Nomograms to predict outcomes after 177Lu-PSMA therapy in men with metastatic castration-resistant prostate cancer: an international, multicentre, retrospective study

BackgroundLutetium-177 (177Lu) prostate-specific membrane antigen (177Lu-PSMA) is a novel targeted treatment for patients with metastatic castration-resistant prostate cancer (mCRPC). Predictors of outcomes after 177Lu-PSMA to enhance its clinical implementation are yet to be identified. We aimed to develop nomograms to predict outcomes after 177Lu-PSMA in patients with mCRPC.MethodsIn this multicentre, retrospective study, we screened patients with mCRPC who had received 177Lu-PSMA between Dec 10, 2014, and July 19, 2019, as part of the previous phase 2 trials (NCT03042312, ACTRN12615000912583) or compassionate access programmes at six hospitals and academic centres in Germany, the USA, and Australia. Eligible patients had received intravenous 6·0–8·5 GBq 177Lu-PSMA once every 6–8 weeks, for a maximum of four to six cycles, and had available baseline [68Ga]Ga-PSMA-11 PET/CT scan, clinical data, and survival outcomes. Putative predictors included 18 pretherapeutic clinicopathological and [68Ga]Ga-PSMA-11 PET/CT variables. Data were collected locally and centralised. Primary outcomes for the nomograms were overall survival and prostate-specific antigen (PSA)-progression-free survival. Nomograms for each outcome were computed from Cox regression models with LASSO penalty for variable selection. Model performance was measured by examining discrimination (Harrell's C-index), calibration (calibration plots), and utility (patient stratification into low-risk vs high-risk groups). Models were validated internally using bootstrapping and externally by calculating their performance on a validation cohort.FindingsBetween April 23, 2019, and Jan 13, 2020, 414 patients were screened; 270 (65%) of whom were eligible and were divided into development (n=196) and validation (n=74) cohorts. The median duration of follow-up was 21·5 months (IQR 13·3–30·7). Predictors included in the nomograms were time since initial diagnosis of prostate cancer, chemotherapy status, baseline haemoglobin concentration, and [68Ga]Ga-PSMA-11 PET/CT parameters (molecular imaging TNM classification and tumour burden). The C-index of the overall survival model was 0·71 (95% CI 0·69–0·73). Similar C-indices were achieved at internal validation (0·71 [0·69–0·73]) and external validation (0·72 [0·68–0·76]). The C-index of the PSA-progression-free survival model was 0·70 (95% CI 0·68–0·72). Similar C-indices were achieved at internal validation (0·70 [0·68–0·72]) and external validation (0·71 [0·68–0·74]). Both models were adequately calibrated and their predictions correlated with the observed outcome. Compared with high-risk patients, low-risk patients had significantly longer overall survival in the validation cohort (24·9 months [95% CI 16·8–27·3] vs 7·4 months [4·0–10·8]; p<0·0001) and PSA-progression-free survival (6·6 months [6·0–7·1] vs 2·5 months [1·2–3·8]; p=0·022).InterpretationThese externally validated nomograms that are predictive of outcomes after 177Lu-PSMA in patients with mCRPC might help in clinical trial design and individual clinical decision making, particularly at institutions where 177Lu-PSMA is introduced as a novel therapeutic option.FundingProstate Cancer Foundation.