Background: Treatment intensification with docetaxel or abiraterone improved survival for advanced prostate cancer starting androgen deprivation therapy (ADT) in the Systemic Therapy in Advancing or Metastatic Prostate Cancer: Evaluation of Drug Efficacy (STAMPEDE, NCT00268476) trial. However, survival and time-to-progression is highly variable on ADT, introducing the risk of unnecessary toxicity from additional treatments for some patients. Here we test the prognostic association of proliferation index using Ki67 scores in the control arm of the STAMPEDE population of high-risk localised (M0) and metastatic (M1) prostate cancer. Methods: Pre-ADT diagnostic needle biopsies were obtained from 517 men randomized in STAMPEDE arm A between 2006 and 2015. These were assessed for proliferation using an analytically optimised Ki67 immunohistochemistry assay. Ki67 was tested for associations with baseline clinico-pathological variables (Grade group, pre-ADT serum PSA and imaging metastatic burden) in univariable linear-regression models, and for associations with survival outcomes in multivariable Cox-regression models adjusted for these and additional confounding variables. Primary outcome measure was overall survival, secondary outcomes were prostate cancer-specific, failure-free, progression-free and metastatic progression-free survival. Results: Ki67 was available for 475 patients who received ADT only for at least 2 years ± radiotherapy. Of 202 M0, 74 were node positive. Of 273 M1, 116, 127 and 30 were respectively low, high and unknown radiological M1 volume. Ki67 score associated with higher Gleason (p=7.15x10^-11) and presence of extra-pelvic metastases (p=1.41x10^-8). Increasing Ki67 scores showed a strong linear association with poorer overall survival, with an estimated 2% increase in the hazard of death per percentage increase in the score (adjusted HR=1.02, 95% CI 1.01-1.02; p=1.04x10^-5). There was also strong evidence that Ki67 associated positively with all secondary outcomes, including prostate cancer-specific survival (adjusted p=5.50x10^-6) and metastatic progression-free survival (adjusted p=3.50x10^-9). Conclusions: Ki67 immuno-score is strongly prognostic in clinically advanced prostate cancer independent of Gleason score and the other clinicopathological variables tested in this study. Ki67 is a clinically scalable assay that could improve selection for treatment intensification and provide a tool for screening patients most likely to benefit from further molecular investigation.