Nuclear morphometry as a predictor of disease outcome in gastric cancer.

Nuclear morphometric features including nuclear area (NA), nuclear perimeter (PE), their variation (SDNA, SDPE), and mean largest and smallest nuclear diameters (Dmax, Dmin) were examined in 116 patients with stage I-II gastric adenocarcinoma. Morphometric measurements were compared with tumour differentiation, depth of invasion (pT), lymph node status (pN), and Lauren classification. All morphonuclear features were significantly larger in the intestinal type than in the diffuse type of cancer. None of the morphometric variables was related to TNM status or tumour size. Nuclear area and perimeter and their variation were closely related to survival in univariate analysis, patients with small and regular nuclei surviving longer. In the multivariate analysis, pT, pN, perineural invasion, and the standard deviation of nuclear perimeter (SDPE) were independent predictors of survival. Nuclear morphometry is a quantitative, objective, and highly reproducible method of revealing malignant features in several neoplasms. The results of the present study suggest that nuclear morphometric data may help in defining prognosis in gastric cancer.