marker panel; (b) delineating genomic regions for each gene target; (c) identifying a 4 to 1500 nucleotide repeat microsatellite and/or a minisatellite in the genomic region that will constitute the marker panel; (d) identifying at least two 4 to 1500 nucleotide microsatellites and/or minisatellites positioned a certain distance from each other and performing genomic deletional expansion; (e) determining the amount of DNA in the biological sample; (f) determining the quality of DNA in the biological sample by quantitative PCR; (g) determining the quality of DNA in the biological sample by means on competitive template PCR (CT-PCR); (h) determining the amplifiability of DNA for each 4 to 1500 repeat microsatellite and/or minisatellite; (i) defining a normal range of allele variation thereby defining allelic imbalance comprising: (i) defining different normal ranges for each allele for two or more quantities of DNA; and (ii) defining different normal ranges for each allele with two or more qualities of DNA; and (j) defining minimum thresholds for significant allelic imbalance thereby obtaining an indication of mutation change in a significant percentage of evaluated cells comprising: (i) defining minimum thresholds for significant allelic imbalance for different amounts of DNA; and (ii) defining minimum thresholds for significant allelic imbalance for different qualities of DNA; and (k) calculating a percentage of mutated cells based upon ratios of a tested sample using a calculated normal for each 4 to 1500 microsatellite and/or minisatellite, thereby creating a panel of molecular markers for detecting a condition in a patient.

27. The method of claim 26, wherein the microsatellite is 4 to 20 nucleotides.

28. The method of claim 27, wherein the microsatellite is 4-nucleotide microsatellite.

29. The method of claim 26, wherein the condition is a neoplasia, a hyperplasia, or a benign growth.

30. The method of claim 26, wherein the condition arises from a germ line mutation, an environmental factor, or a spontaneous mutation, or a combination thereof.

31. A marker or panel of markers identified by the method of claim 26.

32. A method of determining tumor aggressiveness in a patient comprising the steps of: (a) analyzing two or more genes from DNA amplified from a microdissection section of a patient biological sample for the presence of a nucleotide deletion wherein a deletion is the acquisition of genetic mutation; (b) analyzing each gene with an array of markers to determine the extent of nucleotide deletion; (c) determining the order of acquisition of nucleotide deletion in the patient; and (d) collating the data from steps (a) to (d) to determine tumor aggressiveness.

Description:

REFERENCE TO RELATED APPLICATIONS

This application claims benefit of U.S. Provisional Application Nos. 60/620,926 filed Oct. 29, 2004; 60/631,240 filed Nov. 29, 2004; 60/644,568 filed Jan. 19, 2005; 60/679,969 filed May 12, 2005; and 60/679,968 filed May 12, 2005, all of which are herein incorporated by reference in their entirety for all purposes.

FIELD OF THE INVENTION

The application relates to methods, materials, kits and devices for achieving a diagnosis, prognosis, and/or determination of the biological aggressiveness of a tumor based on the temporal profile (time course) of genomic deletion expansion. The materials, methods, kits and devices disclosed herein can also be used for pre-cancerous growths and diagnosis and prognosis of non-malignant tissues and cells.

BACKGROUND

Cancer is characterized by the stepwise accumulation of genetic damage. Such genetic damage includes mutations and aberrant expression of oncogenes and tumor suppressor genes that are causally related to phenotypic expression of malignant characteristics such as rapid uncontrolled growth, local invasion, and metastasis (Weir et al., 2004 Cancer Cell. 6(5): 433-8; and Osborne et al., 2004 Oncologist 9(4): 361-77). As individual mutations are acquired by affected cells, clonal expansion of biologically more aggressive tumor cells ensues (Braakhuis et al., 2005 Semin. Cancer Biol. 15(2): 113-20; Breivik et al., 2005 Semin Cancer Biol. 15(1): 51-60). Thus, cancer cells with increasing mutational change come to replace precursor cells with fewer mutations, which is the driving force for clonal expansion. The process is dynamic one occurring across the entire tumor, with the creation of distinct tumor cell clones in topographic regions that undergo neoplastic progression at different rates. To a large extent, the process is irreversible, with the most biologically aggressive cellular elements of a particular cancer overgrowing the relatively less aggressive cellular components. The majority of human cancers are eventually dominated by clonal expanding malignant cells possessing specific, detectable mutational alterations directly responsible for tumor growth and biological behavior.

One of the most common cancer related genetic alterations is genomic deletion (Bishop A J, et al., 2003 Exp. Mol. Pathol. 74(2): 94-105; Popescu, 2003 Cancer Lett. 192(1): 1-17). This form of damage is commonly associated with tumor suppressor gene inactivation which follows a two-step process (Baker et al., 2003 Genes Chromosomes Cancer. 38(4): 329 and Knudson, 2001 Lancet Oncol. 2(10): 642-5). First, one normal copy of the tumor suppressor is inactivated, typically by means of point mutational damage. The second step is genomic deletion of the remaining normal copy, often with adjacent non-coding DNA being included in the deletional region (Baker et al., 2003 and Knudson, 2001).

A primary objective in the management of cancer is predicting the intrinsic biological aggressiveness of an individual neoplasm, so that it may be treated in an appropriate manner. Cancers that are intrinsically indolent should be treated in a conservative manner to avoid the unnecessary morbidity associated with the side effects of therapy. In contrast, cancers that can be predicted to behave in a more aggressive fashion justifiably demand a more aggressive therapy to avoid under the under-treatment of the patient. Thus, it is essential to effectively prognosticate the biological aggressiveness of each cancer on an individual patient basis.

The prediction of cancer aggressiveness, according to conventional teaching, involves searching for specific microscopic features such as high mitotic rate, cellular anaplasia, nuclear pleomorphism, etc. Conventional teaching does not require that the prediction of cancer aggressive require criteria other than those encompassed by microscopic observation. At the extremes (i.e. near normal cellular appearance with no evidence of high mitotic rate, cellular anaplasia, nuclear pleomorphism versus extensive high mitotic rate, cellular anaplasia, nuclear pleomorphism), these criteria are generally reliable. However, the majority of clinical cases fall in between the polar ends of this spectrum with some but not all microscopic features of biological aggressiveness present. Under these circumstances, the prediction of tumor biological aggressiveness cannot be carried out microscopically with certainty. The use of additional technique to derive correlative mutational information can serve as a powerful means to add discriminating information leading to greater certainty.

Unfortunately, in the majority of cases, such clear-cut discrimination is not possible, and there is significant error upon the part of the diagnosing physician in predicting cancer aggressiveness based on microscopic features alone. In fact, patients with the identical histologic type of cancer, including cellular microscopic features, have proven over time not to behave in an equivalent manner. One patient may die after a short interval from rapid tumor progression, while another patient shows a long disease-free survival. The mortality for equivalent stages of human cancer is ex pressed as a probability, although it is recognized that there will be variation across the spectrum of patient outcomes. This renders individualized cancer patient care less effective.

Progress over the last decade has provided abundant information regarding the molecular changes in cancer (Ross et al., 2002 Hematology 7(4): 239-52 and Tomlinson et al., 2002 Genes Chromosomes Cancer 34(4): 349-53). Cancer aggressiveness can be more effectively but not completely predicted by parameters such as the cumulative amount of mutational damage (Ross et al., 2002 and Tomlinson et al., 2002) or the level of altered gene expression (Hughes, 2004 J. Surg. Oncol. 85(1): 28-35; Risques et al., 2003 Cancer Res. 63(21): 7206-14; and Kwon et al., 2004 Dis. Colon Rectum. 47(2): 141-52). In some cases, specific gene involvement has proven capable of characterizing differential cancer biological aggressiveness (Risques et al., 2003 Histol. Histopathol. 18(4): 1289-99). While great strides have been taken, there is much to be learned concerning the control mechanisms for tumor behavior.

The realities of clinical practice demand that cancer diagnosis be accomplished on minute samples of tumor or fluid aspirate. Moreover, all cell or tissue specimens removed from a patient's body must be thoroughly evaluated by microscopic pathology evaluation to ensure that adequate amounts of lesion material have been obtained, and in order to derive conventional histopathologic diagnosis upon which to base a standard diagnosis. Measures that interfere with this basic approach and fail to complement routine pathology practice run the risk of significant error, with serious medical and legal consequences. Rather, molecular analysis must proceed in series after conventional microscopic examination on the very same samples in order to effectively integrate mutational analysis into pathology practice. As a result, the molecular analysis must also be effective on small-sized, fixative-treated and/or stained cells as well as on cells or DNA material containing in biological fluid samples. The most valuable diagnostic methods are those that meet these operational criteria.

Advances such as the sequencing of the human genome have made it possible to design molecular testing strategies to assist in the detection of cancer, precancer and related conditions (Kreiner et al., 2005 Am. J. Health Syst. Pharm. 62(3): 296-305). Among the information generated by the Human Genome Project is the position of polymorphic markers in the form of minisatellites, microsatellites and single nucleotide polymorphisms (SNPs) that can be used to track the presence of both copies of markers and genes that compose the human genome (Beroud et al., 2005 Hum. Mutat. 26(3): 184-91). While the existence of these markers are known, their value and use in the context of searching for and characterizing human mutation is still unknown. It is the desire of many to use these markers to evaluate pathologic specimens. However, their utilization remains unclear.

When endeavoring to use markers to explore cancer formation and progression, the issue arises as to how to formulate a panel of markers that will serve this function best and provide the most information (Berger et al., 2003 Diagn. Mol. Pathol. 12(4):187-92). Conventional teaching on the subject is minimal and repeats the basic tenet that one should choose a panel of markers that are as close as possible to those genes, which research has already shown to be involved in the molecular pathogenesis of a particular form of cancer (e.g., Yoshino et al., 2005 Respir. Med. 99(3): 308-12). Current convention instructs that the best markers are dinucleotide microsatellites and SNPs, because there are many more of these markers scattered across the human genome and thus one can get closer or even within specific genes of interest. Beyond this recommendation to favor the use of dinucleotide microsatellites and SNPs, there are no further recommendations and one is left to determine how best to then utilize these markers and perform and broad marker panel analysis

SUMMARY

Herein are provided methods, compositions, kits, and devices for assessing dynamic genomic deletional expansion that can be determined using small amounts of tissue, fixative-treated tissue, cellular specimens, and/or fluid specimens that directly and effectively prognosticate cancer biological aggressiveness and combinations thereof. This observation has heretofore not been described, yet its application in everyday clinical practice can be highly effective and easy to perform. The result is a new modality with which to characterize a tumor in a patient on an individual basis. Although the patient is preferably human, such diagnosis can also be performed on other animals for veterinary diagnosis. It may be expected to have profound consequences for future diagnosis of cancer and pre-cancer states and, by consequence, therapeutic management.

The materials, methods, and kits provided herein can be utilized to diagnose and characterize the biological aggressiveness of a tissue or cells, such as a neoplastic tissue, thereby improving patient diagnosis and treatment of the patient with the tissue abnormality.

Provided herein is a method of determining tumor aggressiveness in a patient comprising the steps of: (a) amplifying DNA from a microdissection section of the biological sample; (b) analyzing two or more genes for the presence of a nucleotide deletion wherein a deletion is the acquisition of genetic mutation; (c) analyzing each gene with an array of markers to determine the extent of nucleotide deletion; (d) determining the order of acquisition of each nucleotide deletion in the patient; and (e) collating the data from steps (a) to (d) to determine tumor aggressiveness of the patient. In one aspect, the reagents used to amplify DNA from the biological sample of the patient, when admixed for amplification, comprises about 1 to about 15 mM magnesium chloride (other valency ions such as manganese can be substituted in the form of MnCl2), more preferably from about 6.0 to about 10.0 mM (and every 0.1 mM value in between), and most preferably about 8.0 mM magnesium chloride. The solution preferably also comprises when admixed about 5 to about 20 g/100 mL sucrose (and any 0.1 g/100 mL value there between), more preferably the solution contains about 12.0 g/100 mL sucrose.

The biological sample can be pretreated with a proteinase from about 2 hours to overnight. Preferably, the proteinase treatment occurs at about 37° C. and halted by boiling the sample for about 5 minutes. Suitable proteinases include but are not limited to proteinase K, pronase, subtilisin, thermolysin, papain, or a combination of proteinases. Proteinases are present in an amount of about 0.5% to about 2.0% final volume of lysis buffer (and every 0.1% value there between), and preferably about 1.0%. The lysis buffer further includes a nonionic detergent. Nonionic detergents can include but are not limited to nonidet P40, Tween (e.g., Tween 20), Triton X, or Nikkol. The nonionic detergents are present in the amount of about 0.5% to about 2.0% (and any 0.1% value in between).

The biological sample may be a cell-free fluid sample, a blood sample, a cytology sample, a urine sample, a tissue swab, or resected tissue. The cell-free fluid sample contains non-nuclear DNA. The resected tissue may be frozen, fresh, stained, or fixative-treated, or a combination thereof. More than one type of biological sample can be utilized for each patient. For example, a fluid sample and a resected tissue sample can be analyzed and the data from each combined. Additionally, biological samples can be derived from more than one point of origin in a patient. For example, a biological sample can be obtained from a putative primary tumor as well as a biological sample from a putative second unrelated tumor or from a putative metastatic lesion, or combination thereof. Thus, for example, in Wilm's tumor, wherein nodules may each be a primary, samples of each nodule may be analyzed alone or the data combined.

Step (c) of the method may be repeated to obtain replicate data. The method may further comprise the step of identifying a treatment plan to treat the tumor of said patient which best treats the tumor based on the determination of tumor aggression. The biological sample may be microdissected tissue, and the steps of (b) and/or (c) may be performed on DNA obtained from two or more microdissected sections of the tissue sample. Step (c) may be repeated to obtain replicate data.

The patient may also have a fluid sample, and the fluid sample may undergo analysis comprising: (a) amplifying DNA from the fluid sample of the patient; (b) analyzing two or more genes for the presence of nucleotide deletion from the amplified DNA; (c) analyzing each nucleotide deletion with an array of markers to determine the extent of nucleotide deletion; (d) determining the order of acquisition of each nucleotide deletion in the patient; and (e) validating the collated data with the data from the fluid sample. Markers can be used to determine whether the mutation was environmentally induced, such as by exposure to trichloroethylene or other chemical, a germ line mutation, or a spontaneous mutation.

Also provided is a kit for determining tumor aggressiveness comprising (i) a device for amplifying DNA; and (ii) sets of cancer specific markers for assessing nucleotide deletion and extent of nucleotide deletion. The markers can include markers for environmentally-induced mutations, spontaneous mutations, and/or germ line mutations. The kit may also provide for reagents for amplifying DNA as discussed above. The device for amplifying and determining genomic deletion may also comprise a data storage component for storing patient information regarding sex, age, weight, medical history, family medical history, prior cancer history and genetic analysis of a prior cancer, and genomic deletion acquisition data. This information may further be stored in a relational database, wherein factors are weighted in order to best determine treatment strategy.

Another aspect provided is a method of creating a panel of molecular markers for detecting a condition in a patient comprising the steps of: (a) determining gene targets for detection of a mutation to include in a molecular marker analysis for a marker panel; (b) delineating genomic regions for each gene target; (c) identifying a 4 to 1500 nucleotide repeat microsatellite and/or a minisatellite in the genomic region that will constitute the marker panel; (d) identifying at least two 4 to 1500 nucleotide microsatellites and/or minisatellites positioned a certain distance from each other and performing genomic deletional expansion; (e) determining the amount of DNA in the biological sample; (f) determining the quality of DNA in the biological sample by quantitative PCR; (g) determining the quality of DNA in the biological sample by means on competitive template PCR (CT-PCR); (h) determining the amplifiability of DNA for each 4 to 1500 repeat microsatellite and/or minisatellite; (i) defining a normal range of allele variation thereby defining allelic imbalance comprising:

    • (i) defining different normal ranges for each allele for two or more quantities of DNA; and
    • (ii) defining different normal ranges for each allele with two or more qualities of DNA; and
      (j) defining minimum thresholds for significant allelic imbalance thereby obtaining an indication of mutation change in a significant percentage of evaluated cells comprising:
    • (i) defining minimum thresholds for significant allelic imbalance for different amounts of DNA; and
    • (ii) defining minimum thresholds for significant allelic imbalance for different qualities of DNA; and
      (k) calculating a percentage of mutated cells based upon ratios of a tested sample using a calculated normal for each 4 to 1500 microsatellite and/or minisatellite, thereby creating a panel of molecular markers for detecting a condition in a patient. Preferably, the microsatellite is 4 nucleotides to 1500 nucleotides, more preferably 4 to 200, and more preferably 4 to 20 nucleotides (any integer value between 4 and 1500). The markers can be used to preferably diagnose a neoplasia, a hyperplasia, or a benign growth as well as determine if the mutation is germ line (inherited), spontaneous, or due to an environmental factor.

Also contemplated are markers or marker panels identified by the above method.

BRIEF DESCRIPTION OF THE DRAWINGS

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only, and are not restrictive of the material methods, devices, and kits. The accompanying drawings, which are incorporated herein by reference, and which constitute a part of this specification, illustrate certain embodiments together with the detailed description, serve to explain the principles of the materials, methods, devices, and kits.

The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee. The drawings are exemplary only, and should not be construed as limiting the materials, methods, devices and kits described herein.

FIG. 1. Phenomena of Genomic Deletion Expansion. FIG. 1 depicts a representative ideogram of one of two copies of a chromosome (chromosome 1). The arrow demarcates the position of a tumor suppressor gene on the chromosome. In the two-step process of tumor suppressor gene inactivation, the second normal copy of the gene is deleted leading to clonal expansion of phenotypically more aggressive cancer cells. Over time and through successive replications of the tumor cells, the deleted region undergoes progressive expansion of the deleted region (left to right). The rate of this expansion is proportional to the tumor's biological aggressiveness. In vivo, the two ends of the deletion regions are attached resulting in an intact chromosome that lacks an expanding segment of its genome.

FIG. 2. Determination of the Timeline of Mutation Acquisition. Timeline determination of the temporal sequence of acquired mutations occurs in two independent ways both reliant upon the fundamental concept of clonal expansion. The first is to compare the presence or absence of mutations at topographically separate locations within a tumor. Mutations that are present over a wider topographic distribution were acquired earlier in time those mutations that are focally distributed occurred later in time. The second is to measure the proportion of tumor cells bearing specific mutations. Mutations present in a high proportion of tumor cells occurred earlier in temporal sequence that other mutations affecting a smaller percentage of tumor cells at a specific location.

FIG. 3. Quality Control Analytic Validation. Quality control measures to monitor analytic precision are presented in this schematic. Analytic validation operates at two levels. The first is with respect to the sample which is replicated in a fashion leading that allows a determination of variation between separate samples. The second operates at the individual marker level with replication allowing a determination of the variation for specific markers within individual samples. The variability between individual results that are replicated can be used as a means to understand the significance of results.

FIG. 4. Electropherograms depicting an example of marker replication. In this replicate analysis for two different markers, the ratio of the allele peak heights is very similar providing validation of the analytical precision of the technique. The x-axis reflects the base pair length of the amplicon, and the y-axis is the relative fluorescence units detected.

FIG. 5. Allelic imbalance due to inadequate amounts of amplifiable DNA producing false-positive imbalance is depicted. Allelic dropout—the ratio of the allele peak heights is highly variable, yet these are replicate analysis of the identical specific marker from the identical specific microdissected specimen or source of DNA. Quality control measures are essential to guard against the occurrence of allelic dropout that will result in false positive results. The x-axis reflects the base pair length of the amplicon, and the y-axis is the relative fluorescence units detected.

FIG. 6. Procedure for comparative analysis of tumor-specific mutational profiles, as performed on tissue samples from esophagus and lung.

FIG. 7. Tumor-specific mutational profiling, as performed on tissue, from right upper lobe and right lower lobe of lung from the same patient.

FIG. 8. Tumor-specific mutational profiling, as performed on tissue from patient with squamous cell carcinoma of the tongue and gingival cancer.

FIG. 9. Tumor-specific mutational profiling, as performed on liver tissue from two different patients with cirrhosis.

FIG. 10. Tumor-specific mutational profiling, as performed on renal tissue and lip tissue from the same patient.

FIG. 11. Summary of mutational profiling for discrimination of cancer recurrence versus new primary cancer formation, as performed on samples from 278 patients.

FIG. 12. Bar graphs illustrating discordance, showing tumor recurrent and concordance, showing new primary cancers.

FIG. 13. Summary of results, showing 100% definitive diagnosis using mutational marker panels.

FIG. 14. Stutter band analysis using tetranucleotide or greater microsatellites as polymorphic markers. Use of a dinucleotide renders a series of shorter stutter band peaks, which reduces the exactness of quantitative determination of the amount of fluorescence for the true PCR product.

FIG. 15. Temporal sequence analysis of acquired mutational damage as used to create a molecular classification of gliomas.

FIG. 16. Endoscopic ultrasound images of a pancreatic cancer (left panels) and molecular analysis of the biological sample obtained by fine need aspiration.

FIG. 17. Microdissection of the biopsy fragments of duodenal mucosa showing adenomatous change (upper left panels). The biopsied fragments were paraffin-embedded, fixative treated biopsy samples, which were analyzed for a molecular analysis (lower left panel). In the bottom left panel/table, the markers constituting the panel used to search for evidence for or against the APC gene germ line mutation are shown. The marker panel consists of three intragenic single nucleotide polymorphisms (SNPs) in exons 10, 15, and the 3′ non-translated region of the APC gene. The marker panel also includes three tetranucleotide or large microsatellites situated in proximity, but outside the APC gene. The top right panel shows a representative example of one of the tetranucleotide microsatellites (i.e., D5S592), which shows a concordant pattern of allelic loss in two separate sites of neoplasia formation in this patient. In fact, all the tumor deposits showed the same concordant microsatellite band loss, which supports a germ line mutation in the APC gene. In the top right panel, the x-axis reflects the base pair length of the amplicon, and the y-axis is the relative fluorescence units detected. The bottom right pantel shows the same concordant pattern of SNP imbalance with relative loss of the allele bearing thymidine (c=cytosine; t=thymidine). The normal samples show that the two polymorphic bases are of equal intensity, while the two tumor samples show a relative loss of the thymidine bearing allele.

FIG. 18. Dysplasia of the esophagus of a patient is shown in the left panel. The molecular analysis of the patient is shown in the right hand panel.

DETAILED DESCRIPTION

Novel methods, reagents, kits, and devices that improve the analysis of microdissected tissue, microdissected cytology, and fluid specimens are presented herein. The methods can be applied to DNA extracted from any source, including fresh specimens. However, the methods are especially well suited for those types of small biopsies and minute specimens encountered in clinical practice. Described herein are more effective means of predicting tumor biological aggressiveness, which is essential to planning appropriate therapy. The methods and devices herein can be used on fresh specimens (e.g., biopsied tissue and frozen sections) as well as on minute, fixative-treated, and/or stained specimens.

