Cytometry (Communications in Clinical Cytornetry) 18:79-87 (1994) zy Ploidy Pattern and Cell Cycle zy in Breast Cancer as Detected by Image Analysis and Flow Cytometry192 Pietro Luzi, Alessandra Bmni, Paola Mangiavacchi, Gabriele Cevenini, Dolly Marini, and Piero Tosi3 zyxwv Institute of Pathological Anatomy and Histology, University of Siena, Italy Both image analysis zyxwvutsrq (IA) and flow cytometry (FCM) may be applied to detect ploidy pattern and cell cycle fractions. However, they have different performance characteristics and may yield different results. The two approaches are applied in this study to 66 breast cancers: IA on imprints and FCM on fresh tissue. The percent coefficient of variation (CV) ranged from 2.0 to 7.0 (mean 5.5; SO 1.1) in IA and from 2.0 to 7.0 (mean 4.4; SD 1.1) in FCM. The values were well correlated. With regard to ploidy pattern, the agreement between the two methods was 92.4%; disagreements were due to four cases being aneuploid by IA but not detected by FCM and one case being aneuploid by IA but tetraploid by FCM. This suggests that IA is capable of detecting aneuploidy with more sensitivity than FCM. In diploid cases, the percent values of cells in GO/G1, S-phase (SPF), and G2M phase were concordant and well correlated. In aneuploid cases, IA was more sensitive than FCM in detecting aneuploid fraction as well as G2M phase, whereas FCM was more sensitive than IA in detecting SPF. A good correlation was found between the DNA indexes (01s) obtained with the two methods. E' 1994 Wiley-Liss, Inc. Key terms: DNA content, cell cycle, image analysis, flow cytometry, breast cancer Quantitative information on the proliferative charac- teristics of neoplastic cells is essential in tumor pathology (3,15.18). Additionally. the DNA index (DI) correlates well with chromosome number as determined by kary- otype analysis in solid tumors zyxwvutsrq (34), and aneuploidy appears to be an excellent marker for the malignant pop- ulations (5). The proportion of diploid and aneuploid populations ;is well as the fractions of cells in GO/Gl. S-phase (SPF), and G2M and of actively proliferating cells (i.e., the growth fraction) are especially important in determining transition from incipient to uncontrolled tumor growth (29), to predict prognosis (2,6,9,10,1 I, 17,21,24,27) and to improve cancer therapy ( 16,20,31, 3839). There are also a number of early malignant or premalignant lesions associated with distinctly abnormal DNA stemline (29). Extensive studies have been performed in this area by means of image analysis (IA) and/or flow cytometry (FCM), and since these two methods yield different per- formances, a number of attempts have also been aimed at comparing them (3,4,7,12,13,15,26,36,41). Unfortu- nately, a considerable variation in modalities of cell prep- aration and staining, instrumentation, choice of reference diploid standard, data presentation, and interpretation ex- ists in different laboratories worldwide (29), and data on the coefficients of variation (CV) of results are not easily available. Moreover, to our knowledge, no comparative @ 1994 Wiley-Liss, Inc. studies have been performed on cell subpopulations in dlfferent phases of cell cycle. In our study, we have analyzed a homogeneous series of breast cancers, comparing results of IA and FCM in detecting not only ploidy pattern but also fractions in different phases of the cell cycle, with special attention to Coefficients of variation of the GO/G1 peak. MATERIAL AND METHODS Instrumentation A Zeiss Universal microscope (Zeiss, Oberkochen, West Germany), was used with a 40 X planapo objective (NA 0.65) and a 10 X eyepiece. A stabilizing equipment (Farnell Power Supply, zyxw 30 V, 4 A; Farnell Instruments, Wetherby, West Yorkshire, UK) served to regulate lumi- nosity. An interactive and automatic image analyzer with a control computer (IBAS 1, Kontron, Munich, Germany) and an image processor (lBAS 2000, Kontron) were used 'Received for publication January 26, 1994; accepted January 26, 1994. 2This work was supported by a grant from the Ministry of University and Scientific Research. 3Address reprint requests to Piero Tosi, Institute of Pathological Anat- omy and Histology University of Siena, Via delle Scotte, 53100 Siena, Italy.