Lymph nodes in colorectal carcinoma. The Poisson probability paradigm.

Abstract

This article introduces use of the Poisson probability density function and Bayes probability rule to understand and analyze lymph nodal metastases in colorectal carcinoma. We apply these models to 370 patients who underwent resection of their tumors. In 213 cases with negative lymph nodes, we found that the Bayes-estimated probability of missed metastases was related significantly to subsequent survival (P = .0006). In 157 cases with lymph nodal metastases, we found that the estimated value of the Poisson parameter a was associated more closely with subsequent survival than pN stage (P = 9 10(-6)). Consequently, we believe that the Poisson model provides insight and useful results in colorectal carcinoma.