Abstract
OBJECTIVE: To determine if individual, instead of group, patient progression risk could be predicted using p53, Ki67 and CK20 biomarker percentage values at initial transurethral resection of bladder tumor specimens. METHODS: This was an observational study where biomarkers were measured with no knowledge of tumor outcome. Initial bladder tumor specimens were classified as non-invasive and invasive to sub-epithelium (pT1). Percentages of stained biomarker cells were tested as progression predictors from non-invasive to pT1 and pT1 to pT2. Progression probability was correlated with biomarker percentages resulting in a regression equation. RESULTS: We studied 112 patients (median age = 67, range 37-91, males 83/112 (73%), with median follow-up of 39 months (range 1.7-140). Mean biomarker values were higher in stage pT1 than in non-invasive (all p < 0.001). Cut-off points separating progression from non-progression groups in stage pT1 were higher than in non-invasive for all biomarkers. Correlation R values for progression probability vs. biomarker percentages varied from 0.7 to 0.9 (all p < 0.001), regression slopes from 0.1 to 0.8 and intercepts from 11 to 35. A novel individual progression probability was calculated as the product of biomarker percentage of stained cells and slope, plus the prevalence-adjusted intercept. CONCLUSIONS: Identification of individual risk of progression in patients with non-muscle-invasive bladder tumors was possible using p53- and Ki67-derived progression probability using a regression equation. Combining biomarker-derived progression probability to tumor stage pT1 improves progression to pT2 predictive accuracy.