Multivariate approach for assessing the prognosis of clinical oesophageal carcinoma.

We investigated the prognostic contribution of 15 discrete variables on the survival of 123 patients with carcinoma of the oesophagus who underwent oesophagectomy and oesophageal reconstruction. In the multivariate analysis, three of fifteen variables were associated with the prognosis of patients who underwent oesophagectomy. The variables are: resectability of the malignant lesion, DNA distribution pattern of the cancer cells and postoperative complications. The DNA distribution pattern proved to be one of the most reliable prognostic factors. These three variables were assigned one of two values, 'favourable' or 'unfavourable', indicative of the effect on survival. The 123 patients were classified into four groups, characterized by eight possible combinations of these three variables. The survival rate of those with two or more 'favourable' variables was the highest, that is about 50 per cent at 5 years after surgery. On the other hand, patients with two or more 'unfavourable' variables had a very poor prognosis. Therefore, early detection and postoperative care play key roles in determining the survival of patients with oesophageal carcinoma.