Ki67 Is an Independent Predictor of Recurrence in the Largest Randomized Trial of 3 Radiation Fractionation Schedules in Localized Prostate Cancer

PURPOSE: To assess whether the cellular proliferation marker Ki67 provides prognostic information and predicts response to radiation therapy fractionation in patients with localized prostate tumors participating in a randomized trial of 3 radiation therapy fractionation schedules (74 Gy/37 fractions vs 60 Gy/20 fractions vs 57 Gy/19 fractions). METHODS AND MATERIALS: A matched case–control study design was used; patients with biochemical/clinical failure >2 years after radiation therapy (BCR) were matched 1:1 to patients without recurrence using established prognostic factors (Gleason score, prostate-specific antigen, tumor stage) and fractionation schedule. Immunohistochemistry was used to stain diagnostic biopsy specimens for Ki67, which were scored using the unweighted global method. Conditional logistic regression models estimated the prognostic value of mean and maximum Ki67 scores on BCR risk. Biomarker–fractionation interaction terms determined whether Ki67 was predictive of BCR by fractionation. RESULTS: Using 173 matched pairs, the median for mean and maximum Ki67 scores were 6.6% (interquartile range, 3.9%-9.8%) and 11.0% (interquartile range, 7.0%-15.0%) respectively. Both scores were significant predictors of BCR in models adjusted for established prognostic factors. Conditioning on matching variables and age, the odds of BCR were estimated to increase by 9% per 1% increase in mean Ki67 score (odds ratio 1.09; 95% confidence interval 1.04-1.15, P = .001). Interaction terms between Ki67 and fractionation schedules were not statistically significant. CONCLUSIONS: Diagnostic Ki67 did not predict BCR according to fractionation schedule in CHHiP; however, it was a strong independent prognostic factor for BCR.