The aspects described allow molecular pathologic examination of fixative-treated or fresh tissue specimens (or frozen), stained or unstained cytology preparations, and cell-free fluid samples obtained from biopsy and/or needle aspiration of solid tissue or fluid specimens. These biological samples can be very small in size (i.e., measuring less than about 1 mm in greatest dimension for resected tissue). The methods advance the sensitivity and specificity of prediction of the biological aggressiveness of a tumor that currently is based on specific microscopic and selected molecular features. The methods can be combined with existing pathology practice to provide a more detailed view of a particular patient's condition. The methods also describe characterization of a new parameter, dynamic genomic deletional expansion. This parameter is specifically designed to be highly integrated with microscopic analysis.

In clinical practice, specimens are frequently small in size, and it is technically not feasible to perform a detailed investigative examination of all the possible gene and gene associated abnormalities. Rather, one must assess a subset of the large set of total mutational alterations. Having identified one or more specific alterations it is efficacious to characterize mutational damage in greater detail if such information can be shown to be of clinical utility. The effect will be to derive more information that in turn will improve the detection, diagnosis, management, and prevention of a tumor in a reliable, timely, and cost effective manner in the clinical or veterinary setting. Individual patients will benefit greatly by objective diagnosis and significant financial savings will be realized for the healthcare system. The applicability to cell-free fluid specimens is noteworthy, as currently there are limited molecular means available to analyze cell-free fluid specimens. As a result, this type of specimen is usually discarded and with it potentially valuable diagnostic information. Herein is provided an approach that can obtain more information from even these specimens.

Also provided are an analytical platform and methods of use which enhance current improvements in prediction of tumor biological aggressiveness. The analytical platform includes:

(1) techniques to detect sites of genomic deletion;

(2) a means to further analyze the detected deletion to characterize the full extent of genomic loss;

(3) a means to extend the analysis to adjacent areas in tissue specimens (with normal and/or precancerous or hyperplastic tissue) and/or to separate collections of fluid specimens to determine the extent of genomic deletion according to topographic distribution;

(4) a means to extend the analysis to other genomic deletional expansions;

(5) a means to incorporate the determinations derived from a microscopic review of cellular morphology with the data (1) to (4) obtained via the means of;

(6) a means to determine the time course of deletion expansion thereby providing a dynamic understanding of molecular change;

(7) a means to characterize and predict a tumor's biological aggressiveness based on the data obtained by the means of (1) to (6); and/or

(8) to provide quality control measures to monitor the deletion expansion analysis consisting of recommendations for replicate analysis at both the specimen level (sample replication), and replication of the individual marker results (marker replication) so that it may be most effectively applied to clinical specimens.

A preferred analytical platform is designed to correlate with and significantly improve histopathologic and cytologic observation, thereby complementing and improving existing pathology practice.

These methods permit the objective characterization of tumor biological aggressiveness based on DNA genomic deletional expansion determined at multiple sites in a given patient. The overall process can provide an understanding of the genomic deletional expansion, which in turn serves as an excellent predictor of tumor biological aggressiveness.

Quality control assessment of small amounts of DNA is essential due to paucity of diagnostic material; inadequate amounts of starting material produce both false positive and false negative results for mutation detection (Heinrich et al., 2004 Int. J. Legal Med. 118(6):361-3; Schneider et al., 2004 Forensic Sci. Int. 139(2-3): 123-34; and Dreesen et al., 2000 Mol. Hum. Reprod. 6(5): 391-6). For example, extracted DNA from fixative-treated, stained specimens is very low (e.g., below 1 ng for fixative-treated, paraffin-embedded specimens that measure 1 cm in size of less). Although quantitative polymerase chain reaction (qPCR) has been employed on fixative-treated tissue, qPCR assumes high quality starting nucleic acid content. However, it cannot be assumed that fixative-treated, stained specimens of limited size contain high quality nucleic acids. Thus, methods, kits, and devices which address this need for analytic validation and quality control are needed. The methods and devices described herein address these issues.

Current understanding holds that tumor suppressor genes most often undergo a two-step process of gene inactivation (Baker et al., 2003 and Knudson, 2001). The first step is often point mutational damage in a critical region of a gene leading to activation or inactivation of that copy. The second step is the loss of the second “normal” copy of the gene. This can occur by an interstitial genomic deletion encompassing part or all of the gene and the adjacent non-coding DNA (Baker et al., 2003 and Knudson, 2001). The materials and methods described herein assist to characterize this process.

The current understanding views the second step deletion as a one time event, which is stable henceforth. Newly provided herein is evidence that the second step is not stable. In fact, the genomic interstitial deletion can be a dynamic process that progresses leading to a greater widening of the deleted genomic segment. The deletion can undergo further expansion over time in cells clonally derived from precursor tumor cells. Moreover, the pattern and rate of the expansion provides highly useful information with which to characterize a particular neoplasm. Given the reality of clinical practice that one can only look at a subset of the entire set of accumulated mutational change, it becomes highly efficient to more carefully examine the individual detectable deletions rather than attempt to seek new mutational damage in order to better predict tumor biological aggressiveness.

Materials and methods are provide in order to formulate panels of markers based on tetranucleotide (or pentanucleotide, hexanucleotide or greater in length of repeat nucleic acid sequences including minisatellite) repeats targeting genomic regions associated with, but not necessarily associated with, the location of common tumor suppressor genes and not strictly limited to any particular gene. These repeated domains can be any length from 4 nucleotides to 50 nucleotides (and any integer value in between). Methods describe how markers are to be paired for each genomic region but separated by distance to each other in order to provide information concerning the directionality of deletion expansion (genomic deletion expansion). Having defined a set of markers, instructions are provided whereby normal samples are carefully analyzed in sufficient number to enable a thorough understanding of their behavior under conditions of lower amounts of DNA quantity and under conditions when the DNA is degraded. Methods such as comparative template polymerase chain reaction (CT-PCT) and evaluation of fluorescence generation (or other detectable means) are used to quantify marker function under these conditions. Using normal specimens, thresholds for significant allelic imbalance are defined using statistical methods recognizing relevant attributes of markers behavior under varying conditions of DNA quantity and quality. These methods are entirely new and highly valuable to enhance our use of information derived from efforts such as those which have sequenced the human genome.

1. Acronyms and Definitions

Acronyms and definitions as provided herein are art accepted unless indicated otherwise.

1.1 Acronyms

In accordance with this detailed description, the following abbreviations and definitions apply:

    • ADJ adjusted threshold value of significant allelic imbalance
    • AFP alphafetoprotein
    • ALCL anaplastic large cell lymphoma
    • Alk-SMAse alkaline sphingomyelinase
    • ALL acute lymphoblastic leukemia
    • AML acute myelogenous leukemia
    • APC adenomatosis polyposis coli
    • ATM ataxia telangiectasia
    • AVE statistical average value
    • AXIN1 axin 1
    • BAX BCL2-associated X protein
    • BL Burkitt's lymphoma
    • BRCA1 Breast cancer associated gene 1
    • BRCA2 Breast cancer associated gene 2
    • BRCAX Breast cancer associated gene 3
    • CBKB CBKB protein
    • CDKN2A cyclin dependent kinase 2A, also known as p16, DPC3, MTS1, p16(INK4a)
    • CEA carcinoembrionic antigen
    • CEL carboxy ester lipase gene
    • CML chronic myelogenous leukemia
    • CNS central nervous system
    • COL4A5 collagen type 4 alpha
    • COL4A6 colagen type 6 alpha
    • CTCL cutaneous T-cell lymphomas
    • CT-PCT competitive template polymerase chain reaction
    • CYP2A6 cytochrome P450 2A6 gene
    • DCC deleted in colon cancer gene
    • DLC2 Rho GTPase gene (also known as deleted in liver cancer 1 gene)
    • DLCL diffuse large cell lymphoma
    • DMBT1 deleted in malignant brain tumors 1
    • DNA deoxyribonucleic acid
    • DPC deleted in pancreatic cancer gene
    • DPC1/2 Depleted in pancreatic carcinoma 1/2
    • DPC3 Depleted in pancreatic carcinoma 3
    • DPC4 Depleted in pancreatic carcinoma 4
    • DUTT1 Human homologue of the Drosophila axonal guidance receptor gene, Roundabout
    • EGFR epidermal growth factor receptor
    • FHIT fragile histamine triplet gene
    • FAP familial adenomatous polyposis coli
    • FL follicular lymphoma
    • GDE genomic deletional expansion
    • GSC glucocerebrosidase gene
    • GSTM1 glutathione S-transferase mul
    • GSTT1 glutathione S-transferase thetal
    • Her2-neu v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2 (neuro/glioblastoma derived oncogene homolog)
    • HET heterozygosity rate for a particular polymorphic marker
    • IFNA Interferon alpha
    • IGF-1 Insulin growth factor 1
    • KLE6 Kruppel-like factor 6
    • KLF5 Kruppel-like factor 5
    • K-ras Kirsten ras oncogene
    • LMP-1 Latent membrane protein 1
    • LOH loss of heterozygosity
    • LPL lymphoplasmacytoid lymphoma
    • LRP1 B low-density lipoprotein receptor-related protein 1B
    • LRPDIT low density lipoprotein-related protein 1B (deleted in tumors)
    • MCC Mutated in colorectal carcinoma
    • MCL mantle cell lymphoma
      • mdm2 Mdm2, transformed 3T3 cell double minute 2, p53 binding protein
    • MEN-1 multiple endocrine neoplasia type 1
    • MMAC1 phosphatase and tensin homolog (mutated in multiple advanced cancers 1)
    • MTAP methylthioadenosine phosphorylase
      • mtDNAmitochondrial DNA
    • MTS-1 cyclin-dependent kinase inhibitor 2A
    • MYCL L-myc oncogene
    • MYH a/g-specific adenine dna glycosylase
    • MYH11 myosin, heavy polypeptide 11, smooth muscle
    • NF2 neurofibromatosis type 2
    • NKX3.1 homeodomain-containing transcription factor NKX3.1
    • NLCLC non-small cell lung cancer
    • OAZ1 antienzyme
    • OD optical density
    • p14 (ARF) cyclin-dependent kinase inhibitor 2A (melanoma, p16, inhibits CDK4)
    • p15 (INK4b) cyclin-dependent kinase inhibitor 2B (p15, inhibits CDK4)
    • p21 (WAF1) cyclin-dependent kinase inhibitor 1A (p21, Cip1)
    • p73 tumor protein p73
    • Parkin Parkinson disease (autosomal recessive, juvenile) 2, parkin
    • PCBP4 poly(rC) binding protein 4
    • PCR polymerase chain reaction
    • PTEN phosphatase and tensin homolog (mutated in multiple advanced cancers 1)
    • Pter P chromosomal arm terminal region
    • PYGM muscle glycogen phosphorylase
    • qPCR qualitative polymerase chain reaction
    • RCC renal cell carcinoma
    • RUNX3 runt domain transcription factor
    • SD standard deviation
    • Smad4 (DPC4) SMAD, mothers against DPP homolog 4 (Drosophila)
    • SNF5/INI1 SWI/SNF chromatin remodeling complex involved in transcriptional regulation
    • SNP single nucleotide polymorphism
    • STK11 serine/threonine kinase 11 (Peutz-Jeghers syndrome)
    • T-ALL T cell acute lymphoblastic leukemia
    • TCE trichloroethylene
    • TG topographic genotyping™
    • TOP2A topoisomerase Ilalpha
    • TP53 p53 tumor suppressor gene (also known as p53)
    • TU12B1-TY TU12B1-TY protein
    • VHL von Hippel Lindau gene
    • WT1 Wilms' Tumor 1
    • WWOX WW-domain containing oxidoreductase

1.2 Definitions

It must be noted that as used herein, the singular forms “a”, “an”, and “the” include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to “a patient” includes a plurality of patients and reference to “the dosage” includes reference to one or more dosages and equivalents thereof known to those skilled in the art, and so forth.

By “patient” is meant to include any mammal including, but not limited to, bovines, primates, equines, porcines, caprines, ovines, felines, canines, and any rodent (e.g., rats, mice, hamsters, and guinea pigs). A preferred primate is a human. Patients thus can include agricultural animals and domesticated animals under veterinary care. Although the genes provided herein are generally referring to the human gene or disease, it is evident that this can be applied to the cognate animal gene or cognate animal disease.

By “tissue,” and “cells,” is meant to include resected tissue (fixed, stained, or treated), cytology specimens, blood and blood fractions from a patient or from a tissue bank. By “biological sample” is meant to include “tissue” and “cells” as well as “fluid samples” which contains free floating DNA. Free floating DNA is non-nuclear DNA that is in fluid or that adheres to the walls of cells or cysts. The biological samples can be any sample containing DNA or cells from a patient. Such samples include but are not limited to fine needle aspirates, a cytology sample, a blood sample, a spinal tap, resected tissue, frozen tissue, blood sample, fixed tissue, a urine sample, a tissue swab (e.g., buccal swab or pap smear), and the like.

By “genomic deletional expansion” is meant a defined region of the chromosome often containing all or part of a known or putative tumor suppressor gene that undergoes deletion during cancer development and progression. Expansion pertains to the progressive enlargement of the deletional segments of gene during progressive cycles of tumor cell replication. While this definition applies to tumor suppressor genes it may also apply to other genes that function to control cell growth and/or which have not as yet been formally recognized as tumor suppressor genes.

By “genomic amplification expansion” is meant a defined region of a chromosome which contains a gene that undergoes amplification of a segment of the gene during cancer development and/or progression. For example, Her2/neu is a gene wherein a segment of the genome containing the gene is amplified during breast cancer.

By “tumor” is meant to include any malignant or non-malignant tissue or cellular containing material or cells. By “non-malignant tissue” is meant to include any abnormal tissue or cell phenotype and/or genotype associated with metaplasia, hyperplasia, a polyp, or pre-cancerous conditions (e.g., leukoplakia, colon polyps), regenerative change, physiologic adaption to stress or injury and cellular change in response to stress of injury. Tumor is meant to include both solid tumors as well as leukemias and lymphomas. “Neoplasm”, “malignancy”, and “cancer” are used interchangeably.

By “tumor aggressiveness” or “tumor aggression” is meant the phenotypic expression of a malignancy that is associated with increased adverse biological behavior. This includes phenomenon such as capacity for early metastatic seeding, capacity for wide visceral organ dissemination, rapid growth and invasion, lack of treatment responsiveness, short treated disease free interval and short overall patient survival.

By “genetic acquisition due to familial inheritance” is meant a genomic deletion that is part of the patient's germ line. Such mutations occur in all the cells and are not caused by environmental factors nor do they arise spontaneously.

By “genetic acquisition due to an environmental factor” is meant exposure to a carcinogen or mutagen which causes a change in the patient's genome. This can be caused by viruses, food-borne carcinogens, exposure to chemicals (e.g., occupational hazards), or unknown environmental agents. Environmental factors can include but are not limited to trichloroethylene, tobacco smoke (tobacco use or as second hand smoke), arsenic, asbestos, crystalline silica, benzenes, penzo[a]pyrene, beryllium, bis(chloro)methyl ether, 1,3-butadiene, chromium V1 compounds, coal tar, pitch, nickel compounds, soots, mustard gas, erionite, nickel compounds, heterocyclic amines, aflatoxin, vinyl chloride, thorium dioxide, phenacetin, 4-aminobiophenyl, benzidine, 2-naphthylamine, phenacetin, cadmium, cyclosporine A, ethylene oxide, N-nitroso compounds, nitric oxide, antineoplastic agents, and compounds that cause free oxygen radicals. Other environmental factors can include viruses such as certain RNA viruses (e.g., retroviruses such as human T-lymphotropic virus type 1 and type 2, human immunodeficiency virus, and hepatitis C virus) and DNA viruses (such as hepadnaviruses, papillomaviruses, Epstein-Barr virus, Kaposi's sarcoma-associated herpesvirus, simian vacuolating virus 40 and other polyomaviruses). Environmental factors can also include treatment for cancer by radiation or chemotherapy. Radiation exposure is also meant to be included in environmental factors, such as that which is an occupational hazard, sun exposure, and the like.

By “spontaneous mutation” is meant a mutation that occurs during the development of a cancer or a growth that is progressing towards cancer that is not a germ-line mutation. Such mutations can include environmentally induced mutations.

2. Cancer Types

The malignancies and precancerous conditions that can be diagnosed using the materials and methods described herein include solid tumors as well as leukemias and lymphomas. Thus, the cancers include but are not limited to cancers of the head and neck (e.g., nasal cavity, paranasal sinuses, nasopharynx, oral cavity, oropharynx, larynx, hypopharynx, salivary glands, and paragangliomas), lung tumors (e.g., non-small cell and small cell lung tumors), neoplasms of the mediastinum, cancers of the gastrointestinal tract (e.g., colon, esophageal carinoma, pancreatic carcinoma, gastric carcinoma hepatobiliary cancers, cancers of the small intestine, cancer of the rectum, and cancer of the anal region), genitourinary cancers (e.g., kidney cancer, bladder cancer, prostate cancer, cancers of the urethra and penis, and cancer of the testis), gynecologic cancers (e.g., cancers of the cervix, vagina, vulva, uterine body, ovaries, fallopian tube carcinoma, peritoneal carcinoma, and gestational trophoblastic diseases), breast cancer, cancer of the endocrine system (e.g., thyroid tumors, parathyroid tumors, adrenal tumors, pancreatic endocrine tumors, carcinoid tumors, carcinoid syndrome, and multiple endocrine neoplasias), sarcomas of the soft tissues and bone, benign and malignant mesothelioma, skin cancers, malignant melanoma (e.g., cutaneous melanoma and intraocular melanoma), neoplasms of the central nervous system, pediatric tumors (e.g., neurofibromatoses, neuroblastoma, rhabdomyosarcoma, Ewing's sarcoma and peripheral neuroectodermal tumors, germ cell tumors, primary hepatic tumors, and malignant gonadal and extragonadal germ cell tumors), paraneoplastic syndromes, and solids cancers with unknown primary sites. By cancer or neoplasm is also meant to include metastatic disease and reoccurrence or relapse of a cancer(s). Also contemplated are virally induced neoplasms such as adenovirus, HIV-1, or human papilloma virus induced neoplasms such as cervical cancer and Kaposi's sarcoma, and primary CNS lymphoma.

Conditions that create a germline deletion alteration can be included as having the capacity for dynamic genomic deletion expansion. This would encompass inherited genetic alterations, translocations and inherited or somatically acquired DNA damage.

Lymphomas include but are not limited to Hodgkin's lymphoma, non-Hodgkin's lymphoma (B-cell non-Hodgkin's lymphoma and T-cell non-Hodgkin's lymphoma), cutaneous T-cell lymphomas (CTCL), lymphoplasmacytoid lymphoma (LPL), mantle cell lymphoma (MCL), follicular lymphoma (FL), diffuse large cell lymphoma (DLCL), Burkitt's lymphoma (BL), Lennert's lymphoma (lymphoepithelioid lymphoma), Sezary syndrome, anaplastic large cell lymphoma (ALCL), and primary central nervous system lymphomas.

Leukemias include chronic myelogenous leukemia (CML), acute myelogenous leukemia (AML), acute promyelocytic leukemia, acute lymphoblastic leukemia (ALL), prolymphocytic leukemia, hairy cell leukemia, T-cell chronic lymphocytic leukemia, plasma cell neoplasms, chronic lymphocytic leukemia (CLL), and myelodysplastic syndromes (e.g., chronic myelomonocytic leukemia).

Preferred cancers are those that can be readily screened or found either by visual inspection of the patient's skin (skin cancer), mammography (biopsy of lumps), gynecologic cancers (PAP smears and gynecologic examination), or colorectal examination (identification and removal of polyps). The methods and materials provided herein can also be used on tissue samples that have been surgically resected from the patient (resected bowel, removed lung and liver, or other organs).

The methods and materials described herein can also be utilized in diagnosing non-malignant conditions and rendering a diagnosis of cancer from cell aspirates.

Over time, different markers and genes have been associated with different types of cancers. Listed in Table 1 are a sampling of the types of cancers and associated genes that are frequently found deleted. Markers are used to determine the extent of the gene's deletion. Determination of the marker set to utilize would be known to one of ordinary skill.

TABLE 1
Chromosome
CancerGeneLocation*
Gastrointestinal Cancers
Colorectal cancerAPC5q23
p5317p13
DCC18q25
Rb13q14
Her2-neu17q12
MYH
Smad4(DPC4)18q21.1
Esophogeal cancerLRP1B
tp5317p13
FHIT3p14.2
p16/MTS19p21
CDKN2a (p16(INK4a)9p21
p14(ARF)9p21
CYP2A6
p15(INK4b)
WWOX16q23.3-24.1
Stomach cancerp21(WAF1)
Pancreatic cancerK-ras12p15
EGFR7p11
DPC1/213q12
p16/MTS-19q21
DPC418q21.1
TU12B1-TY12q22-q23.1
RUNX31p36
Liver cancerAFP
WWOX16q23
p5317p13
mtDNAmitochondrial
p14(ARF)9p21
DPC213q12
RUNX31p36
p16-INK4A/p14-ARF9p21
FHIT3p14.2
AXIN1
PTEN10q23.31
NF222q12
STK11
BAX
LRPDIT
CholangiocarcinomaCEA
TP5317p13
p169p21
p159p21
Biliary tree cancerCOL4A5
COL4A6
DPC418q21.1
p14(ARF)9p21
p16(INK4a)9p21
p5317p13
Small intestine cancerMEN-111q13
APC5q23
Colon cancerAPC5q23
IGF-1
Alk-SMAse
Lung Cancer
Non-small cell lung cancerRb13q14
p169p21
p14(ARF)9p21
p5317p13
EGFR7p11
FHIT3p14.2
APC5q23
MCC5q25
DMBTI10q25.2-26.1
CDKN2a9p21
PCBP43p21
DUTT13p12
WWOX16q23.3-24.1
VHL3p26
IFNA
MTAP9p21
CYP2A6
Small cell lung cancerGSTM1
Rb13q14
p169p21
p5317p13
FHIT3p14.2
APC5q23
MCC5q25
DMBTI10q25.2-26.1
CDKN2a9p21
DUTT13p12
CYP2A6
MTAP9p21
PCBP43p21
Head and Neck Cancers
Cancers of the nasal cavity,p16INK4A9p21
paranasal sinuses,p14(ARF1)
nasopharynx, oral cavity,
and oropharynx
Cancers of the larynx andGSTM1
phyopharynxp15/p16 co-deletion9p21
BreastIGF-1
BRCA117q12
Rb13q14
p169p21
CDKN29p21
MTS-19p21
p537p13
DUTT13p12
Parkin6q25-q27
WWOX16q23.3-24.1
BRCAX
TOP2A17q12-q21
FHIT3p14.2
ATM
GSTT1
GSTM1
Cancers of the Genitourinary System
Renal cancerp169p21
VHL3p26
SNF5/INI1
WT1
p537p13
PTEN/MMAC110q23
Bladder cancerH-Ras11p15.1
p537p13
mdm212
Rb-113q14
DCC18q25
MYCL1p34
p16(INK4A)9p21
p14(ARF1)9p21
Prostate cancerc-myc8q24
p537p13
Rb-113q14
PTEN10q23.31
KLF513q21
KLF6
p14(ARF)/p16(INK4A)9p21
co-deletion
NKX3.1
mtDNAmitochondrial
Gynecologic Cancers
Endometrialp537p13
DCC18q25
CDKN29p21
p169p21
p159p21
MTAP9p21
OvarianBRCA117q12
p537p13
BRCA2
c-myc
H-ras
Ki-ras12p15
erbB-217q12
OAZ119p13.3
Parkin6q25-q27
WWOX16q23.3-24.1
Cervicalp169p12
p159p21
Endocrine Cancers
PYGM11q13
MEN-111q13
Mesotheliomas
Malignant mesotheliomap15/p16 codeletion9p21
MTAP9p21
FHIT3p14.2
CDKN2A9p21
p14(ARF)9p21
NF222q12
GSTM1
GSTT1
Skin Cancers
Malignant melanomaCDKN2A (or p16)9p21
PTEN10q23.31
Pediatric Cancers
NeuroblastomaUnknowndel1p32-36
PTEN10q23.31
DMBT110q25.3-26.1
RetinoblastomaRb-1Del13q14
Wilm's tumorWT-1Del11p13
Osteogenic sarcomaDel13q14
MeningiomaDelq22, -22
Lymphomas
Generallyp537p13
LMP-1
6q deletions
Non-Hodgkin'sp731p36
Leukemias
T-ALLp16(INK4A)9p21
p14(ARF)9p21
p159p21
B-CLLp537p13
AMLMYH1116p13.1
CBFB16q22

*The assignment of cytogeneic location may vary depending upon the molecular database used to derive localization information.

