Predicting the risk of metachronous colorectal cancer in patients with rectosigmoid adenomas using quantitative pathological features. A case-control study

The prognostic value of quantitative pathological features in rectosigmoid adenomas was investigated, in search of more precise identifiers of adenoma-bearing patients at high risk of metachronous colorectal cancer. In a 'nested case-control study' design, 21 cases with metachronous colon cancer during an average follow-up time of 16 years (range 3-30 years) after polypectomy and 67 controls were selected from a cohort of 1618 patients. The most advanced adenoma of each patient was analysed. Cases were matched with up to three controls simultaneously for size, grade of dysplasia, histological type, and number of adenomas, as well as for duration of follow-up. The patients did not undergo any post-polypectomy surveillance. Geometric characteristics of tumour nuclei, the arrangement of nuclei in the epithelium, and glandular changes were measured, and mitoses mere counted. Several quantitative features measuring nuclear polymorphism and crowding showed significant prognostic value, while those measuring glandular changes and mitotic activity did not. A multivariate combination of the average distance between nuclei and the standard deviation of nuclear area discriminated an unfavourable group (n=44) with 17 metachronous cancers from a favourable group (n=44) with four metachronous cancers (P=0.001, RR=6.3). With the optimum cut-off, 28 patients without any metachronous cancer were discriminated from a group of 60 patients with 21 metachronous cancers. In conclusion, in the present study, quantitative pathological features assessed in rectosigmoid adenomas showed prognostic value additional to traditional measures. These features may therefore be useful in guiding post-polypectomy surveillance. (C) 1998 John Wiley & Sons, Ltd.