Trial of Pembrolizumab in Metastatic Castration Resistant Prostate Cancer

Open-label, single arm, phase II trial initially pursuing a two-stage Simon Minimax design. The null hypothesis Po will be 0.20; the alternative hypothesis Pa will be 0.40 (type 1 error will be 0.05 and type 2 error will be 0.10). Patients with metastatic CRPC tumours characterised by high mutational load, high microsatellite instability as established, for example, by the Promega MSI 1.2 analysis system, or a DNA repair defect including loss of MMR, identified on archival or fresh tissue biopsy specimens (through a TMG approved sequencing study e.g. the MAESTRO study) who have either exhausted established active anticancer drugs, or preferred not to have established agents, will be offered entry into the PERSEUS1 study. Part A: Part A is an open-label, single arm, two-stage Simon Minimax design phase II trial. Stage 1: Patients will continue their LHRH analogue therapy. This cohort will include 24 patients as the first stage of a two-stage Phase II. If more than 5 responses in these first 24 patients are reported then the study will proceed to stage 2. Anti-tumour activity will be assessed (measured by response rates) by PSA, imaging assessments (CT \& bone scans or when indicated whole body MRI) and CTC count measures. Whole Body MRI (WB-MRI) will be performed instead of CT in patients with contraindications to CT contrast), or instead of bone scan in patients with widespread bone disease at baseline thought to be non-evaluable by PCWG criteria due to inability to reliably identify two new lesions. Stage 2: This will enrol a further 21 patients to a total of 45 patients. Ineffectiveness will be concluded if ≤5 and ≤13 responses are seen in the stage 1 and stage 2 respectively and the null hypothesis will be rejected (i.e. further research would be warranted) if \>13 responses are reported from the first 45 patients. With this Simon Minimax design the probability of early termination of this trial when the true response probability (P0) is \<0.20 is 0.66. Part B: Biomarker enrichment stage: If the null hypothesis is rejected (and with Sponsor approval and confirmation of sufficient funding) the study will continue with stratified recruitment into biomarker defined cohorts in order to increase the precision with which the response rate can be estimated within mCRPC molecular subgroups of interest. It is anticipated that approximately 55 patients will be included in Part B, which will make a total of 100 mCRPC patients in part A and B together, including ≥9 patients for each of: A) MMR defective mCRPC; B) High mutational load mCRPC without MMR defects; C) mCRPC with deleterious homologous recombination (HR) DNA repair aberrations (BRCA2, ATM, BRCA1, PALB2, others); D) mCRPC with deleterious nucleotide excision repair (NER) aberrations; E) mCRPC with deleterious base excision repair (BER) aberrations; F) mCRPC with deleterious aberrations in non-homologous end-joining (NHEJ). Eligibility for more than one cohort is anticipated to be uncommon; where this occurs "allocation" to a cohort will be determined in discussion with the CI based on initial sequencing results and with priority given to driver (rather than sub-clonal) mutations. Patients recruited in Part A will be retrospectively "allocated" to a cohort on the basis of their initial sequencing data (reported prior to trial entry) prior to any response assessments, for the purposes of analysis. Within each biomarker subgroup (A-F) this will allow us to reject the probability of a \>30% response rate if we see no responses in 9-patients, with a 5% false negative risk (Gehan design). If ≥1 responses are seen in 9 patients in a subtype, recruitment will continue to that subtype with a further 20-25 patients such that the final estimate of the response rate has a standard error of 10%. Response rates with confidence intervals will be reported for each subtype. In the absence of any safety concerns from the IDMC, recruitment will continue seamlessly from Part A stage 1 to Part A stage 2 and from Part A stage 2 to the Part B biomarker enrichment cohorts. It is possible that at the time of the Part A stage 1 and 2 overall analysis (based on 45 patients), the graduation of some of the more common subgroups to the enrichment phase will be known (e.g. if more than 1 response from 9 patients recruited to a biomarker subgroup that cohort would continue to at least 20 patients). Conversely, if no successes have been seen in 9 patients of a particular subtype, recruitment to that subtype would not proceed to the enrichment stage. The biomarker defined cohorts may accrue at different rates related to genomic prevalence. At each stop/go decision point (end of enrichment stage 1 in a particular cohort) overall progress of all the ongoing enrichment cohorts will be reviewed by the IDMC/TSC. They will be asked to advise on the value of continued accrual to the cohorts particularly in light of slow recruitment.