3. Methods

The methods described herein are those needed to perform the analysis that will culminate in a more accurate assessment of tumor biological aggressiveness. These methods can be performed on any tissue sample, i.e., resected tissue, drawn blood, cytologic specimen, or fluid specimen (e.g., fluid from a cyst). While there are overall common considerations for all these types of specimens, in fact the specific performance of the analysis will differ. These differences will be described in a subsection for each type of specimen. Preferably, these methods can be used on a combination of specimens such that the data from each specimen can be collated for further validation of the diagnosis and prognosis achieved by the genetic determination. Additionally, if using a microdissected sample, multiple microdissection representing different sections of tumor from the tissue block can be used as controls, and as validation of tumor aggressiveness.

3.1 Techniques To Detect Sites of Genomic Deletion—Common Considerations

The specimen source, be it tissue sections, cytology preparation, and/or fluid containing DNA, is treated in such a fashion that DNA becomes available for nucleic acid amplification. This can involve extraction of DNA or treatment of a rude lysate in such as manner as to render the DNA accessible for amplification. This will vary according to the specimen's type and is described below. Any system that renders DNA into a state suitable for nucleic acid amplification will prove sufficient. Those techniques that accomplish this with minimal effort and rapid time, in general, will be preferred. For example, DNA can be analyzed directly using the technique of Topographic Genotyping, which is designed for clinical use with fixative treated, stained cellular specimens. See U.S. Pat. No. 6,340,563 which is incorporated by reference herein in its entirety for all purposes.

Having rendered the DNA in the specimen suitable for nucleic acid amplification, the following common steps are performed to more simply and directly detect and characterize dynamic genomic deletion expansion. The extracted pure DNA or the crude lysate DNA is treated proteinase K at a concentration of 10 mg/mL at 37° C. for 2 hours. The aliquot of sample containing DNA, generally measuring 1-2 microliters, is added to PCR reaction mixture containing Taq polymerase (or suitable substitute enzyme), deoxyribonucleotides and buffer. Of note is the addition of additional magnesium chloride to a level of 5 mM and sucrose to a concentration of 30%. These two additions, not standard for DNA amplification, will enhance primer binding to fixative treated or poor quality DNA and will also enhance DNA polymerization of target DNA.

Oligonucleotide primers (or other DNA source designed to prime DNA polymerization) are used in nucleic acid amplification reactions to generate DNA from polymorphic regions of the genomes in proximity to sites of potential deletion. If the sites of deletion are unknown, a broad panel of amplification reactions can be used to search for the existence of such deletion. A useful approach in this regard is to target tumor supporessor genes that have been shown to be associated with deletion in neoplastic lesions of the type being evaluated. Other approaches are not excluded such as the used of random markers to detected unexpected sites of genomic deletion.

Genomic sites of deletion may already be shown to exist in which case the analysis can focus on these regions. A site of genomic deletion can be identified using a variety of methods, such as fluorescent in-situ hybridization with suitable molecular probes or comparative genomic hybridization. Other techniques can also be used such as classical cytogenetic analysis of viable tumor cells. With the site of deletion known, one can consult any one of a large number of public or private molecular biology databases that describe the presence of polymorphic DNA markers for specific genomic regions. A series of such markers can then be used to evaluate the extent of genomic deletion at two or more sites within a given neoplasm.

3.1.1 Specific Methodology for Tissue Sections

In the case of tissue section, microdissection targets are selected on the basis of correlative microscopic features. Serial unstained standard four micron thick histologic recut sections are prepared along with a stained slide for determination of precise tissue targets. The tissue blocks that are selected for sectioning are those that contain the most representative tissue targets for microdissection based analysis. The size and configuration of each microdissection target is guided by histopathologic features to provide maximum representativeness and purity of cellular constituents. Less precise formats of target selection such as the designation of targets that are larger in size, less correlated with specific histopathologic features and inclusive of undesirable unintended cellular elements serving to dilute and diminish the effectiveness of desired target cell molecular characterization can be substituted recognizing that such a format can obscure part of all of the collective molecular alterations.

Each microdissection target is manually removed using a pointed scalpel under stereomicroscopic observation. Alterative methods for tissue target removal such as equipment assisted tissue sampling (i.e., laser capture microdissection and other methods) or manual tissue removal can be substituted with varying degrees of effectiveness.

The microdissected tissue is placed into a tube containing 25 microliters of lysis buffer composed of 1% NP-40 in dilute, 10 mM Tris, pH 7. After treatment with proteinase K as describe above, the sample is pelleted at 10,000 rpm in a table top centrifuge. The aliquot for nucleic acid amplification is taken from just above the pellet.

3.1.2 Specific Methodology for Cytology Specimens

The methodology is very similar to that described for tissues. Instead of unstained slides as is used for tissue sections, the pap smear stained cytology smear of liquid cytology slide is used for microdissection. Prior to removing the coverslip by immersion in xylene or other solvent, the precise cellular collections are marked with ink on the reverse surface of the glass slides. This will enable the targets of cells to be identified after the coverslip has been removed in order to allow access to the cells of interest.

The cells are removed by manual microdissection in exactly the same manner as for tissue sections. The microdissected cells are then placed into tubes and prepared as for tissue microdissection. Of note is the importance not to overfill the tube with microdissected material as this is cause the subsequent nucleic acid amplification reaction to fail. The amount of cytology microdissection should be large enough to produce a pellet after centrifugation that is barely visible in size by naked eye examination.

3.1.3 Specific Methodology for Fluid Specimens

Fluid specimens undergo direct DNA extraction by filter collection or other suitable isolation methodology such as DNA precipitation, glass bead separation or other technique. The collected DNA is resuspending in a small volume after which it may be quantified by optical density measurement. The DNA is now ready for nucleic acid amplification.

3.1.4 Specific Methodology for Non-fixative treated Specimens

Nonfixed tissue specimens could exist in several forms. One common form is that of frozen sections of fresh tissue which are handled in an identical fashion to that for fixed tissue sections. Another form is that of pieces of fresh/frozen tissue that require mincing followed by homogenization in order to render the DNA amenable to further isolation. Any standard protocol to extract DNA from a tissue specimen would be suitable.

3.2 Characterization of Extent of Genomic Deletion and Detection of Deletional Expansion

A series of polymorphic markers are selected which may be in the form of microsatellites or single nucleotide polymorphisms or any other polymorphism that enables the status of each individual allele of genomic DNA to be selectively recognized. The concept here is to utilize a series of genomic markers that vary in their specific localization at a gene or genomic site of deletion. The presence of absence of the amount of each polymorphic marker will be used to define the length of genomic deletion. The more serially placed markers the show loss, the greater in length will be the linear length of the genomic deletion.

In order to accomplish an analysis for deletion expansion, at least two separate topographic sites of analysis must be performed so that the results may be compared between the various targets available in a given case. These markers are designed to cover the location of the deletion and adjacent DNA both on the centromeric and telomeric side. A representative example is shown in Table 2 for APC. A similar approach can be applied to any gene of interest across all chromosomes. The source of specific markers can be obtained from molecular biology databases in the public domain (Internet: <URL:www.nih.gov>).

Although the specific markers may vary according to the genomic roadmap, the results will be the same. The greater the number of markers used which cover the DNA on either side of the genomic deletion, the more precise the existence and extent of genomic deletion expansion can be characterized. The size of the deletion can be as short as a single base deletion which is the resolution capacity of most automated capillary electrophoresis units. The size of genomic deletion will vary according to the position of the polymorphic markers used. Markers that are close to each other in genomic distance will provide quantitative information on small genomic deletions. More widely spaced markers can be used when the deletional areas are larger in size. Thus, any number of markers may be used. In Table 2, four markers are shown for simplicity, 2 on either side of the deleted gene. However, any number of markers can be used. The distance relationship between individual deletion markers is generally available in most databases in the public domain (Internet: <URL:www.nih.gov>).

Operating in parallel as a means to define genomic deletion expansion are quantitative values for the degree of deletion according to individual topographic targets. Deletions that affect a larger number of microdissected cells or DNA at a given site indicate that the deletion has expanded to involve that increased proportion of cells. Thus both the topographic distribution and the quantitative degree of LOH deletion are used to define the rate of genomic deletional expansion.

Allelic imbalance analysis (also commonly referred to as loss of heterozygosity analysis) is then performed for each of the polymorphic markers. This involves PCR followed by electrophoresis to determine the balance status for polymorphic alleles. Manufacture's instructions with respect to fragment analysis should be consulted and followed for optimal performance of the deletional analysis (i.e., ABI 3100 Genetic Analyzer, Applied Biosystems, Foster City, Calif.). The cumulative information for all of the topographic targets with respect to each of the polymorphic markers used to assess deletion is then performed. By comparing the results for each marker at the various topographic sites it is possible to define a process of deletion expansion associated with topographic distribution of microdissection targets. Fluid and cytology samples are handled in a similar way with separate targets obtained from different cytology slides and separate fluid specimens.

Allelic imbalance refers to the degree of LOH present for a given microdissection target and for a specific individual mutational marker. Allelic imbalance is preferred over either deletion or amplification since either of the latter two processes could produce an identical value for imbalance. The threshold for significant imbalance is defined as a degree of imbalance that exceeds the range of variation in normal sample allele peak height ratios with 95% confidence. This must be determined for each individual marker and for each combination of alleles since the ratios are specific to each pairing. This requires a sufficiently large patient cohort that encompasses all known allele combinations in adequate numbers of subjects to derive 95% confidence intervals for the variation in peak height ratios. Any system of allelic imbalance analysis can be substituted and will yield the same results. At the conclusion of this step, the extent of the genomic deletion(s) analyzed will be defined according to the status of the location of the various polymorphic markers.

TABLE 2
POLYMORPHIC MICROSATELLITE MARKERS SPANNING A REGION WITH A
GENOMIC DELETION CONTAINING THE APC TUMOR SUPPRESSOR GENE
CENTROMERIC TO THE APCTELOMERIC TO THE APC
GENETUMORGENE
DISTALPROXIMALSUPPRESSORDISTAL
MICROSATELLITEMICROSATELLITEGENEPROXIMALMICROSATELLITE
DISTANCEDISTANCEDISTANCEMICROSATELLITEDISTANCE
pterpterpterDISTANCE pterpter
(megabases)(megabases)(megabases)(megabases)(megabases)
111.10111.85112.10112.20112.40
D5S2027D5S19655q22.1 APCD5S346D5S1170
 21.60 21.80 21.96 22.10 22.20
D9S736D9S9169p21 CDKN2AD9S1814D9S966
 89.40 89.50 89.60 89.90 90.00
D10S579D10S239410q23.31D10S541D10S2339
PTEN
7.10 7.25 7.50 7.60 7.70
D17S2159D17S178317p13.1 TP53D17S655D17S1796

The analysis is then carried out quantitatively to determine the proportion of cells and/or DNA affected by imbalance. The following description can be used for microdissected tissue samples. An equivalent method can be used on cytology cells and DNA extracted from free-fluid.

3.2.1 Nucleic Acid Amplification

For quantitation, amplification is performed as follows. An aliquot of the extracted DNA is removed for PCR amplification of individual polymorphic microsatellite markers. Preferably, only about 1 μL is utilized but other sized aliquots of DNA can also serve equally well. Nucleic acid amplification can be carried out according to methods known in the art, e.g., GeneAmp kit, Applied Biosystems, Foster City, Calif. Other variations on the PCR reaction can be substituted. Fluorescent labeled oligonucleotide primers can be employed for quantitative determination of allelic imbalance based on the peak height ratio of polymorphic microsatellite alleles.

Each peak height ratio will be generated for all specific markers from each of the different microdissection targets. In normal samples without deletional damage, these ratios will vary around a mean value. When ratios are derived from test samples that fall beyond two standard deviations about the means, allelic imbalance or LOH is said to be present. It is also possible to use the value of the imbalance to derive the proportion of cells affected by deletion. When compared between two or more topographic targets it is then a simple matter to detect genomic deletion expansion. Other labeled primers can also be substituted as can other quantitative systems for PCR and/or qualitative approaches.

3.2.2 Allelic Imbalance Determination

Post-amplification products from Section 2.2.1 are electrophoresed and relative fluorescence determined for individual allele peak height (using for example a GeneScan ABI3100, Applied Biosystems, Foster City, Calif.). The ratio of peaks are calculated by dividing the value for the shorter-sized allele by that of the longer sized allele. Thresholds for significant allelic imbalance have been determined beforehand in extensive studies using normal (non-tumor) specimens representing each unique pairing of individual alleles for every marker used in the panel. Peak height ratios falling outside of two standard deviations beyond the mean for each polymorphic allele pairing are assessed as showing significant allelic imbalance. In each case, the non-tumor (normal) tissue targets are used to establish an informativeness status, and then to determine the individual pattern of polymorphic marker alleles. Once significant allelic imbalance is established, it is then possible to calculate the proportion of cellular DNA that is subject to hemizygous loss. For example, a polymorphic marker pairing whose peak height ratio is ideally 1.00 in normal tissue with a standard deviation in non-neoplastic tissue of 0.23, could be inferred to have approximately 50% of its cellular content affected by hemizygous loss, if the peak height ratio was 0.5 or 2.0. This requires that a minimum of 50% of the DNA in a given sample be derived from cells possessing deletion of the specific microsatellite marker. The deviation from the ideal normal ratio of 1.0 indicates which specific allele is affected. In a similar fashion, allele ratios below 0.5 or above 2.0 can be mathematically correlated with the proportion of cells affected by genomic loss. Other algorithms for quantitative determination of allelic imbalance can be used with equal effectiveness.

At this point, the proportion of cells or DNA accounting for the imbalance can be determined for each marker. This information is used with the information derived from microscopic analysis of the tissue section or cytology cells from where the samples originated. Particular attention here is paid to the extent of non-neoplastic cell inclusion, as these cells can provide normal DNA to the analysis. The inclusion of normal cellular elements is not limited and is only mentioned in order to provide greater understanding for the quantitation of allelic imbalance. This provides a baseline value for the maximal degree of allelic imbalance since normal cells and their normal DNA will contribute to the ultimate result for allelic. The inclusion of some degree of normal (non-tumor) cellular elements will not interfere with the final characterization of genomic deletion expansion. The proportion of cells or DNA affected by an imbalance can then be determined for markers pertaining to a specific genomic deletion (Table 3). A gradient of reduced mutation involvement indicates that a proportion of tumor cells with an expanded deletion is present in the area from which the sample was derived. This analysis is sensitive for detected genomic deletion expansions up to the threshold representing two standard deviations above the mean for normal allele peak ratios.

TABLE 3
REPRESENTATIVE RESULTS FOR AN ANALYSIS OF GENOMIC DELETION
FOR THE APC GENE
CENTROMERIC TO THE APCTELOMERIC TO THE APC
GENEGENE
DISTALTUMORDISTAL
MICROSATELLITEPROXIMALSUPPRESSORPROXIMALMICROSATELLITE
DISTANCEMICROSATELLITEGENEMICROSATELLITEDISTANCE
pterDISTANCE pterDISTANCE pterDISTANCE pterpter
111.10111.85112.10112.20112.40
D5S2027D5S19655q22.1 APCD5S346D5S1170
57%83%62%NO
IMBALANCE

As shown in Table 2, the exemplified APC gene region contains deletion damage. Polymorphic microsatellites are varying distances away from the deleted region and show progressively small proportions of affected cells. Thus, the deletion is shown to be expanding, and the rate of expansion is quantitatively expressed by the diminishing proportion of cells manifesting this change in the topographic sites from where the source of DNA has been obtained. The use of additional markers for allelic balance representing genomic sites at different distance from the gene of interest will provide the user with a clearer understanding on how rapidly the genomic deletion is expanding.

In Table 3, it can be seen that 83% of the tumor cells manifest the deletion situated 0.25 Mb centromeric to the APC gene. This falls to 57% of the tumor cells over the subsequent 0.75 Mb. These parameters can be used to quantify the rate of genomic deletional expansion using the following formulae:
RATEGENOMIC DELETIONAL EXPANSION=Δ% AFFECTED CELLS/DISTANCE
RATEGENOMIC DELETIONAL EXPANSION=17%/0.25 Mb for the initial centromeric expansion
RATEGENOMIC DELETIONAL EXPANSION=26%/0.75 Mb for the further centromeric expansion

Similar calculations can be made for the telomeric direction of deletional expansion. This calculation assumes that the APC gene is in fact the sole source for deletional expansion which is supported by the fact that this gene is centered in the deleted region and that expansion is occurring in both centromeric and telomeric directions. The reader will note that this analysis can be very useful in searching for the precise location of novel genes that may be behaving as tumor suppressor genes with deletional damage. Intragenic determination of copy number using either intragenic microsatellites if present, single nucleotide polymorphisms if known or other polymorphic markers can be very helpful in defining the specific gene responsible for deletion and then deletional expansion.

3.3 Determination of Genomic Deletion According to Topographic Distribution by Extending the Analysis to Adjacent Tissue Specimens or Fluid Specimens

The analysis is similarly carried out for samples taken from different topographic regions of the specimen (Table 4). The term “topographic regions” refers to separate targets for molecular analysis that can be referred to specific distance measurements away from each other or in relationship to certain landmarks. For example, microdissection targets that are separated by a distance of 2 cm are topographically related to each by this distance. The greater the topographic distance between two targets, the greater will be the genomic deletion if it is to be detected as different from each other. Since deletion expansion is unidirectional towards progressively greater expansion, cells towards the periphery or margins of a particular solid tumor will show correspondingly greater deleted regions. The topographic distribution of deleted segments over distance from the center of a tumor provides the diagnosing physician/veterinarian with a quantitative measure of how rapidly the deleted segment is expanding. In Table 4, evidence is provided demonstrating that a higher grade neoplasm possesses a high rate of deletion expansion.

TABLE 4
DYNAMIC GENOMIC DELETIONAL EXPANSION ANALYSIS
APPLIED TO VARIOUS TOPOGRAPHIC SITES FOR A SPECIMEN
DistanceTUMOR
fromMIDWAY
pter (Mb)TOTUMOR
Micro-TUMORPERIPH-PERIPH-
satelliteCENTERERYERY
DISTAL111.1057%65%88%
MICROSATELLITED5S2027
DISTANCE pter
PROXIMAL111.8583%80%86%
MICROSATELLITED5S1965
DISTANCE pter
TUMOR112.10N/AN/AN/A
SUPPRESSOR5q22.1
GENEAPC
DISTANCE pter
PROXIMAL112.2062%78%85%
MICROSATELLITED5S346
DISTANCE pter
DISTAL112.40NO51%60%
MICROSATELLITED5S1170IM-
DISTANCE pterBALANCE

Any system for topographic selection of tissue from a particular specimen can be used for this type of analysis. Microdissection according to histopathologic features is especially well suited (Finkelstein et al., 2003 Hepatology 37(4): 871-9; Mohan et al., 2004 Mod. Pathol. 17(11): 1346-58; Sasatomi et al., 2002 Cancer Res. 62(9): 2681-9; Rolston, et al., 2001 J. Mol. Diagn. 3(4): 129-32; and Hunt et al., 2004 Arch. Pathol. Lab. Med. 128(12): 1372-8).

A variety of mathematical formulae can be created to quantify the rate of expansion taking into account the distance away for the tumor suppressor gene or other gene of interest and the proportion of cells or DNA affected by imbalance. The sidedness (centromeric versus telomeric) of the deletional expansion can also be calculated as described above (Table 4). In fact, the pattern of deletional expansion can also be used to test whether a particular gene is in fact the center of genomic deletion as described above (Table 4). For example if an expanding deletion if found to be present in a particular tumor, the smallest size of the deletion will provide the most circumscribed localization for the position of the mutated tumor suppressor gene that is responsible for the initiation of the process. The topographic target that shows this minimal deleted region will contain the initial focus of neoplastic cells that were subject to mutational change for this tumor suppressor gene alteration. The pattern of the genomic expansion will then inform the user on the three dimensional directionality of clonal expansion emanating from that altered mutated collection of cells.

3.4 Extension of Analysis to Other Genomic Deletional Expansions

The same analysis can then applied to additional markers associated with other genomic deletions (Table 5). For example, information for the APC gene on 5q is combined with data for the TP53 gene on 17p. The first boxed region relates to the APC deletional analysis, whereas the bottom 5 rows of Table 5 relate to the TP53 deletional analysis. While the rate of genomic deletion expansion may not necessarily be the same for all deletions in a given tumor, in general, the more malignant neoplasms show similar degrees of deletion expansions for multiple separate genomic deletions. When a large number of extensive genomic deletions are present, this provides strong support that the neoplasm is intrinsically aggressive, whereas the converse supports indolent biological behavior. Benign neoplasms do not show deletional expansion, while cancers that are intermediate and high-grade manifest this alterations. There is a direct relationship between the extent and degree of genomic deletion expansion and tumor biological aggressiveness.

TABLE 5
DYNAMIC GENOMIC DELETIONAL EXPANSION ANALYSIS
INCORPORATING THE DATA FROM MULTIPLE MARKERS
AND GENOMIC DELETIONS IN A PARTICULAR SPECIMEN
TUMOR
TUMORTUMORPERIPH-
CENTERMIDWAYERY
DISTAL111.1057%65%88%
MICROSATELLITED5S2027
DISTANCE pter
PROXIMAL111.8583%80%86%
MICROSATELLITED5S1965
DISTANCE pter
TUMOR112.10N/AN/AN/A
SUPPRESSOR5q22.1
GENEAPC
DISTANCE pter
PROXIMAL112.2062%78%85%
MICROSATELLITED5S346
DISTANCE pter
DISTAL112.40NO51%60%
MICROSATELLITED5S1170IMBAL-
DISTANCE pterANCE
DISTAL 7.10NONO52%
MICROSATELLITED17S2159IMBAL-IMBAL-
DISTANCE pterANCEANCE
PROXIMAL 7.2564%75%79%
MICROSATELLITED17S1783
DISTANCE pter
TUMOR 7.50N/AN/AN/A
SUPPRESSOR17p13.1
GENETP53
DISTANCE pter
PROXIMAL 7.60NO50%61%
MICROSATELLITED17S655IMBAL-
DISTANCE pterANCE
DISTAL 7.70NONONO
MICROSATELLITED17S1796IMBAL-IMBAL-IMBAL-
DISTANCE pterANCEANCEANCE

Mathematical formulae can be created to quantify the rate of expansion as a function of topographic distance for different genomic deletions (Table 5).
Rate of genomic expansion over distance=Adetectable genomic expansion(Mb)/Δdistance between two specific microdissection targets (mM),

The rate of genomic expansion over distance is directly related to biological aggressiveness.

