V S Ghali1, S Liau, C Teplitz, R Prudente. 1. Department of Pathology and Laboratory Medicine, Beth Israel Medical Center, New York, NY 10003.
Abstract
BACKGROUND: Qualitative and quantitative analysis of cellular DNA content may be clinically useful in the prognostic evaluation of certain types of malignant tumors, including breast carcinoma. Flow cytometric (FCM) analysis has been the most frequently used procedure for DNA analysis, but it requires a reasonably large tissue sample. Computer-based image analysis (IA) now allows imprint, cytospin, and needle aspiration smear preparations and other small tissue samples to be used. METHODS: To resolve concern about the diagnostic efficacy of small tissue samples in the use of IA, the authors performed a comparative study of FCM analysis and IA using 115 fresh-frozen breast carcinomas. Feulgen-stained imprint preparations for IA and single-cell suspensions from the same fresh-frozen tissue for FCM analysis were used, and the respective histograms were compared. RESULTS: The results were concordant in 90.4% (104 of 115) of the cases, but 11 specimens yielded discordant data. IA provided histograms with a somewhat lower resolution and a relatively high coefficient of variation for the G0/G1 peak, thus rendering occasional tumors, which were near-diploid aneuploid by FCM analysis (four cases), not amenable to diagnosis by aneuploid characterization. In three additional cases, FCM analysis showed aneuploid hyperdiploid (two cases) and multiploid (one case) histograms, but IA only demonstrated a diploid peak. Conversely, in four other cases, aneuploid peaks were recognized only by IA. CONCLUSIONS: Computerized IA has significant advantages over FCM analysis, including lower cost, the ability to analyze very small specimens, the capability of detecting rare high ploidy cells, the capacity to classify cellular populations according to specific morphologic type, and the fact that no destructive enzyme or chemical digestion is required for specimen preparation, thereby preserving the integrity of fragile cells.
BACKGROUND: Qualitative and quantitative analysis of cellular DNA content may be clinically useful in the prognostic evaluation of certain types of malignant tumors, including breast carcinoma. Flow cytometric (FCM) analysis has been the most frequently used procedure for DNA analysis, but it requires a reasonably large tissue sample. Computer-based image analysis (IA) now allows imprint, cytospin, and needle aspiration smear preparations and other small tissue samples to be used. METHODS: To resolve concern about the diagnostic efficacy of small tissue samples in the use of IA, the authors performed a comparative study of FCM analysis and IA using 115 fresh-frozen breast carcinomas. Feulgen-stained imprint preparations for IA and single-cell suspensions from the same fresh-frozen tissue for FCM analysis were used, and the respective histograms were compared. RESULTS: The results were concordant in 90.4% (104 of 115) of the cases, but 11 specimens yielded discordant data. IA provided histograms with a somewhat lower resolution and a relatively high coefficient of variation for the G0/G1 peak, thus rendering occasional tumors, which were near-diploid aneuploid by FCM analysis (four cases), not amenable to diagnosis by aneuploid characterization. In three additional cases, FCM analysis showed aneuploid hyperdiploid (two cases) and multiploid (one case) histograms, but IA only demonstrated a diploid peak. Conversely, in four other cases, aneuploid peaks were recognized only by IA. CONCLUSIONS: Computerized IA has significant advantages over FCM analysis, including lower cost, the ability to analyze very small specimens, the capability of detecting rare high ploidy cells, the capacity to classify cellular populations according to specific morphologic type, and the fact that no destructive enzyme or chemical digestion is required for specimen preparation, thereby preserving the integrity of fragile cells.
Authors: C J Fabian; C Zalles; S Kamel; B F Kimler; R McKittrick; A S Tranin; S Zeiger; W P Moore; R S Hassanein; C Simon Journal: Breast Cancer Res Treat Date: 1994 Impact factor: 4.872