Statistical Evaluation of Determinants of Progression of Superficial Bladder Cancer by Proportional Hazards Model

To evaluate clinical and pathological factors present at the initial consultation which affect disease progression, we reviewed data from 223 patients with superficial bladder cancer (pTa and pT1) who were initially treated at Nagoya University Hospital between January 1973 and December 1987. The factors included in the present analysis were age, sex, symptoms, interval between initial symptoms and first consultation, location of tumor, size, number, endoscopic shape, histological pattern of growth, grade and stage. The median duration of the follow-up after initial treatment was 46 months. Of the 223 patients, 17 died: 8 (3.6%) of bladder cancer and the remaining 9 (4.0%) of unrelated causes. Disease progression developed in 12 patients (5.4%): muscle invasion of the bladder wall in 11 and lung metastasis in one. The interval between initial treatment and progression ranged from 4 to 108 months, with a median of 11.5 months. Of the 12 patients, 9 (75%) had disease progression within 2 years. Progression was significantly associated with poor prognosis (p<0.001): the 5-year actuarial survival rates were 47.1% and 92.8% in patients with and without progression, respectively. Univariate analysis by Cox's proportional hazards model demonstrated that characteristics such as irritative bladder symptoms, higher-grade tumors, invasion into lamina propria, and nonpapillary growth seen at initial consultation were significantly related to disease progression. Cox's proportional hazards model produced hazard ratios of 10.2 in irritative bladder symptoms (yes vs. no), 6.3 in histological grade (grade 3 vs. grades 0-2), 4.9 in stage (pT1 vs. pTa), and 4.7 in pattern of growth (papillary vs. nonpapillary). Multivariate analysis by the proportional hazards model confirmed that irritative symptom was the most important factor affecting progression (p<0.001), followed by grade (p<0.05), stage and pattern of growth in this order. We conclude that meticulous follow-up is mandatory for patients with irritative bladder symptoms, tumors of grade 3, pT1 and nonpapillary growth which are present at the initial consultation.