Multivariate evaluation of prognostic determinants in bladder cancer patients.

To clarify the relative importance of factors affecting the survival of patients with bladder cancer, a multivariate analysis by Cox's proportional hazards model was performed on 264 patients initially treated from 1973 to 1984 at Nagoya University Hospital. Clinicopathological data included in the analysis were sex, age, symptoms, interval from onset of symptoms to first consultation, smoking history and tumor characteristics (location, size, number, shape, histological grade and stage). The analysis revealed that stage is the most statistically significant determinant for survival, followed by size, irritative symptoms, age and grade in this order. The model composed of the above five determinants yielded hazard ratios of: 4.6 for stage (pT2-pT4 vs. pTa-pT1); 3.1 for size (greater than 3 vs. less than or equal to 3 cm); 2.5 for irritative bladder symptoms (present vs. absent); 2.7 for age (70 years or more vs. younger), and 2.1 for grade (high grade vs. low grade tumors). We conclude that these findings quantitatively confirm previous clinical impressions that accurate staging of the tumors is most important for improving the prognosis of bladder cancer patients, and that each of the above five determinants should be considered when planning an effective initial treatment regimen.