Computerized nuclear morphometry is a useful technique for evaluating the high metastatic potential of colorectal adenocarcinoma

Abstract

BACKGROUND. Nuclear profiles have been reported to be useful prognostic predictors in various cancers. Data from computerized morphometry are objective and are quickly obtained by conventional microscopic analysis. However, this image analysis of nuclear features has been only rarely applied to investigations of colorectal adenocarcinoma. The aim of this study was to evaluate the correlation between the morphologic nuclear features and clinicopathologic parameters in cases of colorectal adenocarcinoma. METHODS. Morphometric nuclear features (nuclear area, perimeter, and shape) were analyzed in 343 patients with colorectal carcinoma and in 57 patients with colorectal adenoma. In each case, 300 nuclei of carcinoma or adenoma cells were analyzed on routine hematoxylin and eosin stained slides by means of a computer-assisted image analysis system that involved tracing the nuclear profiles (magnification X400) on a computer monitor. The morphometric data were compared with patients' survival, clinicopathologic status, and DNA ploidy pattern of tumors. RESULTS. The mean nuclear area (NA) enlarged from normal colorectal mucosa to adenoma and carcinoma (normal mucosa: n = 343, mean NA = 19 μm2; adenoma: n = 57, mean NA = 34 μm2; mucosal carcinoma: n = 15, mean NA = 45 μm2; P < 0.001). In 343 colorectal carcinomas, NAs of cancer cells in tumors with lymphatic invasion, venous invasion, lymph node metastasis, or hepatic metastasis were significantly larger than those of cancer cells in tumors without such factors. The mean NA of DNA aneuploid tumors was larger than that of DNA diploid tumors (P < 0.001). The nuclear area of cancer cells was determined to be one of the independent prognostic factors in multivariate analysis (P < 0.001). Moreover, the large nuclear area of cancer cells was recognized as one of the risk factors of metachronous hematogenic metastasis in patients after curative surgery. CONCLUSIONS. Data from computerized morphometry are objective and can be obtained rapidly by conventional microscopic analysis. The nuclear area of cancer cells appears to predict 1) the ability of cancer cells to invade the microvessels in the colorectal wall and 2) the ability of cancer cells to metastasize to the lymph nodes or liver. Therefore, nuclear morphometry is beneficial in mass screening to select patients who are at risk of hematogenic or lymph node metastatic recurrence after curative surgery for colorectal carcinoma.