Visually imperceptible nuclear abnormalities detectable by high resolution image analysis, may serve as predictors of the ultimate outcome of disease states. The scope of this study is to demonstrate that the cells which we call """"""""moderate dysplasia"""""""" are in some way a pivotal point in the progression of dysplasia. As the majority of cases of cervical intraepithelial neoplasia (CIN) do not progress to life-threatening disease, the therapy of an individual could be based upon the predicted outcome of the lesion, thus avoiding expensive and possibly unnecessary operations, as well as undue psychologic trauma. In constructing a data base of uterine-cervical epithelial cells for an automated cytologic screening system, individual cells were classified according to their disease related cytomorphologic characteristics, independent of the histopathologic diagnosis of the tissue from which these cells originated. The cell class labeled """"""""moderate dysplasia"""""""" demonstrated significant differences among selected nuclear features, and to a lesser extent, so did the cells classified as """"""""severe dysplasia"""""""" and """"""""invasive carcinoma."""""""" In contrast, those cells categorized as most probably originating in a carcinoma in situ showed no significant differences regardless of the tissue diagnosis. In order to be sure that our preliminary data are not influenced by bias or artifact, many more cases and much more in depth analysis is necessary. Algorithms to more clearly define the differences among the nuclei need to be developed. Both of these goals, the collection of more data, and development of algorithms, are the reason for application for further funding. If this most abundant abnormal cell type, the """"""""moderate dysplasia"""""""" cell, can be used as a predictor of ultimate outcome of disease, an automated screening system can be made to function efficiently and economically, as theoretically shown in our previous work. Most importantly, patient care and therapeutic regimens will be optimized, and medical costs minimized. (3)
| Hall, T L; Castleman, K R; Rosenthal, D L (1988) Canonical analysis of cells in normal and abnormal cervical smears. Anal Quant Cytol Histol 10:161-5 |
| Hall, T L; Philippe, A; Rosenthal, D L (1988) Multifactor analysis of intermediate cells from the uterine cervix. The importance of slide effects on variance components. Anal Quant Cytol Histol 10:166-70 |
| Rosenthal, D L; Philippe, A; Hall, T L et al. (1987) Prognosis of moderate dysplasia. Predictive value of selected markers in routinely prepared cervical smears. Anal Quant Cytol Histol 9:165-8 |
| Wheeler, N; Suffin, S C; Hall, T L et al. (1987) Prediction of cervical neoplasia diagnosis groups. Discriminant analysis on digitized cell images. Anal Quant Cytol Histol 9:169-81 |