The sidedness (centromeric versus telomeric) of the deletional expansion can also be calculated as described in the previous section (Table 5). This is accomplished using the same formula but applied to markers providing information either in the direction towards the centrimere or towards the telomeric end of the chromosome on which the marker is located. The information can be used to determine the extent of genomic instability or imbalance that exists for a tumor in a given patient specimen.

3.5 Incorporating Microscopic Determinations for a More Comprehensive Understanding of Deletion Expansion

Having established the presence and rate of deletion expansion, it is very useful to incorporate microscopic cellular analyses in order to derive the greatest understanding of the significance of the deletion expansion. For example, one common use is to distinguish between a low-grade component of cancer that served as the precursor to the subsequent development of a high-grade malignancy versus the infiltrating edge of a high-grade neoplasm that can appear relatively sparse. This can be explained in more detail as follows.

A true low grade precursor component of a high-grade neoplasm would manifest little or no deletional genomic expansion as evaluated using multiple microdissection targets. Thus if one sought the presence of a determined the rate of deletional expansion in sparse and more cellular regions of a tumor, the low grade precursor component would show a lower or absence value compared to the high-grade component.

On the other hand, if the sparsely cellular infiltrating edge of a glioma was evaluated for genomic deletion expansion and compared to a more cellular area, both regions would show a high rate of genomic deletional expansion which may in fact be greater in the sparsely cellular infiltrating edge than in the cellular high-grade areas of cancer. This would establish that the sparsely cellular area is in fact more aggressive in keeping with the infiltrating edge of a malignancy.

This is not the only example, as there are many similar related situations where this information can provide highly discriminating information on cancer behavior, while addressing fundamental diagnostic issues. Table 6 provides other examples where information for dynamic genomic deletional expansion can be used clinically.

TABLE 6
CLINICAL APPLICATION OF DYNAMIC GENOMIC DELETIONAL EXPANSION
Cancer recurrenceIf two tumors possess the same genomic sites of deletional
versus de novodamage, the expansion in the deposit suspected to be a
cancer formationmetastasis provides support that it is in fact a recurrence
and not a de novo tumor. A de novo cancer would be
expected to have a lesser degree of genomic deletional
expansion.
Reaction versusThere is controversy as to whether reactive processes
neoplastic(inflammation, regeneration, hyperplasia, metaplasia) can
acquire mutational damage. Demonstrating that the
deletional mutation is widening provides support that a
neoplasia is present.
Low-grade versusIn borderline cases with threshold levels of acquired
high-grademutational damage, the finding that deletions are undergoing
dysplasiaactive expansion provides support for high-grade dysplasia.
Prognostication ofAs described here, the demonstration of significant
tumor biologicaldeletional expansion provides strong support for increased
aggressivenessbiological aggressiveness.
Characterization ofThe most obvious manifestation of treatment effect is
treatmentnecrosis of the tumor. Not uncommonly, residual viable
responsivenesstumor is found and it is problematic to determine whether the
surviving neoplastic cells have been arrested with respect to
further cancer progression or whether they have in fact
become truly resistant to therapy. The finding of no further
genomic deletion expansion supports the former which the
demonstration of a higher rate of deletion expansion
supports the latter.

3.6 Time Course Determination of Temporal Mutation Acquisition

The time course of mutation acquisition can be derived by defining the topographic distribution of each mutation with earlier mutations manifesting their presence over a wider distance. Later mutations tend to be more focally confined (FIG. 2).

Another means for deriving the timeline of mutation accumulation is based on the proportion of tumor cells with specific mutational damage at a particular location. Mutations present in a higher proportion of tumor cells that are in a particular location were acquired earlier, than different mutations that are present in a lesser proportion. This analysis is then combined with the data from analysis of the expansion of genomic deletions to provide a dynamic understanding of aggregating genomic damage in particular neoplasms. Then the rate of genomic deletion expansion is determined for the various microdissection targets providing quantitative data on the rate of clonal expansion and deletional expansion. The formulae for combining the data is as follows:
RATE OVERALL MUTATION ACCUMULATION=Δnumber of total detectal mutations/Δdistance between two specific microdissection targets (mM).
RATE GENOMIC DELETION EXPANSIONMUTATION 1=Δproportion of cells affected by deletion MUTATION 1/Δgenomic distance separating the markers being used.
RATE CLONAL EXPANSIONMUTATION 1=Δpercentage of microdissected cells bearinMUTATION 1/Δdistance between two specific microdissection targets (mM)
BIOLOGICAL AGGRESSIVITY INDEX is a function of RATEOVERALL MUTATION ACCUMULATION+RATECLONAL EXPANSION+RATEGENOMIC DELETIONAL EXPANSION
These are all parameters which quantify the instability of the deleted genomic segment in relationship with tumor growth. In general, a cancer with rapidly widening genomic deletion that deletes a largers genomic segment and does so over a shorter topographic distance and also shows rapid acquisition of abundant mutational damage will be predictably more aggressive. The higher the parameters are found to be, the greater certainty there will be that the tumor will be biologically more aggressive.

Optionally, the cumulative information is organized in a suitable display form so that it may be reviewed and conclusions drawn concerning the significance of the findings by the diagnosing physician. Preferably, the data can be coupled with various treatment options associated with the patient's diagnosis to permit the attending physician to discuss with the patient. An example of such as display is shown in the examples below.

3.7 Providing Quality Control Measures to Monitor the Deletion Expansion Analysis Consisting of Recommendations for Replicate Analysis of the Specimen and Well as Replication of the Individual Marker Results

As in the case of all clinical analysis, it is important to have quality control measures in place to ensure that recorded results are accurate and reproducible. There are multiple layers of quality control to monitor analytic precision. First, the taking of multiple topographic samples provides a means for separate sample validation (see FIG. 3). Quality control measures are shown in FIG. 3 for cytology preparation however equivalent methods can be applied to tissue specimens and to multiple samples of aspirated fluid or extracted DNA. Samples obtained close to or from overlapping defined sites are more likely to be consistent than samples taken from widely separated regions of a tumor.

The second level of quality control is replication of individual marker analyses (FIG. 3). The same marker performed in triplicate and should yield equivalent quantitative results. The degree of variation may be used as a means to assess analytic precision with respect to marker replication. For determination of deletion expansion, the availability of additional polymorphic markers within the observed segment of expansion is a useful means to independently corroborate the results.

The electropherograms shown in FIG. 4 demonstrate that replicate analyses yield virtually identical results (FIG. 4). In FIG. 5, allelic imbalance is demonstrated due to inadequate amounts of amplifiable DNA. The issue of inadequate DNA should be minimized otherwise it can lead to false positive values for allelic imbalance. False positives will in turn affect the determination of the dynamic mutational profile. Replicate analysis is the most direct way to identify this problem.

A preferred analytical validation method can be achieved by comparing the molecular analysis from biopsy a sampling of a particular tissue or organ to that obtained from analysis of microdissected samples from the resected specimen. A high degree of concordance should be evident using this methodology.

4. Applicability to Detection of Environmental Carcinogens and Environmentally Induced Mutations

The genomic expansion technology can be used to determine whether a cancer or benign growth is a germ line growth or somatic cell related mutation. Additionally, the methods and materials described herein can be used to determine whether a mutation is induced by an environmental factor (i.e., chemical exposure, radiation exposure, exposure to food-borne toxin, viral, and the like).

4.1 Trichloroethylene-induced Cancers

An example of the use of genomic deletional expansion is in the investigation of cancer formation associated with trichloroethylene (TCE) exposure. The carcinogenicity of TCE is much debated and has remained unresolved. There are published reports causally associating TCE and related chlorinated solvents with unique forms of mutational change to specific genes. The best described of these is to the von Hippel-Lindau gene, which is situated on the short arm of chromosome 3, specifically at 3p25.3, 10.16 megabases from the p terminus. In patients with TCE exposure, the most common form of cancer that is seen to develop is that of renal cell carcinoma (RCC). TCE induced RCC appears similar histologically to RCC arising sporadically in patients not exposed to this chemical.

VHL gene damage is common in sporadic RCC, however the precise location of genetic alteration has been described as being unique and distinct in subjects exposed to TCE. More specifically, point mutation is more common and tends to affect exons 1 and 2 of the gene compared to sporadic RCC, where VHL point mutation is relatively less common and if present, is observed affecting exon 3 and the distal part of the gene sequence. Also, in TCE exposed patients with RCC, multiple hits of the VHL have been noted, which is very uncommon in sporadic RCC. Finally, damage to other genes situated on chromosome 3p have been reported as unique to TCE exposure.

By utilizing markers distributed along 3p and taking advantage of the insights derived from genomic deletional expansion, the impact of TCE induced mutation was with its related impact on carcinogenicity. Table 7 outlines a unique marker panel which assesses the issue in patients exposed to TCE. Seven polymorphic microsatellites were used that are distributed across the 3p chromosomal arm from 3p12.3 to 3p26.3. Between the polymorphic microsatellites, specific potential TCE-associated genes with growth regulatory properties are listed in their natural order along the 3p arm. In turn, single nucleotide polymorphisms (SNPs) within each gene are listed when such information is available. This listing is not complete but sufficient to document the presence of genomic and gene deletion to certain genes of interest that may be involved in TCE associated damage. Thus, alternative polymorphic microsatellites and SNPs can be used as they are determined. Note also that the VHL gene is included in Table 7. Primers sequences are shown for effective PCR amplification with the understanding that alternative primer sequences can effectively be substituted. Primers are listed in 5′ to 3′ orientation. In Table 7, “dist” refers to the distance from the p terminus to the location of the marker, and “cyto” refers to the cytogenetic location.

TABLE 7
MARKER PANEL TO STUDY THE RELATIONSHIP OF TRICHLOROETHYLENE EXPOSURE TO HUMAN CANCER FORMATION
GENEMARKERCYTOdistForward oligonucleotide primerReverse oligonucteotide primer
D3S15393p26.31.05CTCTTTCCATTACTCTCTCCCTGGGTAAAAGTAATCCTGG
OGG1SNP intron 43p25.39.76CCAGAGTGAAGGAGAAAGCGTGCCACATATGGACATCC
OGG1SNP exon 73p25.39.76CACOCTCCCAACACTGTCACCTTGGGGAATTTCTTTGTCC
VHLSNP nt 193p25.310.16ATCGCGGAGGGAATGCCCTCGACGCCTGCCTCCTCC
VHLSNP rs16427423p25.310.16CCGCTACGGATGTAGAATGGGTATTTATCAGGAGAAGGTGGTGG
VHLEXON 1 dist seq3p25.310.16AACTGGGCGCCGAGGAGGGCCCGTGCCAGGCGGCAG
VHLEXON 2 seq3p25.310.16CCTTTGCTTGTCCCGATAGCTTACCACAACAACCTTCTC
VHLEXON 3 prox seq3p25.310.16CTTGTACTGAGACCCTAGTCTGCCATCCGTTGATGTGCAATG
XPC3p25.114.16
D3S23033p24.317.92GTGCCTACATGTTAGTATCCCTCCAGAGCTTTGTTTTCAAC
RARBSNP rs11705833p24.225.18CCTGTGGGTGAAATTTGAGCCGGAAGAAAAGGAAGATCAGCAAAG
RARBSNP rs17032183p24.225.18CCATTCCTTCTGGGAAGCTCTGGATTTTGAGTAACTCAATGGCTC
RARBSNP rs76319793p24.225.18TTGGTTGTGTGTTTGAATAGGCCCTCAGTGACTACATCTGC
D3S23273p24.225.54TAGGTAGATAGATAGATAGGTAACCAGAAGACTTTAATAATTTGAG
TOP2BSNP rs10583783p24.226.62CTGGCTCTGTTTGTACATTGAGATTGTGTAGTAAAAGCTCCCACACACC
TOP2BSNP rs22937873p24.226.62TGGCTGGCTCTGTTGCTGCTATGTGTAATTATCATTCTTCAAGAC
TOP2BSNP rs37366013p24.226.62GTACATGATGCATACATATGCAAAACCCATATATGGTTTTTTACC
D3S17453p2333.22TGCTCAGACTTGTGGGTGTCTCTTACCTATCGTGATACACCG
D3S23043p22.142.79CACCACTGCACTCCAGCCCAATATGGGGGTGGGAGG
BLU3p21.350.36
H373p21.3NA
RASSF1A3p21.3NA
D3S15403p14.259.50ACCTGTCCAAGCACCAAAGGGTGGCTGTTCCTGTTGGTATCTGG
FHIT3p14.259.72
FOXP13p1371.09
D3S15423p12.378.58GGAGTAGGAGGTTTCAGTGAGGGGAATCTACCTGACAAACAGGCTC

An analysis of 5 patients with well documented TCE exposure in the work place and with different forms of human cancer are shown in Table 8. The presence of allelic imbalance, reflecting quantitative genomic loss, applied to both microsatellites and single nucleotide polymorphisms, would enable a delineation of the temporal sequence of mutation accumulation in these patients. Results for 5 patients is shown in Table 8.

The assessment of mutations arising from TCE exposure can be utilized with any environmental causative. The technique can be utilized to determine whether diseases are caused, for example, by occupational hazards or by some other causative. Thus, the results applied to the patient pool with TCE exposure can be utilized for any of the environmental factors listed herein.

TABLE 8
RESULTS OF MUTATIONAL ANALYSIS
GENE MARKERCYTO dist PATIENT 1
RENAL CELL CANCER PATIENT 2
COLON CANCER PATIENT 3
LUNG SQUAMOUS CELL CANCER PATIENT 4
UROTHELIAL NEOPLASIA PATIENT 5
RENAL CELL CANCER
D3S15393p26.31.05NO LOHNO LOH73%NI51%
OGG1SNP exon 43p25.39.76
OGG1SNP intron 73p25.39.76
VHLSNP nt 193p25.310.16NI44%NO MUTATION50%NI
VHLSNP rs16427423p25.310.1656%48%NINI66%
VHLEXON 1 dist seq3p25.310.16NO MUTATIONNO MUTATIONNO MUTATIONNO MUTATIONNO
MUTATION
VHLEXON 2 seq3p25.310.16CDN 115 C HIST-GLUTNO MUTATIONNO MUTATIONCDN 81 PRO-SERCDN 81
PRO-SER
VHLEXON 3 prox seq3p25.310.16NO MUTATIONNO MUTATIONNO MUTATIONNO MUTATIONNO
MUTATION
XPC3p25.114.16
D3S23033p24.317.92NO LOH56%63%NO LOHNI
RARBSNP rs11705833p24.225.18NINININO LOH
RARBSNP rs17032183p24.225.18NI49%NININO LOH
RARBSNP rs76319793p24.225.1880%NI77%69%NI
D3S23273p24.225.5496%35%80%95%93%
TOP2BSNP rs10583783p24.226.62NINILOH LOWNINI
TOP2BSNP rs22937873p24.226.62NINO LOHNININI
TOP2BSNP rs37366013p24.226.62NININI57%84%
D3S17453p2333.22NO LOHNO LOHNO MUTATION64%80%
D3S23043p22.142.79NO LOHNINO MUTATION50%NI
BLU3p21.350.36
H373p21.3NA
RASSF1A3p21.3NA
D3S15403p14.259.50NI52%NI44%53%
FHIT3p14.259.72
FOXP13p1371.09
D3S15423p12.378.5845%NINO LOHNO LOHNO LOH

Microsatellite denoted by gray shading;

single nucleotide polymorphisms denoted by absence of shading.

Specific alleles (copy 1 versus copy 2) are denoted by horizontal and vertical hatching respectively.

The following acronyms are utilized above:

LOH = loss of heterozygosity;

NI = noninformative;

BLU = BLU gene;

TOP2B = topoisomerase 2 beta;

RARB = retinoic acid receptor beta;

XPC = XPC gene;

Prox = proximal;

Dist = distal;

nt = nucleotide;

seq = sequence; and

“CDN 115 C HIST-GLUT” = codon 115 histidine-glutamine

While VHL gene deletion and unique point mutation is present in certain patients (i.e., patients #1, #4 and #5 of Table 8), all the patients were observed to manifest an allelic imbalance at 3p24.2. Using genomic deletional expansion to define the origin of temporally early gene deletion, two genes at 3p24.2 were observed to be responsible. These consist of retinoic acid receptor beta and topoisomerase 2 beta. These two genes can be seen to play a critical role in early TCE-associated gene damage and cancer development. Moreover, deletional expansion into the region occupied by VHL may be responsible for analytic evidence of VHL deletion without the finding of point mutational change. By microdissecting lesions at multiple locations and then quantifying the proportion of cells affected by deletion mutation, valuable information can be gathered to assist in defining the molecular pathogenesis of certain forms of human cancer.

4.2 Application of Genomic Deletional Expansion to Resolve the Issue of Primary Versus Metastatic Tumor Formation

Not uncommonly in clinical practice, a patient presents with evidence of cancer at a particular organ site and the issue arises as to whether that single tumor is a primary at that location or metastatic to that location from another primary site of formation. While there may not be overt evidence of tumor in another location, the issue remains that an occult neoplasm may have spread to a site, which has been recognized as bearing cancer. At this point, a dilemma exists as to whether that is primary or metastatic in the site in which it is identified for purposes of rendering an accurate diagnosis and appropriate treatment plan. This is particularly true in instances where the cancer is detected at sites that are uncommon sites for primary cancer formation.

Typically, this problem is addressed by immunohistochemical staining. But often the issue remains unresolved after that form of testing. It is not uncommon in clinical practice to resort to a best guess in order to institute therapy which in turn is likely not to succeed. Additionally, after failure of the “best guess therapy”, it will remain unclear what the resolution is absent the patient's death with a complete autopsy.

Genomic deletion expansion is well suited to resolving this issue. If the tumor is metastatic to the site where it is detected, the advanced nature of tumor progression will declare itself by the presence of abundant concordant detectable mutational change each of which will show evidence of deletional expansion. On the other hand, if the cancer is primary at that site of formation, then multiple microdissection targets of the tumor will show a spectrum of accumulated mutational change with some sites showing few mutations and others showing progressively increasing amounts of mutations. Additionally, there will be only mild degrees of deletional expansion in one or more sites in keeping with the fact that tumor present at those sites are early and therefore the primary site in formation.

As exemplified in Table 9, a patient presented with a tumor mass of the lip. The lip is an uncommon site of cancer formation. Therefore, a metastasis to the lip from an occult primary was suspected, yet no other primary tumor was identified. The tumor displayed microscopic features that suggested possible metastatic origin in lung or gastrointestinal tract. Immunohistochemical staining was inconclusive for determination of the primary site of formation.

TABLE 9
ADENOCARCINOMA OF THE LIP: PRIMARY LIP CANCER VERSUS
METASTASIS FROM AN OCCULT SITE
TUMORTUMORTUMORTUMER
Mb distanceAREA 1AREA 2AREA 3AREA 4
DISTAL111.10NONONONO
MICROSATELLITED5S2027IMBALANCEIMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL111.85NONONONO
MICROSATELLITED5S1965IMBALANCEIMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
TUMOR112.10N/AN/AN/AN/A
SUPPRESSOR GENE5q22.1
(APC)APC
DISTANCE pter
PROXIMAL112.20NONONONO
MICROSATELLITED5S346IMBALANCEIMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL112.40NONONONO
MICROSATELLITED5S1170IMBALANCEIMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL 21.60NONONONO
MICROSATELLITED9S736IMBALANCEIMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL 21.8054%NO63%NO
MICROSATELLITED9S916IMBALANCEIMBALANCE
DISTANCE pter
TUMOR 21.96N/AN/AN/A
SUPPRESSOR GENE9p21
(CDKN2A)CDKN2A
DISTANCE pter
PROXIMAL 22.10NONONONO
MICROSATELLITED9S1814IMBALANCEIMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL 22.20NONONONO
MICROSATELLITED9S966IMBALANCEIMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL 89.40NONONONO
MICROSATELLITED10S579IMBALANCEIMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL 89.5044%NONONO
MICROSATELLITED10S2394IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
TUMOR 89.60N/AN/AN/A
SUPPRESSOR GENE10q23.31
(PTEN)PTEN
DISTANCE pter
PROXIMAL 89.9089%77%69%72%
MICROSATELLITED10S541
DISTANCE pter
DISTAL 90.0052%NONO45%
MICROSATELLITED10S2339IMBALANCEIMBALANCE
DISTANCE pter
DISTAL 7.10NONONO
MICROSATELLITED17S2159IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL 7.25NONONO
MICROSATELLITED17S1783IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
TUMOR 7.50N/AN/AN/A
SUPPRESSOR GENE17p13.1
(TP53)TP53
DISTANCE pter
PROXIMAL 7.60NONONO
MICROSATELLITED17S655IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL 7.70NONONO
MICROSATELLITED17S1796IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter

As indicated in Table 9, two of four markers from the panel do not show mutational change. The remaining markers of a 20 marker panel were without detectable mutation. Only two allelic imbalance (loss of heterozygosity) mutations are detected. The mutation affecting 9p21 was not detected in all four microdissection sites, but in only two of four sites. There was no detectable genomic expansion. The second mutated marker was situated at 10q23, and while present at all four sites, it showed only minimal genomic deletional expansion. Thus, the lip tumor was in fact a primary tumor at the site of detection in the lip, because the genetic profile of acquired mutations was inconsistent with advanced cancer having undergone dissemination.

5. Materials and Methods for Preparing Marker Panels and Sample Marker Panels Identified Thereby

While there are many forms of mutational damage, the most common and simplest to identify in the form of broad panel analysis is allelic imbalance also commonly referred to as “loss of heterozygosity” (LOH). The recommended approach for detection of LOH is to base the analysis on dinucleotide microsatellites and single nucleotide polymorphisms (SNPs), because these polymorphisms are the two most numerous and will enable the selection of a marker closest to a gene of interest (see e.g., Cavalli et al., 2004 Cancer Genet. Cytogenet. 149(1): 38-43; Park et al., 2003 Anticancer Res. 23(3C): 2955-9; and Teh et al., 2005 Cancer Res. 65(19): 8597-603).

The materials and methods described herein is counterintuitive from current practice, because the methods specifically targets tetranucleotide (or larger microsatellites; in all instances when tetranucleotide microsatellines are discussed, it is meant to also encompass 5-nucleotide to 50-nucleotide microsatellites) microsatellites or large sized repeats including minisatellites as the polymorphic markers of choice. Tetranucleotide microsatellites are far less numerous than dinucleotide repeat microsatellites and thus one cannot, in general, obtain a marker closest to a particular gene of interest. This approach is entirely opposite to the approaches described in the literature for the purpose of assembling marker panels. The literature teaches that the markers closest to the gene of interest should be used. The literatures teaches away from the use of markers that are more distant to the gene of interest.

