USES AND LIMITATIONS OF ANCILLARY TECHNIQUES
APPLIED TO CYTOPATHOLOGY

Jeffrey S. Ross, M.D.

USES AND LIMITATIONS OF MOLECULAR-BASED TECHNIQUES IN CYTOPATHOLOGY Diagnostic techniques using molecular biology procedures have recently been applied to cytologic specimens. Given the small cell volume and limitations of target DNA, RNA and protein available for analysis in cytologic preparations the molecular techniques most utilized have included: in-situ hybridization and polymerase chain reaction methods featuring DNA or RNA amplification with or without hybridization. Immunohistochemistry Immunohistochemistry has been extensively utilized in cytopathology to enhance diagnostic accuracy and provide prognosis information as described above. Many of the targets of the immunohistochemical procedures include proteins derived from molecular genetic systems including dominant oncogenes, tumor suppressor genes, invasion and metastasis markers, cell adhesion molecules, cell cycle regulators, and tumor-specific antigens. In-Situ Hybridization Genetic Disease Diagnosis The emerging technique of fluorescence in-situ hybridization (FISH) which allows for the detection of numerical and structural abnormalities of chromosomes has been applied to cytologic samples obtained from amniotic fluid, placental chorionic villi, testicular and ovarian germ cells and somatic cells.122-123  FISH lends itself particularly to use in cytologic preparations in that the cells are disaggregated and in a single layer in most cases allowing excellent technical results.124  A variety of preparations appear to work well including air dried, ethanol, methanol and Carnoy's fixatives uses. Major technical issues are common to all in-situ hybridization techniques and discussed below. Infectious Disease Diagnosis The in-situ hybridization technique has been successfully applied to cytologic samples particularly in the diagnosis of viral infections.125  The in-situ hybridization technique has been applied to the detection of HPV virions in cervico-vaginal cytology specimens.109  The method is now undergoing clinical evaluation in comparison with other non-morphologic-based HPV detection modalities for the ability to identify high grade dysplasia and to triage borderline cases (ASCUS and AGUS) into high and low risk categories (see below under PCR). Diagnosis and Prognosis Assessment of Malignancy Chromosomal Aneusomies The FISH method can detect gains or losses in chromosomes and compares favorably with the flow or image cytometric determination of tumor cell population DNA aneuploidy.126  Chromosomal aneusomies detected by FISH have also correlated with tumor grade, stage and subsequent clinical course127  and the ability to detect micrometastasis.128  They have been used to detect bladder cancer in urothelial cytology samples129-130  and malignant lymphomas in lymph node aspirates.131  Tumor Specific Translocations and Deletions In-situ hybridization has been used recently to confirm the diagnosis of malignancy by identifying clonal chromosomal markers in hematopoietic malignancies and outcome-specific chromosomal abnormalities in some tumors.132-133  The FISH technique has proved to be particularly useful for the detection of bcl-abl gene translocation positive leukemias.134  Recently described tumor-specific translocations for different types of sarcomas have been identified in cytologic specimens using the FISH technique.135-140  Cytologic smears and aspirates may actually be a better sample type than tissue biopsies for the performance f chromosomal translocation analysis by FISH.141  A table (Table 11a) of common chromosomal translocations for sarcomas and round cell tumors is provided below. Table 11a: Important Translocations in Sarcomas and Round Cell Tumors Tumor Translocation Involved Genes Ewings Tumor, PNET, Askin Tumor t(11;22)(q24;q12) EWS/FLI1 Desmosplastic Small Round Cell Tumor t(21;22)(q22;q12) t(12;22)(p22;q12) EWS/ERG EWS/ETV1 Clear Cell Sarcoma t(11;22)(p13;q12) EWS/WT1 Myxoid Chondrosarcoma t(9;22)(q22;q11-12) EWS/ATF1 Alveolar Rhabdomyosarcoma t(2;13)(q35;q14) t(1;13)(q35;q14) PAX3/FKHR PAX7/FKHR Synovial Sarcoma t(X;18)(p11.2;q11.2) SYT/SSX1;SSX2 