&NA;Methods of prognostic stratification for patients with rectal cancer currently rely solely on the extent of tumor penetration in the surgically excised rectum. Thus stratification according to prognostic factors can be performed only after treatment has been carried out and can play no role in pretherapeutic treatment decisions or in the design of trials of preoperative adjuvant treatment. To devise a method of clinical prognostic staging that could be applied to all patients with rectal cancer, regardless of therapy, the medical records were examined of an inception cohort of 282 patients treated by surgical and nonsurgical methods at three Toronto hospitals during the years 1968 to 1974. Six groups of clinical variables were identified that independently influenced prognosis: the presence or absence of metastatic disease, whether the rectal tumor was fixed or mobile, an annular rectal tumor, and the systemic symptoms of weight loss, anorexia, or weakness, and anemia. A composite method was developed for clinical prognostic staging from these variables and it was found that it created substantial gradients in survival. In patients treated by surgery, for whom comparisons of clinical staging with conventional anatomic staging were possible, clinical staging was as effective a method of prognostic staging as was anatomic staging. Furthermore, the prognostic impact of clinical staging persisted after adjustment for the effects of anatomic staging. This method of clinical prognostic staging should prove useful in making therapeutic decisions and in the design and analysis of clinical trials of alternative treatment for patients with rectal cancer.