Implementation and yield of upfront genomic profiling in a clinical prostate cancer diagnostic pathway.

Recent landmark studies have shown a survival benefit from using genomic profiling to guide targeted personalized therapy in men with metastatic castrate-refractory prostate cancer (mCRPC) [1]. However, the point at which genomic profiles should be acquired in the prostate cancer pathway remains unclear. Primary high-grade and/or locally advanced prostate cancer has a significant relapse rate despite radical therapy, with 30%–60% developing distant metastases, cancer-specific and all-cause mortality within a median of 6.5 years’ follow-up [2]. The UK National Prostate Cancer Audit has reported that more than a half of all new cancers diagnosed present with such aggressive disease that is National Institute of Health and Care Excellence [NICE] Cambridge Prognostic Group, CPG 4–5 or de novo metastatic [3]. This represents a large demographic of men with a high probability of progression to the mCRPC state, and in whom knowledge of their genomic status may help in managing their treatment. Here, we explored the evaluation of genomic mutation yield at diagnosis in these men by synergizing data from pre-biopsy imaging, guided biopsies and modern prognostic categorization. Patients referred to a NHS diagnostic service between January 2019 and December 2021 were screened using the pre-biopsy MRI report and presenting PSA level [4]. Eligible men (REC ethics approval 03/018) included those with raised PSA levels and MRI Prostate Imaging – Reporting and Data System (PI-RADS) 4–5 and/or evidence of ≥T3 disease or de novo metastatic disease. Biopsies were taken from image-targeted sites as well as systematic biopsies (no extra or specific research samples added) and prepared as standard (formalin-fixed paraffin-embedded [FFPE]). A uro-pathologist marked all cancer areas from target and non-target samples. Surplus-to-diagnosis FFPE sections matching the marked areas were sent for sequencing. The methods used for extraction, DNA isolation and sequencing are detailed in Appendix S1. For this study, a 350-gene capture panel (TWIST Biosciences) was used. A total of 62 men were recruited, of whom 52 (84%) had biopsy cancer content that were sufficient and met the required quality for sequencing. Cohort details are shown in Table S1. Of the 52 men, 37 (71%) had either CPG ≥4–5 (n = 30) or metastatic disease (n = 7) at diagnosis. These data confirm high yields of aggressive cancer detection by using pre-biopsy selection criteria of high PSA and MRI features and that samples for sequencing can be obtained without disrupting the standard diagnostic histopathology work-up. DNA damage repair (DDR) pathogenic mutations were detected in five individuals, with four harbouring BRCA2 (one germline) and one ATM (Table S2). All were in men with CPG 4–5/metastatic disease, representing 14% (5/37) of this subcohort. Currently based on published guidance the four BRCA2 mutations are potentially actionable with Poly (ADP-ribose) polymerase (PARP) inhibitors [5]. A further four cases were called BRCA2 or ATM loss by the mutation calling algorithm but were of uncertain clinical significance. In the phosphatidylinositol 3-kinase (PI3K)-Akt pathway, PTEN mutations were identified in four men and a further three were present in PI3K-related gene subunits, as well as one mutation in MTOR (total 8/52, 15%; Table S2). Seven were in men with CPG 4–5 or metastatic disease representing 19% of this subcohort (7/37). Based on eligibility for use of the novel AKT inhibitor capiversetib within the Capitello-281 trial (NCT04493853), mutations in PTEN and PIK3CA might be actionable, representing 10% (5/52 men) of the whole cohort and 11% (4/37 men) of the CPG 4–5/metastatic subcohort. Other pathogenic mutations identified are shown in Table S2, with the commonest being p53 (6/52 men; 12%). Focusing on the 37 men with CPG 4–5/metastatic disease, all but two were treated with combined androgen deprivation therapy (ADT) and radiotherapy or systemic therapy alone (ADT and chemotherapy). Follow-up was available for a median (range) of 26 (15–35) months. In this CPG 4–5/metastatic group overall, there were progression events in 12/37 men (32%): nine new or worsening metastases with four who eventually died from prostate cancer, three others developed rapid biochemical relapse after first-line treatment. Of the five men with DDR pathway mutations, two progressed: one with a BRCA2 mutation eventually died from disease and another developed new metastasis. Interestingly, one of the men with ATM loss of uncertain significance, relapsed rapidly after surgery. Of the men with PI3K-Akt mutations there was one death from prostate cancer and one progression despite treatment. Combining men with either a DDR (n = 4) or a PI3K pathogenic (n = 7) mutation, 4/11 (36%) progressed and/or died from disease (one individual had both BRCA2 and PI3K mutations). In men without any reported DDR or PI3K mutations (pathogenic or uncertain), 5/22 (23%) progressed in the follow-up period. We have shown that these steps are possible in a routine NHS diagnostic pathway and offer a rational way to maximize the detection of potentially actionable mutations while avoiding over-investigation. Using this method, DDR and PI3K mutations were detected in 11% and 19% of those with CPG 4–5/metastatic disease and these men seemed to be over-represented in cases that progressed over the short term (observational data). Our study was primarily descriptive as our aim was to assess feasibility and yield and we did not attempt to determine cost effectiveness. Nevertheless, as BRCA1/2 is already commissioned by NHS Genomics, incorporating routine profiling is already possible under NHS provisions. Any test is only of benefit, however, if it will alter management. In this regard, use of PARP inhibitors for treatment of BRCA-mutant tumours (including metastatic prostate cancer) are now approved by NICE (as well as the European Medicines Agency and US Food and Drug Administration) [6]. It could be argued that waiting for men to develop mCRPC before profiling may further reduce the numbers sequenced. However, our approach abrogates issues including (i) having to search off-site archives for archived material (ii) repeat sectioning and reporting and (iii) dealing with degrading DNA quality from stored FFPE. Furthermore, a priori knowledge of BRCA status may allow modifications at earlier stages of management and future proof for new indications for genomic profiling (including trials in earlier disease). BRCA2 status, for instance, is already an independent predictor of poor survival in the Predict Prostate algorithm [8]. In summary, we report a pragmatic method to optimize the selection of men for molecular profiling of prostate cancer at the point of diagnosis. Working within standard diagnostic pathways, our method could be adopted by any NHS (or other) unit without much extra resourcing. The upfront information could not only be used for future drug selection but could also influence decision making and personalized follow-up. We are grateful for help from Mrs Anne George in assisting with the study in its early phases and to the Cambridge Pathology Tissue Bank staff, especially Ms Elizabeth Cromwell who worked to collate and transfer samples to the Cancer Molecular Diagnostics Laboratory. We are also indebted to clincial colleagues who worked with the CUTRACT team to recruit men into this study with particular thanks to Mr Syed Shah and Mr Adham Ahmed. Professor Vincent J. Gnanapragasam acknowledges infrastructure and part funding from the National Institute of Health Research (NIHR) Cambridge Biomedical Research Centre (BRC-1215-20014). The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care. This project and publication are dedicated to the memory of Dr Charlie Massie. A very valued colleague and outstanding scientist, who was committed to the translation of bench to bedside but did not live to see this work come to fruition. Its success would not have been possible without his enthusiasm and championing of the multidisciplinary approach to prostate cancer research. None. Appendix S1. Supplementary methods and tables. Table S1. Demographics of the study population. Table S2. Pathogenic mutations detected in prostate biopsy samples and actionable status. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.