Myxoid Liposarcoma t(12;16)(q13;p11) CHOP/TLS(FUS) Congenital Fibrosarcoma t(12;15)(q13;q25) ETV6/NTRK3 Oncogenes and Tumor Suppresser Genes The FISH method has recently been developed to perform tumor prognosis assays such as n-myc amplification in neuroblastoma and the HER-2/neu (C-erb-B2) oncogene copy number as a prognostic indicator in breast cancer.142  The FISH method produces more sensitive and specific identification of HER-2/neu amplification than does immunohistochemistry when samples have been damaged by fixation and processing or over-aggressive antigen retrieval has been performed.130  HER-2/neu gene amplification detected by FISH correlates with protein overexpression detected by IHC in >95% of cases for frozen tissue sections, but only from 80-85% of cases when formalin-fixed paraffin tissues are used. For a comparison of the advantages and disadvantages of IHC and FISH for the detection of abnormalities in the HER-2/neu gene and protein in breast cancer see Table 11b below. Abnormal expression of the ras oncogene can also be detected by the Fish method and shows promise for increasing the detection rate of small numbers of malignant cells in effusion samples.143  Table 11b: Advantages and Disadvantages of IHC and FISH for HER-2/neu Assessment Immunohistochemistry PROS WIDELY AVAILABLE SHORT PROCEDURE LIGHT MICROSCOPE INEXPENSIVE EXCEPT FOR HERCEPTEST™ FDA APPROVED FOR HERCEPTIN™ SELECTION CONS VARIETY OF ANTIBODIES POLYCLONAL (HERCEPTEST) VS MONOCLONAL (CB11) VARIABLE SENSITIVITY AND SPECIFICITY NO UNIFORMLY ACCEPTED THRESHOLD NO STANDARD SCORING SYSTEM FISH PROS HIGHLY SPECIFIC REAGENTS STANDARDIZED THRESHOLD FOR POSITIVITY BUILT-IN INTERNAL CONTROL QUANTITATIVE RESULTS SMALL NUMBER OF EQUIVOCAL CASES CONS NOT WIDELY AVAILABLE MORE EXPERTISE NEEDED LONGER TIME FLUORESCENT MICROSCOPE REQUIRED MORE EXPENSIVE (EXCLUDING HERCEPTEST™) NOT OFFICIALLY LINKED TO HERCEPTIN™ RESPONSE* Pitfalls and Errors in In-Situ Hybridization Techniques The major causes of errors in in-situ hybridization techniques are summarized in Table 12. Technical errors include: the loss of target DNA or RNA due to extensive degradation resulting from improper preparation, preservation and fixation of cells and tissues; insufficient cellular digestion preventing probe penetration for hybridization; over-digestion degrading the target DNA or RNA and preventing hybridization; the loss of probe specificity resulting from use of a probe with excessive length (too many base pairs in the oligonucleotide probe); the loss of probe sensitivity due to too short length of a probe (non-specific hybridization); excessive stringency washes (salt concentration or temperature) resulting in removal of significant hybridization; insufficient stringency washes allowing non-specific hybridization to remain; failure to utilize proper control systems to confirm technique specificity and sensitivity; selection of the wrong probe (DNA versus RNA probes); improper detection system (colorimetric versus radioactive versus fluorescence); improper immunohistochemistry detection technique results in false negative or false positive hybridization. Table 12: Errors and Pitfalls In-Situ Hybridization Techniques loss of target DNA or RNA due to extensive degradation resulting from improper tissue preparation unsatisfactory preservation and fixation of cells and tissues insufficient cellular digestion preventing probe penetration for hybridization overdigestion degrading the target DNA or RNA and preventing hybridization loss of probe specificity resulting from use of a probe with excessive length (too many base pairs in the oligonucleotide probe) loss of probe sensitivity due to too short length of a probe (non-specific hybridization) excessive stringency washes (salt concentration or temperature) resulting in removal of significant hybridization insufficient stringency washes allowing non-specific hybridization to remain after washing failure to utilize proper control systems to confirm technique specificity and sensitivity selection of the wrong probe (DNA versus RNA probes) improper detection system (colorimetric versus radioactive versus fluorescence) improper