Yet, set forth herein is the discovery that tetranucleotide, and longer, microsatellites produce less artifact stutter bands when DNA containing microsatellites are amplified by the polymerase chain (PCR) reaction (FIG. 14). Stutter band formation reduces the ability to perform quantitative analysis of DNA content based on peak height content. While specific algorithms can be applied to estimate the effect of stutter band formation on DNA content, this correction factor can produce significant errors, which in turn reduces the ability to achieve quantitative DNA content analysis. Moreover, the minimal stutter bands associated with tetranucleotide formation are more reproducible and can be accommodated much easier by mathematical correction factors (Perlin et al., 1995 Am. J. Hum. Genet. 57(5): 1199-210).

An advantage of dinucleotide repeat microsatellites and SNPs when searching for specific gene imbalance is that they are closer and often time situated within specific genes of interest. This is not the case with tetranucleotide polymorphisms (or larger), which tend to be found outside specific genes. However, carcinogenesis induced genomic loss or gains cannot be so effectively predicted and genes other than the ones being targeting may be in fact responsible for allelic imbalance. Thus, the apparent advantages attributed to dinucleotide microsatellites and SNPs are greatly outweighed by the advantage of more exact quantitative analysis of tetranucleotide repeats.

Having chosen tetranucleotide (or greater) polymorphisms as the marker of choice, contrary to conventional teaching, a set of genomic regions are selected for evaluation of allelic imbalance. The conventional teaching is to select such regions based on the existence of cancer specific associated gene mutations linked to a particular cancer of interest which is being studied. While this recommendation has merit, the novel approach provided herein emphasizes the use of markers that have already been developed and employed to study genomic regions in other forms of cancer. The reason for this is that quantitative data will already be in place to more thoroughly evaluate the significance of threshold levels of allelic imbalance, whereas with the introduction of totally new markers, the process of accruing sufficient quantitative information to be thoroughly familiar with the behavior of a particular marker will be delayed. If complete different markers are used for each and every form of human cancer, then it may never be possible in a reasonable amount of time to accumulate detailed experience with any one marker. As stated before, this is contrary to conventional teaching, which places great emphasis on the uniqueness of individual markers.

Having decided on a list of specific genomic regions, tetranucleotide (or larger) microsatellite markers are selected so that they are not immediately next to each other, but rather separated by a distance of 10-20 megabases (Mb) although other distances can function equally well. This approach is entirely counterintuitive to conventional thought and practice in the selection of markers, where it is strongly recommended that markers that target a particular regions should be as close to each other as possible, preferably next to each other. The reasoning behind conventional practice is that close proximity of both markers can be expected to produce exactly the same allelic imbalance information. In the case of dinucleotide microsatellites and SNPs, this objective can be met due to the greater number of such markers distributed along the human genome, which in turn makes it simple to follow this recommendation.

The methods disclosed herein provide spaced markers in order that genomic deletional expansion (GDE) can be determined in those patients found to be informative for both markers to a certain region. Recognition and quantitative of GDE is extremely valuable when searching for and characterizing allelic imbalance in pathology specimens. In addition to providing evidence for allelic imbalance, the methods provide information on tumor biological aggressiveness, because more aggressive cancer show greater degrees of GDE. Also GDE provides directionality information on the genome. Thus, GDE provides information on where the particular gene is situation that is actually responsible for mutation and not merely assumed to be responsible.

Having created a panel of markers, the next step of the method is to evaluate each marker on a large number of normal samples prepared in the same manner as those biologic samples that are to be evaluated in the context of disease states. Conventional teaching would proceed directly with the analysis of pathologic samples, but in the method described here the analysis of a sufficiently large sample of normal specimens is needed in order to evaluate each marker for their ability to provide quantitative information. Because there is effort involved in accumulating this information on every new marker, one can appreciate the value in using the same markers for different organ panels, as it makes best use of the accumulated data. While each marker must be capable of detecting mutational damage and addresses the clinical needs for which the analysis is being performed, the use of the same marker whenever possible is highly advantageous. Conventional guidelines teach away from this, because conventional guidelines emphasize the fact that each marker panel must be different for each tissue, each organ, and/or clinical situation.

Conventional teaching would not perform the steps herein, which serves as the foundation for quantitative interpretation of mutational results. The standard procedure is to obtain a sample of normal tissue, or DNA containing material, such as biopsied tissue, from the same patient or tissue, and determine the ratio of peak heights for this normal specimen. Then conventional teaching instructs that the ratio of peak height values for polymorphic alleles from the lesional sample be divided by the value so obtained from the normal sample. This approach will not work in many pathologic states, because often a “normal” specimen is simply not available. For example in the surgical treatment of brain tumors, removal of normal tissue is highly discouraged as it could lead to permanent neurological deficits. Also, the medical staff may simply forget to obtain a blood sample while the patient is in their care and since many procedures are now done in the outpatient setting, the patient may no longer be present to obtain a “normal” specimen.

It is common practice in the field of molecular analysis of cancer specimens to obtain a “normal” sample from adjacent, histologically normal appearing tissue (see, e.g., Zauber et al., 1999 Mol. Diagn. 4(1): 29-35). This practice is also not acceptable, because very often such adjacent tissue may actually contain infiltrating cancer cells that will affect the “normal” tissue values. Also, DNA from tumor cells that are turning over more rapidly and releasing their mutant DNA into the extracellular fluid, may migrate into “normal” appearing areas of tissue and be included in the “normal” specimen that ultimately will be used to normalize the lesional values. Free DNA cannot be seen under the microscope, yet its presence cannot be discounted. Application of these concepts into an analytic system to analyze lesional specimens provides for a method that is different than that which is part of conventional pathology practice. Also provided is a novel step to normalize lesional data without the need of an actual normal specimen from the subject at the time when the lesional sample is obtained.

Data is collected on each marker as shown in a representative table below for a single marker. The distribution of specific alleles (columns A1 and A2 representing the two alleles, i.e., allele 1 and allele 2, in individual subjects) is determined for the general population. This information is not available in molecular biology database repositories such as GenBank. For example, the data presented in Table 10 can be compared to that from the human genome project (Internet:<URL:http://www.gdb.org>). In fact, the data present in those sources is based on far a smaller number, which invariably is not as complete as the information provided herein. As shown in Table 10, the total population of subjects upon which this is based is over 400 people (column marked #) and is both unique and essential in defining individual marker characteristics.

TABLE 10
MARKERA1A2R#% [HET]AVESDLOW%HIGH%ADJ
D10S.520-HEX143740968.7%1.41320.33210.749047.0%2.077332.0%2.6664
D10S.520-HEX12210.2%NANA
D10S.520-HEX1381460.998210.2%0.9982NA
D10S.520-HEX1421661.876010.2%1.8760NA
D10S.520-HEX14610.2%NANA
D10S.520-HEX1461541.261210.2%1.2612NA
D10S.520-HEX1461582.037741.0%1.53800.38910.759850.6%2.316333.6%3.1136
D10S.520-HEX1461621.812941.0%1.39960.28550.828540.8%1.970729.0%2.3643
D10S.520-HEX1461701.848710.2%1.8487NA
D10S.520-HEX15051.2%NANA
D10S.520-HEX1501581.333920.5%1.2242NA
D10S.520-HEX1501621.729241.0%1.5817
D10S.520-HEX1501662.086410.2%2.0864NA
D10S.520-HEX154133.2%NANA
D10S.520-HEX1541581.9518215.1%1.42770.32330.781145.3%2.074231.2%2.6094
D10S.520-HEX1541621.9600174.2%1.54300.27920.984636.2%2.101426.6%2.4181
D10S.520-HEX1541662.0904256.1%1.52870.34730.834145.4%2.223331.2%2.8016
D10S.520-HEX1541702.026351.2%1.56450.40270.759151.5%2.370034.0%3.2246
D10S.520-HEX1541742.086430.7%1.4277NA
D10S.520-HEX1541821.713610.2%1.7136NA
D10S.520-HEX158225.4%NANA
D10S.520-HEX1581622.0949409.8%1.47060.39900.672554.3%2.268635.2%3.2156
D10S.520-HEX1581662.0873276.6%1.31380.35270.608453.7%2.019334.9%2.8373
D10S.520-HEX1581701.540171.7%1.32210.18740.947328.3%1.696822.1%1.8451
D10S.520-HEX1581741.873351.2%1.41090.33860.733648.0%2.088232.4%2.7134
D10S.520-HEX162409.8%NANA
D10S.520-HEX1621661.91264511.0%1.33750.30550.726645.7%1.948531.4%2.4622
D10S.520-HEX1621702.0019184.4%1.42230.29750.827341.8%2.017329.5%2.4452
D10S.520-HEX1621741.826982.0%1.48410.34410.795946.4%2.172231.7%2.7672
D10S.520-HEX166348.3%NANA
D10S.520-HEX1661701.9447235.6%1.30020.24550.809337.8%1.791127.4%2.0889
D10S.520-HEX1661741.617292.2%1.24660.18750.871730.1%1.621623.1%1.7828
D10S.520-HEX1661781.586810.2%1.5868NA
D10S.520-HEX1661861.887610.2%1.8876NA
D10S.520-HEX170112.7%NANA
D10S.520-HEX1701741.497651.2%1.19450.19520.804132.7%1.584924.6%1.7744
D10S.520-HEX17410.2%NANA
D10S.520-HEX17810.2%NANA
D10S.520-HEXΔ012931.5%NA
D10S.520-HEXΔ0413332.5%1.37960.33200.715648.1%2.043632.5%2.6597
D10S.520-HEXΔ087618.6%1.37640.31330.749945.5%2.003031.3%2.5264
D10S.520-HEXΔ124912.0%1.49820.31620.865742.2%2.130629.7%2.5926
D10S.520-HEXΔ16153.7%1.50410.35820.787747.6%2.220632.3%2.8722
D10S.520-HEXΔ2041.0%1.54270.56620.410273.4%2.675242.3%5.8021
D10S.520-HEXΔ2420.5%1.8623NA
D10S.520-HEXΔ2810.2%1.7136NA

A1 = shorter length allele 1;

A2 = shorter length allele 2;

R = ratio of peak heights of polymorphic allele bands;

# = number of subjects for each marker;

% [HET] = observed heterozygosity rate;

SD = standard deviation;

LOW = average minus two standard deviations representing lower threshold for significant allelic imbalance;

% = observed allelic imbalance ration converted into a percentage of mutated cells;

HIGH = average plus two standard deviations representing a higher threshold for significant allelic imbalance; and

ADJ = adjusted threshold value of significant allelic imbalance of the higher allele based on the threshold determined for the lower threshold.

For each allele pairing, the ratio of peak heights is determined on each specimen (column labeled AVE in Table 10). These average values form a distribution which can be evaluated mathematically to define ranges of normal values within certain degrees of confidence. Many different forms of statistical analysis can be applied and we have chosen to regard the overall values as having a typical bell curve distribution. This then allows 95% confidence intervals to be defined as two standard deviation around the mean (columns labeled SD, low, and high [2 standard deviations above and below the average value] in Table 10). Values within the range are considered “normal”, because small populations of mutated cells cannot be detected above the background of normal peak height ratio variation. Values beyond the range for normal variation are considered as bearing mutational change in a significant proportion of cells that exceed the minimum threshold for peak height ratio variation.

At this point, the advantages of detailed and accurate quantitative analysis exert great value. The ratio of peak heights for polymorphic microsatellite alleles can be converted into a percentage of mutated cells for the individual sample that has been taken. This turns a value that is difficult to visualize (ratio of peak heights) into a value that has real meaning, i.e. percentage of mutated cells in a given area of a specimen. The conversion is based on a model of allelic deletion as it applies to tumor suppressor genes in which one copy of the target region contains a point mutation or other cause for functional inactivation. The second event is genomic deletion of the remaining normal gene copy to create a total functional deficit of tumor suppressor gene activity. This has been known as the Knudsen hypothesis and is now well established in our understanding of human carcinogenesis.

To calculate the percentage of mutated cells using a knowledge of average value for peak height variation, the following formulae may be used:
Percentage % mutated cells=[1 minus(NAVERAGE/Value)]×100% if the ratio of peak heights for the sample of interest is greater than NAVERAGE
Percentage % mutated cells=[Value/NAVERAGE]×100% if the ratio of peak heights for the sample of interest is less than NAVERAGE
These formulae have never been reported, and their use with calculated normal values is novel. The concepts are not apparent from a reading of the current literature, which teaches away from the concepts used to develop the above formulae. Current teaching states that all lesional values should be divided by the normal specimen value but when the latter is not available, there are no further recommendations.

The ability to detect mutational change in a population of analyzed cells will depend upon the expected variability in the peak height ratios. Using the example described above with 95% confidence limits, values falling outside 2 standard deviations may be considered as bearing mutated cells. The critical threshold value is determined by the distribution of normal values (NAVERAGE) and the statistical algorithm (i.e. 2 standard deviations) that is applied to the distribution to determine what is within the normal range of variation. This approach is not designed to detect one mutated cell amongst hundreds or more of non-mutated cells. Rather the method detects significant clonal expansion of acquired mutations that confer a favorable growth advantage as determined by the ability to exceed values that will fall outside of the normal range of variation of normal values.

Conventional teaching instructs that after selection, the marker can be applied to lesional specimens using the normalizing value of the “normal” sample taken with each case. Before this can be applied to lesional specimens, the issue of limiting amounts of DNA must be addressed. Conventional teaching disregards this consideration, because conventional teaching assumes that there is unlimited amounts of lesional DNA for analysis.

While the distribution of normal values for a given type of specimen are defined as discussed above, the next step required accounting for the effect of DNA quantity and quality. Conventional teaching ignores this important factors and simply recommends that the internal normal sample (which may not exist as described above) be used as a normalizing factor. This is not appropriate when dealing with clinical specimens, which may be small in physical size (i.e. biopsies, cytology cell clusters) and be subject to chemical fixation, which may induce DNA degradation.

If the DNA is present in adequate amount, the phenomenon of allelic dropout may occur. With insufficient DNA, the two allele copies may not be accurately represented in the completed PCR reaction due to asymmetric initial template amplification. Instead of two equal copies being equivalently amplified in the first few critical cycles when template amounts are limiting, one copy is preferentially amplified and interpreted falsely as the sample having allelic imbalance. Thus, the distribution of normal average values (NAVERAGE) may be greatly influenced by the amount of amplifiable DNA present in the sample. Lower absolute amounts of DNA can often create a wider distribution of normal average values, and so requires that suitable adjustments be made in the definition of threshold levels for significant allelic imbalance. Using the statistical format of a normal bell curve distribution, with lower DNA content, standard deviations may be expected to be greater; thus, the thresholds for mutation detection must be correspondingly widened.

Quantity of DNA, usually measured by optical density (OD), is not the only factor. Quality of DNA is equally critical since degraded DNA, while present when measured by OD, may actually be totally incapable of being effectively amplified in the extreme cases of DNA degradation. An appreciation of DNA degradation is important to detecting mutational change, because samples typically are exposed to chemical fixatives and treatment during histologic preparation of tissue section of cytology slides that act to increase DNA degradation. Unlike samples in the laboratory which have the luxury of being immediately acted upon in the fresh state, clinical samples do not share this advantage and in fact must be assumed to have some degree of DNA degradation. Conventional teaching fails to recognize this reality and offers no useful recommendations in this regard. The recommendation to simply divide the lesional value by the normal value is valueless, because the “normal” value may be obtained under conditions of low amount and quality of DNA and so may itself show significant variability. This variability will then be introduced into all subsequent determinations of lesional values.

When the specimen is primarily a fluid that contains free DNA or a few cells, the DNA may be extracted in a conventional manner, and the quantity of extracted DNA measured by means of optical density (OD). Refer to commonly owned U.S. utility application entitled “Molecular Analysis of Cellular Fluid and Liquid Cytology Specimens for Clinical Diagnosis, Characterization, and Integration with Microscopic Pathology Evaluation,” filed Oct. 22, 2005 which claims priority to U.S. Provisional Application Nos. 60/620,926, 60/644,568, 60/631,240, and 60/679,968, all of which are herein incorporated by reference in their entirety for all purposes. Alternatively, quantitative nucleic acid amplification (qPCR) can be performed to measure the amount of extracted DNA.

When the specimen is a solid tissue specimen of small size such as a microdissected tissue sample, the DNA cannot simply be extracted. Nor can the amount of DNA be easily measured in small biopsy samples, because the usual methods for DNA extraction suffer from low yield when the specimen has been subject to chemical fixation. In such circumstances, it is better not to attempt to extract pure DNA and make it amplifiable DNA from a crude lysate of the specimen. To this end the method is Topographic Genotyping™ (TG) is strongly recommended (Topographic Genotyping™, U.S. Pat. No. 6,340,563) as well as the techniques discussed in the U.S. patent application entitled “Enhanced Amplifiability of Minute Fixative-Treated Tissue Samples, Minute Stained Cytology Samples, and Other Minute Sources of DNA” filed on Oct. 24, 2005, which are commonly owned and both of which are herein incorporated by reference in their entirety for all purposes. These methods are for amplifying minute quantities of DNA from minute clinical specimens typically encountered in routine clinical practice in the form of biopsies or fine needle aspirations. DNA amount is not measured in minute samples, because in doing so the specimen would be entirely consumed with none left for subsequent analysis.

Quantitation of DNA degradation of the sample can be performed in a novel and simple manner termed the competitive template polymerase chain reaction (CT-PCR). To perform this method, a gene/pseudogene pair is required that share close homology in genomic sequence but differ by the presence of a deletion of varying size in one of the sequences. This is usually encountered in gene/pseudogene pairing. However, it may be seen in other situations as well, such as between genes that are members of a large gene family. Such a situation can also be constructed in vitro by manipulating genomic sequences to possess deletions of predetermined size. Two such gene/pseudogene pairing involves the glucocerebrosidase gene (GenBank D13286) and its pseudogene (GenBank D13287), and the human carboxyl ester lipase (CEL) gene (GenBank M94579) and its pseudogene (GenBank M94580). Other similar examples can also be employed. In a preferred embodiment, the glucocerebrosidase gene and its pseudogene is sufficient, as it affords specific deletional regions of 19 bases and 55 bases in the exon 1 and 9 respectively.

The competitive template PCR (CT-PCR) reaction for the glucocerebrosidase gene/pseudogene pairing (exon 1 and 9) provide a novel and sensitive means to quantitate the degree of DNA degradation. The CT-PCR reaction, which is performed in triplicate (or more) at varying concentrations, provides quantitative information on the quality of DNA with respect to degradation, and quantitative information on the representative amplifiability of the DNA. The CT-PCR reaction provides a sensitive means to quantitatively characterize DNA degradation in a fashion that is essentially independent of the various phases of nucleic acid amplification (exponential phase, plateau phase) and is effective on minute specimen samples. By means of gene/pseudogene sequences that are nearly identical, but differ by varying lengths of base deletion, the effect of DNA degradation can be accurately characterized. By performing CT-PCR in triplicate (or more) at different concentrations of starting template, the effect of allelic dropout can be effectively controlled, and the degree of degradation can be accurately defined. Adequacy of amplifiable DNA is reflected by replicate reliable amplification of both the shorter and longer sized amplicons. Low amounts of inadequate levels of amplifiable DNA for mutational analysis is represented by variability in the effectiveness of relative amplification of the longer sized amplicon. The use of CT-PCR as described herein is a approach which provides the information needed to effectively analyze minute clinical tissue specimens subject to tissue fixation and staining.

A second method is used to quantify both the amount and degradation of DNA and is based on an analysis of the fluorescence distribution generated by nucleic acid amplification. The primers used for amplification possess fluorescent labels on their 5′ end as a means for detection and quantification by capillary electrophoresis. The amount of fluorescence produced for each peak height is related to the amplifiability of the DNA, which in turn is greatly influenced by the amount of starting DNA effective for amplification and its quality. The greater in amount that each amplicon is produced, the greater will be the fluorescence values that are generated for each polymorphic allele. Other factors influence total fluorescence in addition to template amount and quality and relate to the kinetics of the individual PCR reaction such as efficiency of primer hybridization and polymerization. These factors can be optimized to a degree by selecting primers for the amplification that are most robust for amplicon production. Having settled on a defined set of primers, fluorescence generation, a measure of genomic target amplifiability for the type of specimen being studied, is related to the amplifiability of the starting template. The combination of CT-PCR and fluorescence generation derived from normal samples will enable the practitioner to thoroughly understand critical factors such as DNA amount, quality, and amplifiability for any type of sample that is about to be studied. Of importance is that these features must be thoroughly worked out first on normal samples, and the information derived there from, before the data can be used in mathematical algorithms to define minimum threshold values for significant allelic imbalance.

To exemplify the use of these principles, the guidelines described in Table 11 may be applied first to normal samples and then to lesional specimens. This example is for a microdissected tissue specimen and so the amount of extracted DNA is not available as described above. It would be apparent that this can also be applied to a fluid cytology sample or other biological sample. Table 11 serves as a general template that must be refined for each marker and corresponding primer pair. Levels of CT-PCR and fluorescence that define states of reduced DNA content and increased DNA degradation can be derived for each marker. Markers that are sensitive to low amounts of DNA or degradation should not be used, because they are likely to give spurious values in clinical specimens that are less than optimal.

TABLE 11
FLUORESCENCE
CT-PCRUNITSCONCLUSION
0.50-1.00over 1000Sufficient good quality DNA is present in the
system to afford accurate, representative
amplification.
0.20-0.50Over 1000DNA amount is sufficient but it is mixed with
poor quality degraded DNA. This accounts
for the low CT-PCR values but high
fluorescence generation. Final values likely
not to be affected high DNA degradation.
0.20-0.50200-1000DNA amount is borderline, and there is a
significant admixture with degraded DNA.
Threshold values for significant allelic
imbalance must be widened to accommodate
greater variation in normal values (NAVERAGE).
0.00-0.20200-1000Small amount of good quality DNA is present,
but the sample DNA is highly degraded. Only
lesional samples that generate over 1000
fluorescence units can be trusted for
accuracy and at that the normal range of
values must be widened.
0.00-0.20Less than 200Totally inadequate quality and quantity of
DNA. Values cannot be interpreted.

A panel is then formulated of markers based on tetranucleotide (or greater) repeats targeting genomic regions associated with the location of common tumor suppressor genes but not strictly limited to any particular one. Markers are paired for each genomic region but spatially separated to provide information concerning the directionality of deletion expansion (genomic deletion expansion). Having defined a set of markers, normal samples are then carefully analyzed in sufficient number to enable a thorough understanding of their behavior under conditions of lower amounts of DNA quantity and under conditions when the DNA is degraded. Methods such as CT-PCT and evaluation of fluorescence generation are used to quantify marker function under these conditions. Using normal specimens, thresholds for significant allelic imbalance are defined using statistical methods recognizing the critical attributes of DNA quantity and quality. At this point the marker panel is ready to application to lesional samples.

