USES AND LIMITATIONS OF ANCILLARY TECHNIQUES
APPLIED TO CYTOPATHOLOGY

Jeffrey S. Ross, M.D.

DNA PLOIDY ANALYSIS IN CYTOPATHOLOGY: DIAGNOSTIC ERRORS AND PITFALLS Gynecologic Cytology DNA ploidy analysis has been unsuccessful as a method for automating Pap smear screening. Although aneuploidy in cervical-vaginal cytology has been associated with high grades of cervical epithelial neoplasia, associated with clinically more significant types of human papilloma virus, and associated with progression to squamous cell carcinoma in-situ, it does not specifically identify invasive squamous cell carcinoma and may spontaneously disappear in Pap smears independent of specific treatment. DNA aneuploid status has been associated with the progression of low grade intraepithelial lesions to high grade lesions.97  DNA analysis has an unacceptable high false negative rate (diploid specimens positive for dysplasia) and an unacceptable false positive rate (aneuploid specimen negative for dysplasia).98  Potential uses for properly performed DNA ploidy and morphometric analyses in cervical-vaginal cytology include the differentiation of radiation reaction from recurrent squamous cell carcinoma99  and in the prediction of prognosis of squamous cell carcinoma of the female genital tract.100-101  Urothelial Cytology Flow cytometric DNA content evaluation in urothelial cytology was the first major clinical application of the technique. Whether performed by flow or image cytometry, significant errors and pitfalls must be considered. Virtually all grade I and many grade II urinary bladder transitional cell carcinomas feature diploid cell populations. Thus, false negative results in which diploid histograms are obtained from specimens that contain well differentiated transitional cell carcinoma are common. Combining DNA ploidy measurement with urothelial cytologic screening does not significantly improve sensitivity for the detection of low-grade transitional cell carcinoma, and thus cannot be recommended as an overall screening method.102  An additional pitfall in ploidy analysis of urinary cytology specimens is the occasional case of false positive aneuploidy in a benign specimen featuring nuclear changes characteristic of polyoma virus infection.103  DNA ploidy measurements may be relatively insensitive for the detection of small aneuploid subpopulations of cells when these cells are diluted out by an abundant diploid reactive cell population.104  DNA content analysis has, however, been valuable in differentiating chemotherapy effect from recurrent transitional cell carcinoma; in the prognosis assessment for cytologically positive specimens indicating risk for high stage disease; and as an adjunct to cytologic screening in high risk populations including patients with prior papillary transitional cell carcinoma in-situ diagnoses.105-106  Serous Effusion Cytology Studies of flow and image cytometric measurement of DNA content in serous effusion samples have featured varying success with false negative (diploid cases proved malignant) rates as high as 38% and false positive rates (aneuploid cases proved benign) rates as high as 14%.107-110  In general, ploidy measurements have been useful for the detection of malignancy in effusuin samples,111  although a variety of metastatic carcinomas may involve the body cavities that feature a diploid histogram and thus cause a "false negative" result. False positive rates are low, but on occasion aneuploidy has been detected in reactive mesothelial cells in a 2-5% range. Although the sensitivity of identifying a malignant fluid by aneuploid DNA content is low, averaging 44% for flow cytometry and 55% for image analysis with image analysis more capable of identifying small aneuploid populations112  two color methods employing epithelial markers and DNA analysis of the same specimen have significantly improved the detection rate.113  In a recent static image cytometry study, abnormal DNA content was a sensitive and specific indicator of malignancy in effusion cytology.114  However, in another study the contribution of image cytometry for the resolution of a borderline effusion specimen was judged to be limited.115  The positive predictive value of DNA aneuploidy in effusion samples equivocal for malignancy on light microscopy was nearly 98%.116  Flow cytometric DNA ploidy measurements have also been used in conjunction with immunocytochemical results with resulting improvements in diagnostic accuracy especially for difficult cases.117-118  Finally, it has been rerported that image analysis is more sensitive than flow cytometry for the detection of aneuploidy in serous effusion samples.118  Cerebrospinal Fluid Cytology Although hindered by low specimen volume, aneuploidy is highly specific for the presence of malignancy in spinal fluid specimens with false positive (aneuploidy in benign CSF samples) rates of less than 2%.119  Nearly all metastatic solid tumors and malignant gliomas detected in spinal fluid are aneuploid and the false negative (diploidy in malignant CSF samples) rate for the detection by flow or image cytometry of these specimens is quite low.119  A potential pitfall is the presence of leukemic or lymphomatous involvement of the spinal fluid featuring a diploid histogram, although most hematopoietic malignancies in spinal fluid are aneuploid. Gastrointestinal and Respiratory Cytology Although aneuploidy is a relatively specific indicator of malignancy in lung and gastrointestinal cancer, there has been limited use of the cytometric quantification of DNA content in brushings and washings. In Barrett's esophagus, brushing specimens indicating aneuploidy are associated with either high grade dysplasia or adenocarcinoma.120  Preliminary evidence suggests that aneuploidy may complement and confirm the diagnosis of malignancy in these specimens.121  Errors and pitfalls in interpretation of histograms obtained from brushing and washing specimens are related to specimen tumor cell count, presence of acellular debris, and the dilution of tumor cells by benign surface epithelial cells or inflammatory cells particularly when analyzed by flow cytometry. Fine Needle Aspiration Biopsy Cytology Both flow and image cytometry have been utilized to evaluate fine needle aspiration specimens both as an adjunct to diagnosis and for prognosis assessment of malignant lesions. The major pitfall for these specimens is that aneuploidy is not specific for the diagnosis of malignancy and may be encountered in a wide variety of benign neoplastic and non-neoplastic specimens (Table 10). Among non-malignant specimens associated with aneuploidy are: endocrine neoplasms; benign soft tissue proliferations, neoplasms and reactive conditions; colorectal adenomas; pigmented skin lesions; foreign body granulomas; viral infected cells; and normal seminal vesicles. Thus, over-reliance on ploidy pattern to confirm a diagnosis in a fine needle aspiration may lead to a serious diagnostic error and inappropriate conclusion that an ultimately benign lesion is malignant only because it features abnormal DNA content. Table 10: DNA Aneuploidy Reported in Non-Malignant Specimens Leiomyomas Colonic Adenomas Melanocytic Nevi Adenomas of Thyroid, Adrenal, Pituitary, Parathyroid Schwannomas Fibromatosis Foreign Body Granulomas Normal Seminal Vessicles Benign Tumors with High Lymphocyte Content