immunohistochemistry technique underdetects hybridization or results in non-specific staining and false positive hybridization. Polymerase Chain Reaction Methods for Cytologic Samples A wide variety of molecular biologic techniques have been applied to cytologic specimens for purposes of evaluating specimens for the presence of cancer, the prognosis assessment of cancer and the diagnosis of infectious diseases. Molecular techniques appropriate for cytology use have been continuously reviewed.144-150  The in-situ polymerase chain reaction method has been considered ideal for cytologic specimens containing small numbers of cells, but extensive utilization of the technique in cytology practice has not been prevalent. In a recent report, PCR was listed as a current cytology-related activity in the assessment of the epidemiology of HPV in Pap smears, the identification of HPV in squamous carcinomas, detecting ras mutations n pancreatic FNA specimens, the detection of occult lymphoma by bcl-2 gene assay, and the identification of mycobacteria in effusion samples.151  A modified PCR method to identify telomerase activity, the telomere repeat amplification protocol (TRAP), has been applied to a variety of cytologic samples including cervical-vaginal smears, effusion specimens,152  fine needle aspirates153  and, to the largest extent, urinary cytology samples.154-156  Telomerase levels have challenged conventional cytology in sensitivity and specificity for the detection of recurrent urothelial malignancy. Infectious Disease Diagnosis The pitfalls associated with PCR based techniques for the diagnosis of infectious disease have been reviewed.157  Laboratory contamination of the PCR reaction creating false positive results is the most significant pitfall. The in-situ PCR method has been particularly successful in the diagnosis of viral infections of single cells.158-159  The PCR method has recently been applied to Papanicolaou-stained smears for the detection of trichomonas DNA.160  For HPV Detection, PCR based methods such as the hybrid capture technique and on slide methods such as in situ hybridization have been used to detect the so-called "high risk" types of HPV infections in both conventional and monolayer cervico-vaginal cytology samples.161-162  Continued studies of large cohorts of patients will be needed to learn the relative value of the techniques for the prediction of outcome especially for patients originally diagnosed with borderline atypical smears. Other molecular targets have been identified and studied for their ability to characterize borderline atypias for their capability to progress to a high grade intraepithelial lesion.163-164  Detection of Recurrent Urinary Bladder Neoplasia Recently, a number of biomarkers have been developed to detect recurrent urothelial neoplasia and have been compared with conventional cytology.165-166  A comparison of the methods is shown in Table 13. Table 13: Biomarkers for the Detection of Recurrent Bladder Cancer Biomarker Technique(s) Sensitivity Mean/Range* Specificity Mean/Range* Comments Urothelial Cytology Conventional and Liquid-based Varies with Grade Varies with Grade High sensitivity and specificity for high grade disease only Bard BTA Test Latex Agglutination 60% 32-100% 77% 40-96% Significant promise as a cytology adjunct to increase sensitivity NMP22 Test Immunoassay 67% 47-81% 72% 60-86% Higher sensitivity than cytology for monitoring patients Telomerase Assays TRAP Bioassay RT-PCR 77% 62-93% 85% 60-99% Higher reported predictive value than BTA and NMP22 Microsatellite Assays LOH Mutation Analysis 89% 83-95% 100% 100% Highest reported sensitivity and specificity; needs automated assay Aneuploidy Aneusomy Flow Cytometry Image Analysis FISH Variable High Many reports without sensitivity and specificity data. Therapy-associated atypia is diploid. Keratins Proteins IHC Immunoassays 60-81% 80-97% CYFRA 21-1 test promising. Other monoclonal antibody tests are under development Hyaluronidase Assays Immunoassays 70-100% 89% Sensitive/specific for high grade disease; predicts progression Growth Factor Assays IHC Immunoassays Variable Variable bFGF and autocrine motility