The materials and methods described herein are those needed to perform the analysis that will used on the constructed panels to evaluate their effectiveness to address needs for their use. These methods can be performed on any tissue sample, ie resected tissue, drawn blood, cytologic specimen, or fluid specimen (e.g., fluid from a cyst). While there are overall common considerations for all these types of specimens, in fact the specific performance of the analysis will differ. These differences will be described in a subsection for each type of specimen. Preferably, these methods can be used on a combination of specimens such that the data from each specimen can be collated for further validation of the diagnosis and prognosis achieved by the genetic determination. Additionally, if using a microdissected sample, multiple microdissection representing different sections of tumor from the tissue block can be used as controls, and as validation of tumor aggressiveness.

The insight gathered through temporal sequence analysis of acquired mutational damage was used to create a molecular classification of gliomas (FIG. 15). This classification has the advantage of providing more accurate information concerning the pattern of treatment responsiveness (FIG. 15). The classification also reconciles many of the difficulties pathologists have had classifying gliomas based only on microscopic cellular appearance.

5.1 Specific Panels of Markers for Defined Indications

Table 12 lists primers that can be used to amplify tetranucleotide microsatellites with a high degree of robustness and accuracy. The stutter band formation with each is minimal and precise quantification is easily performed. These primers are used in combination with each other to produce highly efficient panels for molecular pathology evaluation of precancer, cancer and related conditions. Also included in the panel are specific non-microsatellite targets such as k-ras-2 exon 1 point mutation detection. These additional sets can be combined with tetranucleotide panels for further optimization. The list below is just a sample as other primer sets and tetranucleotide or longer targets can be added. All primers in Table 12 are listed in a 5′ to 3′ direction.

TABLE 12
Primer NumberPrimer NameSequence
FORWARDD10S1173FAM-TCA TGC CAA GAC TGA AAC TCC
REVERSED10S1173GCT GGC CAT GAC TGT TTT AC
REVERSED10S520GTC CTT GTG AGA AAC TGG ATG C
FORWARDD10S520HEX-CAG CCT ATG CAA CAG AAC AAG
FORWARDD17S1161HEX-AAC AGA GCA AGA CTG TCC AG
REVERSED17S1161GTC CCT CTA ACC TTT AGG AGA G
REVERSED17S1289FAM-CTG CCT CTA AGC AGT CAT TTA GA
FORWARDD17S1289GCA TGG TCT TCT TCC ATT CC
FORWARDD17S974NED-AGC CTG GGT GAG AGT GAG AC
REVERSED17S974GCC ATT GTT AAC AGG TTG GTG
REVERSED18S814TET-CCC ACT ATA TGT ATG TTC ACC
FORWARDD18S814CTC TCT GCC TCT CCC ACC
REVERSED19S400NED-CAG GGT TCT TAT TTC CTG TC
FORWARDD19S400GCC TCT ATA AAT AAA TAA AGA C
REVERSED19S559TET-TTG AGG TAT CTA TGT GGA TAT C
FORWARDD19S559GAG TGA GAC CCT GTC TTT AC
FORWARDD1S1193TAMRA-TCG GCG ACA TAG CCA GAG
REVERSED1S1193CTT TGA TCT AAG GAT TAC CTA C
REVERSED1S407TGG GCG GGG GAT AGA AGG
FORWARDD1S407FAM-TGC TAA CCA CAT GGA GAG G
FORWARDD21S1244HEX-TCT TCT ATC TCA TAT GTG TAT C
REVERSED21S1244GGA GGA ACT TGA GGA TGT G
FORWARDD22S532FAM-CCT GGG CAA CAG AGC GAG
REVERSED22S532GTC TGA GAA GAT ACT TGA TAT AG
FORWARDD3S1539NED-CTC TTT CCA TTA CTC TCT CC
REVERSED3S1539GTT CTC CAT CTA TCT TTC TCT C
REVERSED3S2303TET-CTC GAG AGC TTT GTT TTC AAC
FORWARDD3S2303GTG CCT ACA TGT TAG TAT CC
REVERSED5S592TAMRA-TGG ATA CAT ATT GTT TTC TGC TG
FORWARDD5S592GGT GTC AAC AAA GTA ATG TAA AG
REVERSED5S615FAM-TCC ACA GTG GTA AGA ACC AG
FORWARDD5S615GAG ATA GGT AGG TAG GTA GG
REVERSED7S1530FAM-ACA CAG TCT TCA GCC CTA C
FORWARDD7S1530ACA GAG CTA AAC TCT GTC TC
REVERSED8S373HEX-TGT ATG CAT TAT TTC CTT CAA TC
FORWARDD8S373GTA GGC AGG TGG GTG GGC
REVERSED9S251HEX-CAA TAC TTT TTA AGG CTT TGT AGG
FORWARDD9S251GTT TTA TGT GCA CTA ACT AAT GG
FORWARDD9S252NED-TTG TCA ACT CCT AAT ATG GAC
REVERSED9S252GAT ATC CCC AAG TTC TCA TAC
REVERSED9S254GAG GAT AAA CCT GCT TCA CTC AA
FORWARDD9S254FAM-TGG GTA ATA ACT GCC GGA GA
FORWARDL-MYCFAM-TGA ACC GTA GCC TGG CGA G
REVERSEL-MYCGCT GTT CTT CCT TTT AAG CTG
FORWARDD1S1172GAA CAG AAC CTG GTA CCT TC
REVERSED1S1172FAM-CTG CAC CCG GCT GAT GTT C
REVERSEKRAS 2 Exon1TCC TGC ACC AGT AAT ATG CA
FORWARDKRAS 2 Exon1TAA GGC CTG CTG AAA ATG ACT GA
FORWARDD5S346HEX-ACT CAC TCT AGT GAT AAA TCG GG
REVERSED5S346AGC AGA TAA GAC AGT ATT ACT AGT T
FORWARDD5S1170FAM-AAA TTA ACT GTA ACC CTA GTG TG
REVERSED5S1170TCT ACC CTG AAG TAG CTC CAA AC
FORWARDD5S2027FAM-ACT TGG CAG ATT TTC CAC TC
REVERSED5S2027CAC CTC ATT GAC TGG GAC
FORWARDD5S1965FAM-TGT CCC GTT GAT AAA AAT TAC TGC G
REVERSED5S1965GTG TCT GGG ATT TCC TAC GCA ATG
FORWARDD9S916HEX-ATC GAT CAG GTG CCA GAG
REVERSED9S916AAA GGG AGA ATA TAT TAC TTG TGC AG
FORWARDD9S736HEX-TTC TAG ACC TCT CAG CAG AC
REVERSED9S736GAT AGT GTT GGA GAC ACC AG
FORWARDD9S1814FAM-CAT GGT TCT TCT ACT CAG G
REVERSED9S1814HEX-TGG GCC TGT GA ACCT ACT GAC
FORWARDD9S966GAT TGT ACC AAC AGC ACT GAG
REVERSED9S966FAM-CAG GCT TGA TCA GCT TCC TGG
FORWARDD10S541HEX-AGA ACT GGA ATA GAA GGA AC
REVERSED105541AGG GAT GGG TAC AGA AAC TC
FORWARDD10S2339FAM-CTC TAG GCT CAT TTT CCT CTT CC
REVERSED10S2339CAC AAA TGC AGA TAA AAT AAA GAC A
FORWARDD105579HEX-ACT AGG AAG GTT CAT ATT CC
REVERSED10S579TCA TCC CCT GAA ATA CAC C
FORWARDD10S2394HEX-CCA ACG AGA TGA TGT TGC C
REVERSED10S2394AAG AGC TGG GTT GGG GTA CT
FORWARDD17S655FAM-TGA ACC CGG GAG GCA GAG C
REVERSED17S655GTG TGT GAG GCT GGT AGA ACA TG
FORWARDD17S1796GCT GAT GTC AGA CTG AGC
REVERSED17S1796FAM-CAC AGA TTA ATA GAG ATA TTT TCC C
FORWARDD17S2159HEX-AGA TTA CCC CAT GTC TTC ATC C
REVERSED17S2159TCT ATT GTT TGT GTT AGC CTT ACC C
FORWARDD17S1783HEX-ATG TGT ACA GAG CCG CAA AG
REVERSED17S1783CAT GAT TCA GCC AGG GAC TC

Panels may be constructed as follows for specific applications:

Glioma molecular analysis: D1S1193, D1S407, D1S1172, LMYC, D9S254, D9S251, D10S1173, D10S520, D17S974, D17S1289, D17S1161, D19S400, and/or D19S559. This panel is based upon the use of four tetranucleotide/pentanucleotide microsatellites on 1p to differentiate interstitial from telomeric 1p deletion. The markers for 9p, 10q and 17p are in proximity to the DKN2A, PTEN, and TP53 genes which are very commonly mutated in gliomas. 19q markers are included as they are lost in close correlation with loss of 1p markers. The 17q marker is included, because it has been found to be frequently imbalanced in gliomas.

Discrimination between de novo cancer formation versus recurrence/metastasis of cancer: D1 S1193, D1 S407, D3S2303, D3S1539, D5S592, D5S615, D9S254, D9S251, D10S1173, D10S520, D17S974, D17S1289, D17S1161, D21S1244, and/or D22S532+/−K-RAS-2 point mutation when lung, pancreatic and colon cancer are being evaluated. These markers correspond to 1p36, 3p25, 5q23, 9p21, 10q23, 17p13, 17q21, 21q22, and 22q12 which are situated in proximity to important tumor suppressor genes.

Discrimination between reactive versus neoplastic proliferation: D1 S1193, D1S407, D3S2303, D3S1539, D5S592, D5S615, D9S254, D9S251, D10S1173, D10S520, D17S974, D17S1289, D17S1161, D21S1244, and/or D22S532+/−K-RAS-2 point mutation when lung, pancreatic and colon cancer are being evaluated. These markers correspond to 1p36, 3p25, 5q23, 9p21, 10q23, 17p13, 17q21, 21q22, and 22q12, which are situated in proximity to important tumor suppressor genes.

Molecular analysis of aspirated pancreatic fluid: D1S1193, D1S407, D3S2303, D3S1539, D5S592, D5S615, D9S254, D9S251, D10S1173, D10S520, D17S974, D17S1289, D17S1161, D18S814, D21S1244, D22S532 and/or K-RAS-2. These markers correspond to 1p36, 3p25, 5q23, 9p21, 10q23, 17p13, 17q21, 18q22, 21q22, and 22q12, which are situated in proximity to important tumor suppressor genes associated with pancreatic adenocarcinoma formation.

Discrimination between low versus high grade glioma: D1S1193, D1S407, D1S1172, LMYC, D3S2303, D3S1539, D5S592, D5S615, D9S254, D9S251, D10S1173, D10S520, D17S974, D17S1289, D17S1161, D19S400, and/or D19S559. This panel is based upon the use of four tetranucleotide/pentanucleotide microsatellites on 1p to differentiate interstitial from telomeric 1p deletion. The markers for 9p, 10q and 17p are in proximity to the DKN2A, PTEN and TP53 genes which are very commonly mutated in gliomas. 19q markers are included as they are lost in close correlation with loss of 1p markers. 3p, 5q, and 17q marker is included because it has been found to be frequently imbalanced in gliomas.

Trichlorethylene (TCE) associated mutational damage: D1S1193, D1S407, D3S2303, D3S2327, D3S1745, D3S1540, D3S1542, D3S1539, D5S592, D5S615, D9S254, D9S251, D10S1173, D10S520, D17S974, D17S1289, D17S1161, D21S1244, and/or D22S532+/−K-RAS-2 point mutation when lung, pancreatic and colon cancer are being evaluated. The emphasis on 3p markers reflects the sensitivity of the region for mutational change. The remaining markers are used as a general survey for acquired mutational damage.

Discrimination between benign versus malignant: Discrimination between reactive versus neoplastic proliferation: D1S1193, D1S407, D3S2303, D3S1539, D5S592, D5S615, D9S254, D9S251, D10S1173, D10S520, D17S974, D17S1289, D17S1161, D21S1244, and/or D22S532+/−K-RAS-2 point mutation when lung, pancreatic and colon cancer are being evaluated. These markers correspond to 1p36, 3p25, 5q23, 9p21, 10q23, 17p13, 17q21, 21q22, and 22q12, which are situated in proximity to important tumor suppressor genes.

Discrimination between reactive gliosis versus glioma: Discrimination between low versus high grade glioma: D1S1193, D1S407, D1S1172, LMYC, D3S2303, D3S1539, D5S592, D5S615, D9S254, D9S251, D10S1173, D10S520, D17S974, D17S1289, D17S1161, D19S400, and/or D19S559. This panel is based upon the use of four tetranucleotide/pentanucleotide microsatellites on 1p to differentiate interstitial from telomeric 1p deletion. The markers for 9p, 10q, and 17p are in proximity to the DKN2A, PTEN, and TP53 genes which are very commonly mutated in gliomas. 19q markers are included, as they are lost in close correlation with loss of 1p markers. Markers for 3p, 5q, and 17q are included, because it has been found to be frequently imbalanced in gliomas. In this case, a patient with brain cancer (with a glioma) is being distinguished from an individual with reactive gliosis (caused by multiple sclerosis, an infarction, an infection, and the like).

It will be apparent that as additional markers for each category listed above are delineated they can be added to the list in any form or combination. Optimally, all the polymorphic markers are utilized for each category.

It will be apparent that as additional markers for each category listed above are delineated they can be added to the list in any form or combination. Optimally, all the polymorphic markers are utilized for each category.

EXAMPLES

Example 1

Clinical History. An elderly patient developed anemia and hematuria, and was found to have a sigmoid colon cancer. The colon cancer was resected and mutational analysis was performed at three microscopic sites: center, 1 cm away from the center and 2 cm away from the center at the periphery of the tumor.

Materials and Methods. The tumor sections of each cancer and the metastatic lymph node deposit were reviewed by microscopy and the most representative microscopic sites of cancer were identified (FIG. 6). Multiple sites were selected for microdissection and mutational analysis.

For microdissected samples from tissue sections or cytology slides, the following procedure was used. A lysis buffer (Buffer B) was made by adding 2 mL 5M NaCl (Sodium Chloride, Fisher Scientific Cat. # AC32730-0010), 5 mL 0.5M EDTA pH 8.6 (ethylenediaminetetraacetic acid disodium salt, Fisher Scientific Cat. # BP120-500 with sodium hydroxide, Fisher Cat. # BP359-500 to pH the EDTA to 8.6), 1 mL 1M Tris (Tris Ultra Pure ICN, Fisher Scientific Cat. # 819638), and 500 μL 0.5% NP-40 Surfact-Amps 20 solution in water (Surfact-Amps 20, Pierce Cat. # 28320) into a sterile bottle. Alternative nonionic detergents can be substituted such as and aqueous 10% Tween 20 solution. PCR Grade Distilled Water (PCR Grade Distilled Water DNAse, RNAse Free 1 micron filtered, Fisher Scientific Cat. # BP2470), was added to bring the volume up to 100 mL (approximately 91.5 mL). This buffer was stored in the refrigerator, but it is also stable and can remain at room temperature for a least 1 year or longer. Storing in the refrigerator reduces the tendency for bubbles to form and increases accuracy in pipeting of the buffer.

Then 25 μL of lysis buffer, Buffer B, was pipeted into 0.5 mL-0.65 mL PCR polypropylene tubes which were RNase- and DNase-free, as well as DNA-free. The tubes were then tightly capped and then stored at room temperature for at least 1 year or longer. These tubes were used to process and store DNA microdissected off tissue on glass slides (histology or cytology) or cells from cytology brushes, buccal brushes or cells from cell lines. The concentrations of these reagents may be varied both upwards and downwards with similar or reduced effectiveness. Other nonionic detergents besides NP-40 or Tween may be substituted for breaking up fixative-treated tissue in the crude lysate.

Using six serial four- to five-micron thick deparaffinized, unstained, re-cut histologic sections, the precise areas of cancer were removed from the slides. Tissue sections may be sectioned thicker or thinner with equivalent effect; however 4-5 microns is recommended as it is the standard thickness for sectioning fixative treated tissue. The cap of the tube containing 25 mL of lysis buffer (Buffer B), into which the microdissected tissue is deposited for processing and storage, should be snapped open. A pointed tipped scalpel (No. 11 Feather scalpel) was used for microdissection. The tip of the scalpel was dipped into the lysis buffer (Buffer B) to wet the tip. The wetted tip was used to remove the tissue of interest under stereomicroscopic observation (Olympus SZ-PT stereomicroscope). The reverse (unsharpened) side of the pointed tip was the most advantageous orientation for microdissected tissue or cytology slides. Other aspects of the scalpel can be used with varying effectiveness. The tissue, as it is scrapped off the slide, will adhere to the wetted tip. The wetted tip and tissue thereon were lifted off the glass slide and the tip dipped gently back into the tube with lysis buffer (Buffer B), where the tissue was released from the tip and put into the buffer solution. The tip was wet when removed. With the wetted tip more tissue was collected and placed in the tube, and the blade was carefully and thoroughly wiped and set aside for the next target. Tissue may be added until the tissue begins to appear to be opaque when viewed through the tube. The tube then was capped tightly.

The volume of lysis buffer (Buffer B) was adjusted based on the visible estimation of the amount of tissue in the tube. If no cellular material was present, no additional buffer was added. If a small amount was seen at the bottom of the tube, another 25 μL of buffer was added. When the amount of cellular material was seen to fill approximately half the volume of fluid in the tube, 50 μL of buffer was added. When the amount of cellular material appears to almost fill or fill the total volume in the tube, 75 μL of buffer was added. The range of this volume adjustment to match the amount of lysis buffer to the amount of microdissected tissue can vary with equal effectiveness, as long as the microdissected tissue is not excessively diluted or heavily concentrated. Proteinase K, up to a concentration of 2 mg/mL, was added and the lysate incubated at 37° C. for two hours. The duration of digestion may be greater than 2 hours; convenient overnight digestion may work effectively. Proteinase digestion was then stopped by boiling the sample for 5 minutes. Once brought down to room temperature, the processed sample tubes were centrifuged at 5,000-10,000 rpms for 5-10 minutes. The duration of centrifugation may vary considerably beyond five minutes; however greatly excessive pelleting is not useful. At this point the sample had been processed into a crude lysate which was ready to proceed to PCR.

For analysis of free DNA in fluid specimens, the following procedure was used. 200 μL of fluid was treated and passed through a Qiagen spin column (QIAamp DNA Mini Kit, Valencia, Calif.) according to manufacturer's instructions designed to extract DNA. The DNA was resuspended in 50 μL of 10 mM Tris-EDTA buffer, pH 7.0. One microliter of the resuspended DNA was measured for DNA content and purity by optical density (NanoDrop Technologies, Wilmington, Del.). This step was performed twice consuming 2 μL. The average of the two values for DNA content was used as the value for the sample. The resuspended DNA was then diluted to a value of 5 ng/μL if the content of DNA was above that level. All subsequent analysis was performed on this normalized DNA content sample.

A qPCR reaction for the first exon of the k-ras-2 gene was performed in duplicate to assess the amplifiability of the DNA (see PCR procedure described below). Two microliters of DNA were used in each of these duplicate reactions (Icycler, BioRad, Hercules, Calif.). The Ct values for the duplicate reactions, representing the quantity of amplifiable starting DNA, were averaged and used as a measure of DNA amplifiability. As an independent measure of DNA quality, duplicate 1 μL aliquots of the normalized DNA were used in a PCR reaction to amplify short (99 base pair, pseudogene) and long (154 base pair, true gene) segments of the glucocerebrosidase geneipseudogene using the same pair of amplification primers. The ratio of the peak heights for 99 and 154 base pair products was obtained using the shorter product for the denominator. Optimal amplification of high quality DNA was expected to yield a value of 1.0. Values below this value are proportional to the degree of DNA degradation. Other genomic targets may be substituted for the determination of DNA degradation based on the principle that the longer product will be present in lesser abundance in relation to the degree of degradation.

PCR mix was prepared using reagents from the AmpliTaq Gold DNA Polymerase with Gold Buffer kit (Applied Biosystems, Foster City, Calif.) with some modifications. A 12.5 μL volume PCR was performed in 96-well PCR plates (Fisher Plate, 96-Well, semi-skirted Cat. # 12 566 134). Following manufacturer's kit instructions, 1.25 μL Gold Buffer and 1.25 μL magnesium chloride were added.

0.25 μL GeneAmp dNTP Mix with dTTP (Applied Biosystems, Foster City, Calif.) of dATP, dGTP, dCTP, and dTTP from other manufacturers was then added. Equivalent dNTPs can be substituted. In addition, 5.75 μL of PCR Grade Distilled Water (PCR Grade Distilled Water DNAse, RNAse Free 1 micron filtered, Fisher Scientific Cat. # BP2470) was added (equivalents can be substituted). Modifications included adding additional 1.25 μL 2.5 mM MgCl2 and 2.5 μL of a 60% sucrose solution (Sucrose, Fisher Cat. # AC41976-0010) in PCR Grade Distilled Water (PCR Grade Distilled Water DNAse, RNAse Free 1 micron filtered, Fisher Scientific Cat. # BP2470). 0.125 μL of the upstream and downstream primers were added to complete the mix. To this mix, 1 μL of the crude lysate was added. Then the plate was sealed with Amplification tape (Fisher Scientific Cat. # 12 565 491) and placed on a preheated thermal cycler for polymerase chain reaction to occur. Thermal cyclers in current use include Thermo Electron Multi-Block Cyclers and Bio-Rad's iCycler. Cycling parameters ranged from annealing temperatures between 50 to 60° C. and cycling between 30 to 40 cycles. Incubations range from 15 seconds to 15 minutes. One microliter of crude lysate was used in each singleplex reaction with primers designed to amplify a specific genomic segment within or outside of a genes or genomic regions of interest. Oligonucleotide primers were added as desired. A stock mixture buffer was then used to add Taq polymerase, deoxyribonucleotides, and salts to the final amplification mix.

The following modifications were important to enhance amplifiability. The magnesium concentration was significantly increased to a level of 8 mM in order to improve hybridization of the probe to the template DNA. Also important was the addition of sucrose to a final concentration 12 g/100 mL (12 gram percent). The addition of sucrose at this level may assist in the relaxation of double strands in turn encouraging primer annealing. Other substitutions using similar reagents may be used with equivalent effects expected. Other equivalent sugars may be substituted for sucrose. Further, other ions of equivalent valency, such as manganese, may be substituted for magnesium. Also, the concentration of the reagents provided herein can be adjusted with correlative beneficial effects.

The 1 μL of amplified DNA was mixed with 10-18 μL of deionized Formamide and 1 μL of a 40%-50% Rox Calibrator Solution. Samples were loaded onto the capillary electrophoresis (ABI 3100 Genetic Analyzer, Applied Biosystems, Foster City, Calif.) and electrophoresed through POP4 polymer. POP6 polymer from ABI is suitable, as well as any polymer which can discriminate the size of base pairs. Samples were run through a 36 cm capillary array, although the 50 cm capillary array could be substituted. GeneScan software may be used, with the settings Dye set D and GS400HD Analysis Module. Injection times varied between 22-27 seconds and run times varied between 1020 and 1500 seconds. Other software that allows the user to identify and measure peak height may be substituted, such as 3730xl DNA Analyzer software. Other capillary equipment from, for example, Beckman and Agilent may be used as well.

