A clinical nomogram for predicting long-term survival in advanced colorectal cancer

From our prospectively accrued database of patients with gastrointestinal c ancer, 1057 patients with advanced colorectal cancer were identified with t he aim of determining predictive factors for survival of greater than 2 yea rs and to use this information to develop a predictive nomogram. Patient's baseline characteristics, type and number of chemotherapy regimens received , and response to chemotherapy were assessed by univariate and multivariate logistic regression comparing those who survived greater than or less than 2 years. A total of 161 (15.2%) patients survived more than 2 years, so-ca lled long survivors (LS). In multivariate analysis. positive predictive fac tors for LS were: good performance status (PS), normal serum carcinoembryon ic antigen (CEA), rectal primary, Dukes' stage A-B, well ol moderate differ entiation, two or less disease sites, response to chemotherapy and treatmen t used protracted venous infusion (PVI) 5-fluorouracil (5-FU) in first-line chemotherapy, and the increasing number of chemotherapy treatments receive d. From these PS, CEA, number of sites and response to first-line chemother apy were used to develop a nomogram capable of predicting the probability o f survival beyond 2 years for an individual patient. This large study confi rmed the relevance of known prognostic factors in metastatic colorectal can cer and demonstrated the importance of response to chemotherapy as an indep endent factor to predict LS. By combining these, we developed a nomogram wh ich provides informatiom which is likely to prove useful in the management of patients with advanced color ectal cancer. (C) 2000 Elsevier Science Ltd . All rights reserved.