factor assays are most promising Cell Adhesion Molecules IHC Immunoassays Variable Variable E-cadherin upregulated in papillary tumors, down regulated in invasive flat high grade lesions Fibrinogen Markers ELISA 81% 75% AuraTek FDP Test™ has high sensitivity for low grade lesions Cell Cycle Regulators LOH IHC Not Known Not Known G1-S checkpoint assays. p16 gene associated with papillary lesions p53 and other molecular markers PCR Sequencing IHC Not known Not known p53 measurements guiding follow-up therapy: positive cases treated more aggressively TRAP = telomere repeat amplification protocol LOH = loss of heterozygosity FISH = fluorescence in situ hybridization IHC = immunohistochemistry ELISA = enzyme linked immunosorbent assay Applications in Hematopathology When used in the evaluation of potential clonal tumor cell expansions such as in the malignant lymphomas, PCR methods may be less sensitive than Southern blotting techniques for identification of gene rearrangements characteristic of clonality. Methods have been devised to isolate DNA from archival previously Papanicolaou stained cytologic smears.167  The potential for PCR based techniques using DNA extracted from fine needle aspirates and other cytologic materials to identify the presence of hematopoetic malignancy is considerable.168-170  The identification of clonality for lymphomas as well as oncogene mutations or amplifications; tumor suppressor gene deletions or mutations and other abnormalities may complement and, in some cases, out perform conventional classic morphologic assessment. Errors and pitfalls in these procedures will include contamination of PCR products, improper techniques in which sensitivity is lost, failure to use appropriate positive and negative as well as method controls, and over reliance on the molecular results when morphologic evidence is lacking. In the next several years many new molecular applications to cytologic diagnosis will be presented and this field appears to hold virtually unlimited promise. Oncogene Abnormalities A major example of potential use for molecular cytopathology is the application of PCR based assays for ras gene mutations in the diagnosis for ras gene mutations in the diagnosis of pancreatic and biliary tract carcinomas.80-82  Ras mutation assays may prove more sensitive and specific for the presence of malignancy than conventional cytology for fine needle aspiration specimens and fluid samples.171-172  A variety of molecular assays for oncogene activation (RAS), tumor supressor gene expression loss (p53), DNA instability (microsatellite analyses), and other abnormalities have been studied for their ability to detect early bronchogenic neoplasia in sputum samples.173  Tumor Suppresser Gene Mutations The identification of p53 gene mutations by PCR and sequencing methods in a variety of cytologic samples has shown promise in significantly increasing the sensitivity of cytologic diagnosis and detecting cancerous and precancerous lesions prior to the evidence of recognizable morphologic changes.174  Other Molecular Markers Among the additional PCR-based assays with or without gene sequencing assays used to complement cytologic diagnosis, measurements of telomerase activity and microsatellite instability have received the most attention. The telomere repeat amplification PCR procedure (see above) has identified telomerase, an enzyme most often detected in rapidly proliferating cells, in a variety of cytologic samples as an indicator of malignancy or pre-cancerous dysplasias.175  Microsatellite analysis (see above in urothelial malignancy detection) has been used to detect DNA instability and presumably faulty DNA repair mechanisms as a molecular indicator of malignancy.176  Chromosomal translocations specific for sarcomas, as described above, can also be accomplished using RT-PCR on FNA samples.177  Many molecular genetic techniques, new probes and markers and emerging targets are under investigation as to their potential clinical value in the further improvement on the identification and characterization of malignant processes encountered in the typical small sample submitted to the cytopathology laboratory.