Amplification primers were designed with fluorescent labels attached to their 5′ end so that they may be detected and quantified. Fluorescent labels include 5-HEX, TAMRA, 5,6-FAM (Integrated DNA Technologies, Inc.; Coralville, Iowa) or (Synthegen, LLC; Houston, Tex.), and NED (ABI 3100 Genetic Analyzer, Applied Biosystems, Foster City, Calif.). As the fluorescently labeled DNA flows through the capillary electrophoresis, it is sorted by size and as it passes a window, the laser fires to cause the labels to fluoresce. The fluorescent is captured by a camera as it passes the window. The time in which the fluorescent is captured determined the length of the patient's alleles, and the amount determines the peak height which indicates the amount of PCR product that was amplified in the PCR reaction. The ratio of polymorphic peak heights or specific peak heights representing defined amplification products was then used to detect and quantify mutational change that may be present. The software allows the user to chose peaks, and once a peak is chosen, it generates a row of data including the allele lengths and peak heights.

The data was copied into an Microsoft Excel® table. The Excel®table was then imported into a database, which merges the capillary electrophoresis data with the sample information, so that results may be searched by patient's name or other identifiers, as well as used to generate a report of molecular results linked to each patient's target.

The first exon of the k-ras-2 oncogene was PCR amplified. The PCR product underwent a PCR clean-up step using Microcon YM50 (Millipore Corp., Bedford, Mass.) and following manufacturer instructions to remove amplification primers. Cycle sequencing (BigDye Terminator Cycle Sequencing Kit V1.1, Applied Biosystems, Foster City, Calif.) by the dideoxy technique was performed following the kit's instructions. Any method which allows for the separation and discarding of the primers from the PCR product may be substituted. DyeEx 2.0 Spin Kit (Qiagen, Hilden, Germany) was used for post-cycle sequencing, to remove excess unincorporated fluorescent dye and primers following manufacturer's instructions. There are many kits on the market that can be substituted.

A 96-well standard PCR plate was used to load samples on the capillary electrophoresis equipment. Each well held 15 μL of distilled or PCR Grade Distilled Water (PCR Grade Distilled Water DNAse, RNAse Free 1 micron filtered, Fisher Scientific Cat. # BP2470) and 5 μL of the post cycle sequencing cleaned up sample. Samples were then entered into the Sequencing software program and electrophoresed by capillary electrophoresis on the 3100 Genetic Analyzer (ABI 3100 Genetic Analyzer, Applied Biosystems, Foster City, Calif. on the following settings: Dye set E, mobility file DT3100POP4LR(BD)V1.mod, using the Run Module UltraSeq36_POP4DefaultModule and analyzed with BC-3100POP4UR_SeqOfftOff.saz module.

Fluorescent labeling was used and the sequence was read by fluorescent capillary electrophoresis. The iQ SYBR Green Supermix (Bio-Rad Cat #170-8880, Hercules, Calif.) was used to make a 25 μL volume reaction mix. Following manufacturer's instructions, 12.5 μL of 2×SYBR Green Supermix was added, as well as 0.25 μL of the upstream primer and 0.25 μL of the downstream primer. The primers were at a concentration of 0.5 to 1.0 μM. In addition, 2.5 μL of a 60% sucrose in PCR grade water (i.e., the same water used for the other PCR reactions mixes) was added. The concentration of magnesium chloride was increased by adding 2 μL of 25 mM magnesium chloride (Applied Biosystems, Foster City, Calif.). 5.75 μL PCR grade water was added, to bring to the volume up to 23 μL. 2 μL of the DNA lysate (prepared from cytology or histology microdissected tissue or cells or buccal brush or other solid tissue as appropriate) or 2 μL of DNA extracted (from fluid specimens) was added to the mix, using the Qiagen or equivalent kit or other DNA method of extracting DNA from cells. The iQ SYBR Green Supermix contains fluorescein which was used by a real-time PCR thermal cycler “iCycler” as a “well factor” background check. If another real-time PCR machine is used, it would be necessary to use another kit containing the fluorescence used by that company. The mix may be prepared using the PCR mix described above with the addition of background fluorescence and the SYBR Green. There are also many detection methods known in the art, other than SYBR green, such as TaqMan Probes. However, 2.5 μL of a 60% sucrose in PCR grade water should be added and the concentration of magnesium chloride doubled, regardless of the detection method used.

The thermal cycling used is identical to that of PCR, except it is performed on a real-time thermal cycler such as the iCycler (Bio-Rad Cat. #170-8880, Hercules, Calif.). This real-time thermal cycler allows the immediate viewing of the increase in fluorescence, indicating an increase in PCR product. After PCR is complete, the real-time thermal cycler allows for the analysis of the amount of fluorescence detected (here, the detection of SYBR green). The software sets a threshold above background which indicates when the PCR reaction is in the exponential rate of growth phase. When each sample's fluorescence passes that threshold, the software assigns a “Ct” or Cycle Threshold value. The Ct is calculated as the point in the PCR cycle when the curve passes the threshold. The lower the Ct value, the more effective and more PCR product is being made. The higher the CT value, then the less effective and less PCR product is made. If the sample does not pass the threshold, it is read as “NA” or “not applicable”, indicating that very little to no PCR product was made. The Ct infers how much and/or quality of the DNA was present at the starting point. DNA of poor quality (broken into small sections, cross-linked to proteins etc) or DNA present in very small amounts has a high to “NA” Ct value. DNA of good quality and/or present in larger amounts has lower Ct values. These results are printed out and used in the determination of the patient's diagnosis.

Genetic sequence information was obtained from the Genome Database (Internet:<URL:http://www.gdb.org>). Previously amplified DNA was labeled with radioactivity (35S, 33P, 32P) and electrophoresed through 6-8% polyacrylamide bis-acrylamide gel.

The ratio of peaks was calculated by dividing the value for the shorter sized allele by that of the longer sized allele. Thresholds for significant allelic imbalance were determined beforehand in extensive studies using normal (non-neoplastic) specimens representing each unique pairing of individual alleles for every marker used in the panel. Peak height ratios falling outside of two standard deviations beyond the mean for each polymorphic allele pairing were assessed as showing significant allelic imbalance. In each case, non-neoplastic tissue served as a source of DNA to establish informativeness status and then to determine the exact pattern of polymorphic marker alleles. Having established significant allelic imbalance, it was then possible to calculate the proportion of cellular DNA that was subject to hemizygous loss.

For example, a polymorphic marker pairing whose peak height ratio was ideally 1.00 in normal tissue with a 95% confidence intervals from 0.78 to 1.62, could be inferred to have 50% of its cellular content affected by hemizygous loss if the peak height ratio was 0.5 or 2.0. This requires that a minimum of 50% of the DNA in a given sample be derived from cells possessing deletion of the specific microsatellite marker. The deviation from ideal normal ratio of 1.0 indicated which specific allele was affected. In a similar fashion, allele ratios below 0.5 or above 2.0 could be mathematically correlated with the proportion of cells affected by genomic loss.

Table 13, below, shows the various genomic targets that are being interrogated by oligonucleotide primer sets. Of importance is the delineation of the specific thermocycler profiles that are to be used to most effectively amplify these individual genomic targets using the primers detailed in the accompanying table. The denaturing temperature in all cases was 95° C. The polymerization temperature in all cases was 72° C. The annealing temperature is shown by the first number and the total number of cycles needed is indicated by the second number. Thus D17S974 is amplified with an annealing temperature of 55° C. for 35 cycles. This provides the exact conditions by which each PCR is to be performed.

TABLE 13
The first number is the profile, the second number is the
annealing temperature and the third is the # of cycles
Panels/
TestsICYCLER
Sep. 14, 20043 M 55 301 55 402 55 355 53 356 53 301 55 40
MajorD1s.407D9s.254D17S.1161D9s.251***D3s.2303*
Analysis
D10s.1173D5s.592***D1s.1193***D5s.615**D22s532*
GAUCHERD10s.520***D17s.974***D3s.1539*
E9 FAM
D21S.1244***D17s.1289***
FluidD1s.407*D9s.254*D17S.1161*D9s.251D3s.2303*Kras2 E1
Analysis
PancreaticD10s.1173*D5s.592***D1s.1193***D5s.615**D22s532*
Fluid
GAUCHERD10s.520***D17s.974***D3s.1539*
E9 FAM*
D21S.1244***D17s.1289***
GliomaD1s.407*D9s.254*D17S.1161*D19s.400*D1s1172*
D10s.1173*D5s.592***D1s.1193***D9s.251***
GAUCHERD10s.520***D17s.974***D19s.559***
E9 FAM*
Lmyc.5NT***D17s.1289***D5s.615*
ID QAD1s.407*AMELOGENIND17s.1289***D19s.400*D3s.1539*
Routine
D10s.1173*D5s.592***
Lmyc.5NT***
MSID17s250***D2s123*
1st 5D5s346*
markers
BAT-25*
BAT-26***

Cells in Table 13 are marked to reflect the type of fluorescent marker that has been attached to the primers for detection by capillary electrophoresis.

*for FAM label;

**for NED and TAMRA labels;

***for HEX label; and no label-for direct sequencing.

As discussed above, “3 M 55 30” would be profile 3, the anneal temperature of 55° C., for 30 cycles.

Cells in Table 13 are marked to reflect the type of fluorescent marker that has been attached to the primers for detection by capillary electrophoresis. “*” for FAM label; “**” for NED and TAMRA labels; “***” for HEX label; and no label-for direct sequencing. As discussed above, “3 M 55 30” would be profile 3, the anneal temperature of 55° C., for 30 cycles.

Results. Two sets of markers were used for each of the genes studied for the patient sample. For the APC gene located at 5q22.1, the first set of markers used was D5S2027 and D5S1965, and the second set of markers used was D5S346 and D5S1170. See Table 17 for the primers utilized. The proximal and distal markers respectively flank the APC gene and assist in determining the extent of the gene's deletion. Additional markers may be utilized to further validate and determine the deletional extent for the APC gene. Table 14 shows the two sets of proximal and distal markers used to assess the gene deletions for the particular patient same for APC, CDKN2A, PTEN, and TP53.

Colorectal cancer exhibits prominent genomic deletional expansion affecting the APC gene region located at 5q23. Tumor cells show a dynamic expansion in going from the center of the tumor to the periphery, which supports determination that this neoplasm was biologically aggressive. The results for the CDKN2A gene indicated that there was a dynamic genomic deletional expansion of the gene, but the expansion was not centered on this gene. Rather, the geneomic deletion was seen to expand into the territory of the gene. From the CDKN2A data it was concluded that genomic deletion does in fact exist. The analysis also indicated that an alternative target gene that is centromeric to the CDKN2A gene may be part of the developing aggressiveness of the cancer. This technology, allows the diagnosing physician to more precisely define the most likely target gene or genomic area representing the center for expanding deletional damage. Moreover, by performing the analysis at multiple sites, the affected tissue sites for the origin of cellular clones with specific deletions can be determined with greater precision.

The results for the PTEN gene were also of interest. There was clearly dynamic genomic deletional expansion that was centered on this gene, but the cellular origin for the neoplastic clone of affected cells was not at the tumor center. Rather, it was located in the midway region of the tumor. Clonal expansion from that topographic location is then seen to result in genomic deletional expansion directed towards the center and periphery of the tumor. This is a surprising result because the molecular analysis was designed with the assumption that the PTEN gene would be the affected gene and that deletional expansion would emanate for its genomic location. The analysis however indicated that an alternative gene was the one affected by deletion and that deletional expansion emanated for the location of that alternative gene target.

The results for TP53 analysis showed no imbalance, and thus no evidence for genomic deletion expansion. The molecular features of this patient's are those of an aggressive form of colorectal cancer capable of generating multiple independent and overlapping clones of tumor cells. Three of four (75%) target genomic regions showed mutational damage. This level of mutation acquisition is highly predictive (over 95% confidence) of aggressive colon cancer biological behavior.

TABLE 14
TUMORTUMOR
TUMORMIDWAYPERIPHERY
Mb distanceCENTER1 cm2 cm
DISTAL111.10 99%100% 100% 
MICROSATELLITED5S2027
DISTANCE pter
PROXIMAL111.85 98%99%100% 
MICROSATELLITED5S1965
DISTANCE pter
TUMOR112.10 N/AN/AN/A
SUPPRESSOR5q22.1
GENE (APC)APC
DISTANCE pter
PROXIMAL112.20 63%77%91%
MICROSATELLITED5S346
DISTANCE pter
DISTAL112.40 NO49%66%
MICROSATELLITED5S1170IMBALANCE
DISTANCE pter
DISTAL21.60NO60%78%
MICROSATELLITED9S736IMBALANCE
DISTANCE pter
PROXIMAL21.80NONO53%
MICROSATELLITED9S916IMBALANCEIMBALANCE
DISTANCE pter
TUMOR21.96N/AN/AN/A
SUPPRESSOR9p21
GENE (CDKN2A)CDKN2A
DISTANCE pter
PROXIMAL22.10NONONO
MICROSATELLITED9S1814IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL22.20NONONO
MICROSATELLITED9S966IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL89.40100% 100% 100% 
MICROSATELLITED10S579
DISTANCE pter
PROXIMAL89.5084%73%100% 
MICROSATELLITED10S2394
DISTANCE pter
TUMOR89.60N/AN/AN/A
SUPPRESSOR10q23.31
GENE (PTEN)PTEN
DISTANCE pter
PROXIMAL89.9063%49%79%
MICROSATELLITED10S541
DISTANCE pter
DISTAL90.0045%NO57%
MICROSATELLITED10S2339IMBALANCE
DISTANCE pter
DISTAL7.10NONONO
MICROSATELLITED17S2159IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL7.25NONONO
MICROSATELLITED17S1783IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
TUMOR7.50N/AN/AN/A
SUPPRESSOR17p13.1
GENE (TP53)TP53
DISTANCE pter
PROXIMAL7.60NONONO
MICROSATELLITED17S655IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL7.70NONONO
MICROSATELLITED17S1796IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter

Example 2

Clinical History. An elderly patient developed anemia and hematuria, and was found to have a sigmoid colon cancer. The colon cancer was resected and mutational analysis was performed at three microscopic sites: center, 1 cm away from the center and 2 cm away from the center at the periphery of the tumor.

The materials and methods used to obtain the molecular diagnosis are as set forth in Example 1, supra.

Results. As in Example 1, four genes (i.e., APC, CDKN2A, PTEN, and TP53) were analyzed in the patient sample for the extent of deletion using two sets for each gene of distal and proximal markers. Additional markers can be used to further validate and describe the extent of the genetic deletion. Additionally, different genes and/or more genes can be analyzed using this method to further describe the aggressive character of the disease. Additionally, this data can be coupled with data from different microdissected portions (e.g., normal cells, cells located on the margin of the putative tumor). It would be evident that the genes analyzed would vary depending on the type of cancer and the involvement of a particular gene in that cancer (see Table 1), as well was the combination of number of genes analyzed.

Genomic deletional expansion analysis can be applied to a single genomic target are any number of separate genes or genomic markers. The larger the number of markers analyzed, the greater detail will be generated for assessment of tumor biological aggressiveness.

With this patient, there was evidence of a dynamically expanding genomic deletion for the APC gene centered in the periphery of the tumor. A similar alteration was identified in the center of the tumor for the CDKN2A gene. No allelic imbalance was detected for PTEN or TP53 in any of the tumor samples taken from the patient. The data is presented in Table 15 as sets of marker analysis by gene.

However, in contrast to Example 1, the same here was determined to have a lower rate of deletion expansion compared to the sample from the patient in Example 1. Thus the patient with the tumor here was shown to have a tumor that is less biological aggressive than the tumor of the patient in Example 1 based on the fact that there is less acquired detectable mutational change. In this example, the tumor was shown to have fewer allelic imbalances and those imbalances were expanding at a slower rate.

TABLE 15
TUMORTUMORTUMOR
Mb distanceCENTERMIDWAYPERIPHERY
DISTAL111.10 NONO69%
MICROSATELLITED5S2027IMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL111.85 NONO73%
MICROSATELLITED5S1965IMBALANCEIMBALANCE
DISTANCE pter
TUMOR112.10 N/AN/AN/A
SUPPRESSOR5q22.1
GENE (APC)APC
DISTANCE pter
PROXIMAL112.20 NONONO
MICROSATELLITED5S346IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL112.40 NONONO
MICROSATELLITED5S1170IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL21.6048%52%56%
MICROSATELLITED9S736
DISTANCE pter
PROXIMAL21.8085%92%99%
MICROSATELLITED9S916
DISTANCE pter
TUMOR21.96N/AN/AN/A
SUPPRESSOR9p21
GENE (CDKN2A)CDKN2A
DISTANCE pter
PROXIMAL22.1053%57%62%
MICROSATELLITED9S1814
DISTANCE pter
DISTAL22.20NO48%55%
MICROSATELLITED9S966IMBALANCE
DISTANCE pter
DISTAL89.40NONONO
MICROSATELLITED10S579IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL89.50NONONO
MICROSATELLITED10S2394IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
TUMOR89.60N/AN/AN/A
SUPPRESSOR10q23.31
GENE (PTEN)PTEN
DISTANCE pter
PROXIMAL89.90NONONO
MICROSATELLITED10S541IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL90.00NONO63%
MICROSATELLITED10S2339IMBALANCEIMBALANCE
DISTANCE pter
DISTAL 7.10NONONO
MICROSATELLITED17S2159IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL 7.25NONONO
MICROSATELLITED17S1783IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
TUMOR 7.50N/AN/AN/A
SUPPRESSOR17p13.1
GENE (TP53)TP53
DISTANCE pter
PROXIMAL 7.60NONONO
MICROSATELLITED17S655IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL 7.70NONONO
MICROSATELLITED17S1796IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter

Example 3

Clinical History. An elderly patient developed anemia and hematuria, and was found to have a sigmoid colon cancer. The colon cancer was resected and mutational analysis was performed at three microscopic sites: center, 1 cm away from the center and 2 cm away from the center at the periphery of the tumor.

The materials and methods used to make the molecular diagnosis are as set forth in Example 1, supra.

Results. The hepatocellular carcinoma was determined to have two detectable allelic imbalance mutations affecting the CDKN2A and PTEN genes. The results of the markers for each gene analyzed are presented in Table 16

In both instances of mutation, it was determined that stability was present in the deleted genomic segment over the topographic distance. As a consequence, the patient's tumor was characterized as indolent. The failure to detect dynamic expansion of the deleted genomic segment is evidence of intrinsic indolent biological behavior. The prognosis of the patient is excellent.

TABLE 16
TUMORTUMORTUMOR
Mb distanceCENTERMIDWAYPERIPHERY
DISTAL111.10 NONONO
MICROSATELLITED5S2027IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL111.85 NONONO
MICROSATELLITED5S1965IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
TUMOR112.10 N/AN/AN/A
SUPPRESSOR5q22.1
GENE (APC)APC
DISTANCE pter
PROXIMAL112.20 NONONO
MICROSATELLITED5S346IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL112.40 NONONO
MICROSATELLITED5S1170IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL21.60NONONO
MICROSATELLITED9S736IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL21.8073%71%69%
MICROSATELLITED9S916
DISTANCE pter
TUMOR21.96N/AN/AN/A
SUPPRESSOR9p21
GENE (CDKN2A)CDKN2A
DISTANCE pter
PROXIMAL22.10NONONO
MICROSATELLITED9S1814IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL22.20NONONO
MICROSATELLITED9S966IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL89.40NONONO
MICROSATELLITED10S579IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL89.5049%52%NO
MICROSATELLITED10S2394IMBALANCE
DISTANCE pter
TUMOR89.60N/AN/AN/A
SUPPRESSOR10q23.31
GENE (PTEN)PTEN
DISTANCE pter
PROXIMAL89.90NONONO
MICROSATELLITED10S541IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL90.00NONONO
MICROSATELLITED10S2339IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL 7.10NONONO
MICROSATELLITED17S2159IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
PROXIMAL 7.25NONONO
MICROSATELLITED17S1783IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
TUMOR 7.50N/AN/AN/A
SUPPRESSOR17p13.1
GENE (TP53)TP53
DISTANCE pter
PROXIMAL 7.60NONONO
MICROSATELLITED17S655IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter
DISTAL 7.70NONONO
MICROSATELLITED17S1796IMBALANCEIMBALANCEIMBALANCE
DISTANCE pter

Example 4

A 59 year old man with longstanding heartburn underwent upper gastrointestinal endoscopy, which included a biopsy of the esophagus (FIG. 6). This revealed invasive adenocarcinoma that was treated by an esophagectomy.

One year later he was found to have a solitary nodule in the left upper lobe of the lung. He underwent fine needle aspiration of the lung nodule (FIG. 6). The cytology revealed adenocarcinoma. Despite immunohistochemical staining, there was uncertainty whether the lung nodule represented metastatic esophageal cancer or de novo primary adenocarcinoma of the lung. Treatment could not proceed until this issue was resolved. Resolution was definitive using microdissection based genotyping described herein, and as provided below (i.e., topographic genotyping).

Materials and Methods. The esophageal biopsy specimens were reviewed by microscopy and the most representative microscopic sites of cancer were identified (FIG. 6). Three such sites were selected for microdissection and mutational analysis. The materials and methods used to make the molecular diagnosis are as discussed in Example 1, supra.

Results. Striking discordance was seen between the mutational profiles of each tumor. The esophageal cancer possessed allelic imbalances of 1 p, 3p, 9p, 17q and 21q. These mutations were not present in the lung cancer, except for the 9p imbalance which affects the opposite allele. Similarly, allelic imbalance mutations and k-ras-2 codon 12 point mutation present in the lung cancer were discordant with mutations in the esophageal cancer. By using the ratio of peak heights as a measure of allele content, the proportion of mutated cells in a given microdissected sample could be calculated and compared. Mutations present over a wider distance (in multiple rather than focal cellular targets) and affecting a greater proportion of tumor cells, can be organized in a temporal profile (timeline) of mutation accumulation as being acquired earlier in time. The 9p imbalance, affecting alternative alleles, can thus be seen to have been acquired later in time in the lung cancer but earlier in time in the esophageal cancer. The information derived by microdissection genotyping provides unequivocal evidence that the two cancers are independent primary tumors enabling lung surgery to proceed for curative intent.

Example 5

A 58 year old woman was found on routine chest X-ray examination to have a mass in the right upper lobe of lung (FIG. 7). More detailed imaging revealed an additional mass in the right lower lobe. A right pneumonectomy was performed with both neoplasms found to be adenocarcinomas, containing areas of well, moderate, and poorly differentiated growth. At issue was whether the overall process represented mutlicentric cancer formation of lung cancer with intrapulmonary metastasis. This question could not be resolved by microscopic observation or immunohistochemical staining.

Materials and Methods. The lung tumor sections were reviewed by microscopy, and the most representative microscopic sites of cancer were identified for the upper and lower lobe cancers (FIG. 7). Multiple sites were selected for microdissection and mutational analysis (FIG. 7). The materials and methods used to make the molecular diagnosis are detailed in Example 1, supra.

