EVALUATION OF BREAST-CARCINOMA CHEMOSENSITIVITY BY FLOW CYTOMETRIC DNA ANALYSIS AND COMPUTER-ASSISTED IMAGE-ANALYSIS

Flow cytometric (FCM) DNA and S-Phase (S%) analyses were compared to computerized image analysis (SAMBA 2005) in 27 breast carcinomas (T3, N0-N1, M0) treated by 3 cycles of preoperative Adriamycin, vincristine, cyclophosphamide, methotrexate, 5-fluorouracil (AVCMF) chemotherapy (CT). Twelve carcinomas had shown objective regression and 15 no regression. Samples studied were obtained by sequential fine-needle cytopunctures. Comparing DNA profiles obtained by both methods before and after the first cycle, it appears that tumors can be divided into 3 groups. In the first group (10 cases), no changes were observed after the first cycle of CT. These tumors before treatment had either single DNA peak without cells in S% and G2M or a major peak with a small S% and G2M peak. The second group (9 cases) showed some changes in DNA profiles with an increased G2M peak but no additional values; these tumors before treatment had a small S% and a G2M peak. In the third group (8 cases), before treatment, all were non-diploid with high S% and high G2M. After the first cycle, all showed obvious changes in DNA profiles with a decrease of the G0/G1 peak and an increased S% and G2M with dispersed additional values along the scale in (G2M) x 2 and (G2M) x 4 regions. When changes were compared to tumor regression in the 1st and 2nd groups, 1/10 and 3/9 cases, respectively, were evaluated as objective regression. In the third group, all had objective regression (p < 0.001). In most cases, a good correlation was observed with both methods. Nevertheless, peak position definition was better with FCM than with SAMBA, but SAMBA permits a better estimation of changes in DNA patterns induced by CT. Furthermore, SAMBA gives information on changes in other morphometric parameters, such as nuclear area that significantly increased in chemosensitive cells (p = 0.003), and in chromatin textural parameters. Therefore, FCM and SAMBA give additional information on the early effect of CT on primary tumor cells and could help to select responsive tumors.