Objectives: To analyze the predictive power of Ki67 area% (Ki67), mitotic activity index (MAI), p53 area% (p53), and the mean area of the 10 largest nuclei (MNA10) for progression of stage in 195 primary consecutive TaT1 urothelial cell carcinomas of the urinary bladder.
Methods: Ki67- and p53-positive versus negative nuclei, MAI, and MNA10 using motorized systematic random sampling morphometry were determined. Kaplan-Meier curves and multivariate survival analysis (Cox model) were used to assess the prognostic value of the quantitative and classic clinicopathologic risk factors (age, sex, stage, grade, carcinoma in situ, multicentricity).
Results: Thirteen (6.7%) of the 195 patients had progression (0 [0%] of 36 low-risk, 1 [1.1%] of 85 intermediate-risk, and 12 [16.2%] of 74 high-risk patients). In univariate analysis (all variables), the strongest predictors with the highest hazard ratios were Ki67 (threshold 25.0%), MAI (threshold 30), and MNA10 (threshold 170 microm2). In multivariate analysis, the strongest independent combinations for progression--MNA10 (170 microm2) plus MAI (threshold 30) and MNA10 (threshold 170 microm2) plus Ki67 (threshold 25.0%)--overshadowed all other features. p53 was weaker but, combined with Ki67, still predicted progression fairly well. In the total group, the sensitivity, specificity, and positive and negative predictive values of MNA10-MAI and MNA10-Ki67 at the thresholds mentioned were 100%, 89%, 38%, and 100%, respectively. These feature combinations were also strongest prognostically in the high-risk treatment group.
Conclusions: The combined biomarkers MNA10-MAI or MNA10-Ki67 are accurate, well reproducible, and easy to assess progression predictors in all patients with TaT1 urothelial cell carcinomas, as well as in high-risk (bacille Calmette-Guérin-treated) patients.