Results. Comparative mutational profiling clearly indicated that the right lower lobe tumor was a metastasis of the right upper lobe cancer. The allelic imbalance mutations acquired early in tumorigenesis in the right upper lobe were all present in the lower lobe tumor. This time line concordance provided strong support for the metastatic origin of the right lower lobe cancer. Moreover, it was demonstrated that metastatic seeding occurred from an area of moderately differentiated right upper lobe tumor and that progression into poorly differentiated cancer was a phenomenon that took place in each deposit after metastatic seeding. This exemplifies the ability of topographic genotyping to not only resolve the issue of metastasis versus new primary cancer formation, but the method also delineates the unique time course and molecular pathogenesis of each individual cancer.

Example 6

A 70 year male was found to have an anterior gingival ulcerated mass discovered at the time of dental work. A biopsy revealed invasive squamous cell carcinoma (FIG. 8). Six months later, an enlarged lymph node was detected in the submandibular region and considered to represent metastasis from the anterior gingival cancer. However, a base of tongue squamous cell carcinoma was discovered at the time of lymph node enlargement. Thus, it could not be determined whether the metastasis came form the anterior gingival or the tongue cancer. Microscopic appearance and immunohistochemical staining were not of value in the discrimination.

Materials and Methods. The tumor sections of each cancer and the metastatic lymph node deposit were reviewed by microscopy and the most representative microscopic sites of cancer were identified (FIG. 8). The materials and methods used to make the molecular diagnosis are as discussed in Example 1, supra.

Results. As shown in FIG. 8, the lymph node metastatic cancer matched the anterior gingival squamous cell carcinomaalmost precisely, whereas it was discordant with respect to the base of tongue malignancy. Not only was the concordance seen at the level of specific affected markers, but also in terms of specific alleles affected and furthermore in the time course of mutation acquisition. This provides powerful support for the conclusion regarding the origin for metastatic cancer spread.

Example 7

Two patients were awaiting liver transplantation for advanced cirrhosis (FIG. 9). Both patients were identical with respect to clinical features, cause of cirrhosis and biochemical dysfunction. Both patients were found to have two nodules appear in the cirrhotic liver, in right and left lobes. Metastatic liver cancer is a contraindication for autologous liver transplantation, and therefore both patients might have been excluded from this therapeutic option at this point. However, liver cancer can arise in a multicentric fashion in this context and may not preclude the use of autologous liver transplantation for early stage cancer without spread. Biopsies were taken of both nodules in each patient with the purpose to differentiate liver cancer metastasis versus de novo multicentric cancer formation (FIG. 9).

Materials and Methods. All biopsy sections of cancer were reviewed by microscopy and the most representative microscopic sites of cancer were identified (FIG. 6). The materials and methods used to make the molecular diagnosis are as discussed in Example 1, supra.

Results. It was clear that the discordant mutational profiles in the first patient indicate that multicentric cancer is occurring. This patient received a liver transplant and has continued to do well. The concordant mutational profile in the second patient indicates metastatic liver cancer which precludes the use of liver transplantation. The use of this technology has helped to optimize the management of precious transplant liver organs and reduce health care costs by avoiding surgery in a patient destined to quickly evolve towards metastatic disease.

Example 8

A 58 year old man underwent resection of clear cell carcinoma of the kidney 5 years previously. The tumor was confined to the kidney without evidence of spread beyond Gerota's fascia or metastasis. Five years later a solitary lip nodule was noticed and biopsied. The histology was consistent with metastatic clear cell carcinoma of the kidney however a benign tumor of skin adnexal origin or minor salivary gland origin could not be excluded. The relationship between the prior kidney cancer the current lip tumor was sought using microdissection and comparative mutational analysis. See FIG. 10.

Materials and Methods: Sections of the renal cell carcinoma and lip tumor were reviewed by microscopy and the most representative microscopic sites of cancer were identified (FIG. 10). The materials and methods used are as discussed in Example 1, supra.

Results. A panel of tetranucleotide microsatellite markers were gathered as shown in FIG. 10. Only the markers on 3p were known from the literature to have been associated with renal cell carcinoma. There is scientific work that renal cell carcinoma is associated with trisomy 7. However, rather than introduce new markers which would need to be validated and thoroughly worked up, no chromosomal 7 markers were used. Instead other markers used for other forms of cancer were used as shown in FIG. 10. The behavior of these markers was well characterized. Conventional teaching would instruct that this panel is poorly designed and should not be used. Virtually every marker of this panel except the 3p markers would be considered inappropriate. Even the 3p markers would be considered unsuitable, because they are not situated close to the von-Hippel Lindau (VHL) gene which is mutated in renal cell carcinoma (RCC). Because the 3p markers are tetranucleotide microsatellites, conventional teaching would discourage their use in favor of dinucleotide microsatellites, which can be selected closer to the gene.

Notwithstanding conventional teaching, abundant mutational change is detected. Because the markers are tetranucleotide repeat polymorphisms, precise quantification of allelic imbalance can be achieved with more precise conversion into percentage of mutated cells affected by allelic imbalance. As a result, one is able to clearly show that not only the mutational profiles matched between the prior renal cell carcinoma and the current lip tumor, but the time course of mutation accumulation is identical between the two deposits affirming with certainty that the lip nodule of tumor is a metastatic deposit from the prior renal cell carcinoma.

Example 9

Experiences with 278 cases similar to those set forth in Examples 4-8 supra, were reviewed, where mutational profiling served as the basis for discrimination (FIG. 11). This series consisted of problematic cases in which the distinction between metastasis versus new cancer formation was not readily apparent. While epithelial malignancies dominated the series (i.e., adenocarcinoma and squamous cell carcinoma) many different histologic subtypes of cancer were included. A broad panel of 15-18 was sufficient in all cases to render a definitive discrimination. The marker panel was adjusted according to the tumor types under consideration with allelic imbalance serving as the main type of mutation being detected with lesser contributions from point mutational damage and microsatellite instability. In 71 patients, the results could be validated by outcome information based on clinical course.

The distinction between metastasis versus de novo cancer was straightforward using the degree of concordance and discordance. Metastatic tumors were fully concordant in most cases with only slight differences in mutational profile accounting for the remaining cases (FIG. 12). Similarly, discordant cancers were fully discordant to a degree that enabled the distinction to take place in a highly objective fashion. Concordance extended to the level of specific alleles affected and time course of mutation accumulation rendered the analysis highly sensitive, specific and objective.

The results showed that 100% discrimination could be achieved by mutational profiling making it the method of choice to accomplish this task (FIG. 13). The applicability to biopsy samples, cytology specimens and even free DNA in fluids (not shown) makes this a most powerful molecular pathology technique.

Example 10

The present example illustrates how the determination of the timing of sequential mutation acquisition in colorectal cancer can provide discriminating information regarding tumor aggressiveness. Current microscopic evaluation of colorectal cancer is designed primarily to confirm the presence of tumor, classify the degree of cellular differentiation and define pathology stage. Morphologic analysis cannot define inherent variability in tumor biological aggressiveness or treatment responsiveness on an individual patient basis, essential for optimizing oncologic therapy.

Based on the belief that phenotypic expression of colorectal cancer is determined by a specific constellation of acquired mutational damage on a per case basis, a microdissection approach targeting multiple microscopic sites for detailed molecular characterization of colorectal cancer was developed. This system permits delineation of the unique timeline of specific mutation accumulation for each cancer.

Methods. The predictive value of mutational profiling at multiple sites with time course determination was evaluated in a cohort of colorectal cancer patients. 21 colorectal cancer patients with T1N1 (N=4), T2N1 (n=5), T3N1 with liver metastasis at presentation (n=6) and T3N1 disease and long survival disease were identified. Each cancer was microdissected directly from unstained formalin fixed, paraffin embedded standard tissue sections at one non-neoplastic site (for determination of marker informativeness), two sites of greatest primary site invasion, and sites of lymph node and hepatic metastasis.

Each microdissected sample was quantitatively genotyped for a broad panel of mutational damage, including k-ras-2 point mutation and LOH at 1p36, 3p26, 5q23, 9p21, 10q23, 17p13, 17q21, 18q25, 21q23 and 22q12 (16 markers). The detailed mutational profile at multiple sites was used to define each tumors unique timeline of mutation acquisition.

Results. All colorectal cancer showed a corcordant constellation of mutations between primary and metastatic tumor sites ranging from x-y mutations representing the profile of the precursor cell at the time point of metastasis. There was a progressive increase in independent acquired mutational change in primary and metastatic sites correlated with increasing T stage supporting metastasis from precursor cells at an early stage of invasion. Total mutational load was higher in T3N1 patients with poor survival than for those with good survival (p<0.001). For T3NI neoplasms, precursor cell at the time point of metastasis was deeper in the wall than for patients with poor overall survival. Liver metastasis arose directly from precursor cells in the primary site rather than from metastatic lymph nodes at an early stage of invasion.

Thus, microdissection based mutational profiling at multiple sites in primary and metastatic colorectal cancer using clinical fixed tissue specimens in routine pathology can delineate the time course of mutation acquisition. More aggressive forms of colorectal cancer possess a higher mutational load early in tumorigenesis and are characterized by greater discordance in newly acquired mutations between primary and metastatic sites. In turn, this provides the basis for a new integrated molecular pathology approach to classify colorectal cancer in clinical practice.

Example 11

The present example illustrates the discrimination of cancer recurrence/metastasis versus de novo cancer formation. Distinguishing recurrent cancer from independent formation of a new cancer can be problematic, especially when the neoplasms share similar histologic features. Discriminating multicentric de novo cancer from intraorgan spread of malignancy can be difficult using staining techniques. Major treatment decisions may depend upon the certainty of this differentiation.

To provide definitive, objective discrimination, an integrated molecular pathology approach using tissue microdissection and genotyping of fixative treated tissue was used.

Methods. 278 problematic cases involving new primary versus recurrence/metastasis were collected over a four year period. Adenocarcinoma (n=211, 76%) squamous cell carcinoma (n=47,17%) were the most common shared histologies. When available, a minimum of 2 microdissection targets were obtained from each tumor using standard 4 μm recut sections. The mutational profile for each neoplasm, as well as the time course of mutation acquisition, was determined using PCR and quantitative genotyping for a broad panel of LOH markers targeting 1p, 3p, 5q, 9p, 10q, 17p, 17q, 21q, 22q, and k-ras-2 sequencing when suitable. The extent of concordant mutational markers and specific alleles affected was correlated with clinical and pathologic features of use in determining relatedness.

Results. Primary criteria for relatedness (recurrence/metastasis) between neoplasms was degree of concordance in LOH marker and specific allele affected by mutation and/or specific point mutation present (minimum two mutations present for evaluation). 159 paired tumors of this series showed concordance which ranged from 74% to 100%, supporting recurrence/metastasis. In 43 of these cases, where independent means existed to determine relatedness, all cases proved to be recurrence/metastasis (100% accuracy). In the remaining 119 patients, concordance ranged from 0-27% supporting de novo cancer formation. In 28 verifiable cases, this proved to be correct (100%).

Thus, discrimination between cancer recurrence/metastasis versus new primary cancer formation can be objectively and definitively accomplished by comparative mutational profiling.

TABLE 17
Primers
Primer NumberPrimer NameSequence
FORWARDD10S1173FAM-TCA TGC CAA GAC TGA AAC TCC
REVERSED10S1173GCT GGC CAT GAC TGT TTT AC
REVERSED10S520GTC CTT GTG AGA AAC TGG ATG C
FORWARDD10S520HEX-CAG CCT ATG CAA CAG AAC AAG
FORWARDD17S1161HEX-AAC AGA GCA AGA CTG TCC AG
REVERSED17S1161GTC CCT CTA ACC TTT AGG AGA G
REVERSED17S1289FAM-CTG CCT CTA AGC AGT CAT TTA GA
FORWARDD17S1289GCA TGG TCT TTT TCC ATT CC
FORWARDD17S974NED-AGC CTG GGT GAG AGT GAG AC
REVERSED17S974GCC ATT GTT AAC AGG TTG GTG
REVERSED18S814TET-CCC ACT ATA TGT ATG TTC ACC
FORWARDD18S814CTC TCT GCC TCT CCC ACC
REVERSED19S400NED-CAG GGT TCT TAT TTC CTG TC
FORWARDD19S400GCC TCT ATA AAT AAA TAA AGA C
REVERSED19S559TET-TTG AGG TAT CTA TGT GGA TAT C
FORWARDD19S559GAG TGA GAC CCT GTC TTT AC
FORWARDD1S1193TAMRA-TCG GCG ACA TAG CCA GAG
REVERSED1S1193CTT TGA TCT AAG GAT TAC CTA C
REVERSED1S407TGG GCG GGG GAT AGA AGG
FORWARDD1S407FAM-TGC TAA CCA GAT GGA GAG G
FORWARDD21S1244HEX-TCT TCT ATC TCA TAT GTG TAT C
REVERSED21S1244GGAGGAACTTGAGGATGTG
FORWARDD22S532FAM-CCTGGGCAACAGAGCGAG
REVERSED22S532GTCTGAGAAGATACTTGATATAG
FORWARDD3S1539NED-CTCTTTCCATTACTCTCTCC
REVERSED3S1539GTTCTCCATCTATCTTTCTCTC
REVERSED3S2303TET-CTCCAGAGCTTTGTTTTCAAC
FORWARDD3S2303GTGCCTACATGTTAGTATCC
REVERSED5S592TAMRA-TGGATACATATTGTTTTCTGCTG
FORWARDD5S592GGTGTCAACAAAGTAATGTAAAG
REVERSED5S615FAM-TCCACAGTGGTAAGAACCAG
FORWARDD5S615GAGATAGGTAGGTAGGTAGG
REVERSED7S1530FAM-ACACAGTCTTCAGCCCTAC
FORWARDD7S1530ACAGAGCTAAACTCTGTCTC
REVERSED8S373HEX-TGTATGCATTATTTCCTTCAATC
FORWARDD8S373GTAGGCAGGTGGGTGGGC
REVERSED9S251HEX-CAATACTTTTTAAGGCTTTGTAGG
FORWARDD9S251GTT TTATGTGCACTAACTAATGG
FORWARDD9S252NED-TTGTCAACTCCTAATATGGAC
REVERSED9S252GATATCCCCAAGTTCTCATAC
REVERSED95254GAGGATAAACCTGCTTCACTCAA
FORWARDD95254FAM-TGGGTAATAACTGCCGGAGA
FORWARDL-MYCFAM-TGAACCGTAGCCTGGCGAG
REVERSEL-MYCGCTGTTCTTCCTTTTAAGCTG
FORWARDD1S1172GAACAGAACCTGGTACCTTC
REVERSED1S1172FAM-CTGCACCCGGCTGATGTTC
REVERSEKRAS 2 Exon1TCCTGCACCAGTAATATGCA
FORWARDKRAS 2 Exon1TAAGGCCTGCTGAAAATGACTGA
FORWARDD5S346HEX-ACTCACTCTAGTGATAAATCGGG
REVERSED5S346AGCAGATAAGACAGTATTACTAGTT
FORWARDD5S1170FAM-AAATTAACTGTAACCCTAGTGTG
REVERSED5S1170TCTACCCTGAAGTAGCTCCAAAC
FORWARDD5S2027FAM-ACTTGGCAGATTTTCCACTC
REVERSED5S2027CACCTCATTGACTGGGAC
FORWARDD5S1965FAM-TGTCCCGTTGATAAAAATTACTGCG
REVERSED5S1965GTGTCTGGGATTTCCTACGCAATG
FORWARDD9S916HEX-ATCGATCAGGTGCCAGAG
REVERSED9S916AAAGGGAGAATATATTACTTGTGCAG
FORWARDD9S736HEX-TTCTAGACCTCTCAGCAGAC
REVERSED9S736GATAGTGTTGGAGACACCAG
FORWARDD9S1814FAM-CATGGTTCTTCTACTCAGG
REVERSED9S1814CAGAATGCTGGYYCTTAGCCTAGG
FORWARDD9S966GATTGTACCAACAGCACTGAG
REVERSED9S966FAM-CAGGCTTGATCAGCTTCCTGG
FORWARDD10S541HEX-AGAACTGGAATAGAAGGAAC
REVERSED105541AGGGATGGGTACAGAAACTC
FORWARDD10S2339FAM-CTCTAGGCTCATTTTCCTCTTCC
REVERSED10S2339CACAAATGCAGATAAAATAAAGACA
FORWARDD10S579HEX-ACTAGGAAGGTTCATATTCC
REVERSED105579TCATCCCCTGAAATACACC
FORWARDD10S2394HEX-CCAACGAGATGATGTTGCC
REVERSED10S2394AAGAGCTGGGTTGGGGTACT
FORWARDD17S655FAM-TGAACCCGGGAGGCAGAGC
REVERSED17S655GTGTGTGAGGCTGGTAGAACATG
FORWARDD17S1796GCTGATGTCAGACTGAGC
REVERSED17S1796FAM-CACAGATTAATAGAGATATTTTCCC
FORWARDD17S2159HEX-ACATTACCCCATGTCTTCATCC
REVERSED17S2159TCTATTGTTTGTGTTAGCCTTACCC
FORWARDD17S1783HEX-ATGTGTACAGAGCCGCAAAG
REVERSED17S1783CATGATTCAGCCACGGACTC

Primers are listed in 5′ to 3′ direction.

Example 12

A middle aged woman was being evaluated for pancreatic cancer using endoscopic ultrasound imaging and fine needle aspiration (FIG. 16). Two cystic lesions were found, next to each other in the head of the pancreas. Fine needle aspiration biopsy was performed of each cyst (cysts #1 and #2), and the fluid sent for cytology evaluation and molecular analysis. A buccal brush was also obtained to serve as a source of non-neoplastic DNA. The materials and methods used to obtain the molecular diagnosis are as set forth in Example 1, supra.

Results. Cytology examination of both cysts was similar and described as “inadequate for diagnosis; nondiagnostic”. The molecular profile of detectable mutational damage is shown in FIG. 16. The two cysts were sharply different in all molecular aspects. Cyst #1 had low amounts of DNA (2.60 ng/μL) that was of poor quality (qPCR above 30) and a CT-PCR value of 0.00 indicating that virtually all of the DNA was degraded. Cyst #2 had very high amounts of good quality DNA with small amounts of acceptable DNA degradation (CT-PCR value of 0.56 and 0.61 performed in duplicate. The molecular character of cyst #1 did not permit detection of allelic imbalance mutations which would be subject to a high rate of false positive results.

Allelic imbalance analysis of cyst #2 could be performed in assurance that values would be representative of the true status of the cyst fluid DNA. Three allelic imbalance mutations were detected and accurate quantification enabled their unique temporal profile of mutation accumulation to be determined. Cyst #2 could be shown to be malignant from a molecular perspective and the patient then sent for suitable treatment. In this case pancreatic cancer was discovered at an early and highly treatable stage.

The marker panel contains tetranucleotide microsatellite markers with well known beyond and abundant statistical information using normal samples. Only the markers for 18q22 and k-ras-2 point mutation could be considered as associated with genes known to be linked to pancreatic ductal cancer formation. Conventional teaching would view this panel as inferior and unsuitable for use. However, the advantages of the panel as described in this application as well as the results which clearly benefited the patient validate is efficacy.

Example 13

A 43 year old woman was found to have four colonic adenomatous polyps, fundic gland polyps of the stomach and a small solitary adenomatous polyp of the duodenum. The issue was raised where this patient might possibly have an attenuated form of familial adenomatous polyposis (FAP). Family history was not feasible as the patient was adopted and nothing was known of her biological parents. Sequencing of the APC gene was not considered an option as their was not a strong indication of FAP, and financially there were no funds for such a test without stronger indication of their being the good likelihood of point mutation. The materials and methods used to obtain the molecular diagnosis are as set forth in Example 1, supra.

Results. Because the APC gene is the focus of attention, three intragenic SNPs were utilized for allelic imbalance analysis. At the same time, tetranucleotide microsatellites at a distance from the APC gene were also included in the panel analysis (FIG. 17, lower left panel). Allelic imbalance was proven for the SNPs (FIG. 17, bottom right panel). However, the data from the tetranucleotide microsatellites indicated that each of the areas of tumor, available as fixative-treated, paraffin-embedded biopsy samples only (FIG. 17, top left panel) has different lengths of genomic deletional extension (FIG. 17, top right panel). Thus, all the tumor deposits were provided to be independent primary neoplasms, each with a concordant pattern of APC gene deletion. This result is exemplary of familial polyposis and was determined without sequencing of the gene, and the costs associated with such gene sequencing.

This example highlights the fact that the method described here does not necessarily operate against conventional teaching. When there is a good reason to include markers, such as the SNPs for APC here, that can be combined with the methods described in this application which together yield the best results for the patient.

Example 14

A 75 year old male with a long history of severe heartburn and Barrett's esophagus underwent biopsy surveillance of the Barrett's mucosa (FIG. 18). The histologic diagnosis based on microscopy was “low grade dysplasia with focal areas suspicious for high grade dysplasia”. Molecular analysis was required to arrive at a definitive diagnosis given that low grade dysplasia is treated conservatively, while high grade dysplasia requires ablation of the mucosa or surgical excision of the esophagus. The materials and methods used to obtain the molecular diagnosis are as set forth in Example 1, supra.

Results. The marker panel used in this example was based on tetranucleotide repeat microsatellites. None of the tetranucleotide repeats used are uniquely associated with specific cancer genes linked to esophageal cancer. The markers are situated close to well known tumor suppressor genes such as APC, CDKN2A, PTEN, and TP53. While mutations in these genes have been described in esophageal cancer they are not unique to that tumor type and can be seen in many different forms of human cancer. Rather, these primers sets are well characterized for molecular analysis and were successfully applied to the problem of Barrett's dysplasia.

Areas of low grade dysplasia showed two and three mutations, while the foci that were worrisome for high grade dysplasia showed 5 and 6 allelic imbalance (LOH) mutations (FIG. 18). The significant increase in acquired mutations in this panel is quite sufficient to affirm the presence of true definitive high grade dysplasia and enabled implementation of appropriate application of treatment for the patient.

Although the materials and methods have been described in detail with reference to examples above, it is understood that various modifications can be made without departing from the spirit of the materials and methods, and would be readily known to the skilled artisan.

This application also claims benefit of U.S. Provisional Application No. 60/620,926 filed Oct. 29, 2004; 60/631,240 filed Nov. 29, 2004; 60/644,568 filed Jan. 19, 2005; 60/679,969 filed May 12, 2005; and 60/679,968 filed May 12, 2005, all of which are herein incorporated by reference in their entirety for all purposes. This application also incorporates herein in their entirety for all purposes the following U.S. Patent Applications filed Oct. 24, 2005 entitled: “Molecular Analysis of Cellular Fluid and Liquid Cytology Specimens for Clinical Diagnosis, Characterization, and Integration with Microscopic Pathology Evaluation” and “Enhanced Amplifiability of Minute Fixative-Treated Tissue Samples, Minute Stained Cytology Samples, and Other Minute Sources of DNA.”

All cited patents and publications referred to in this application are herein incorporated by reference in their entirety for all purposes.