A pathologist’s primer on implementing new standard-of-care molecular biomarker testing for precision prostate cancer management

Contemporary national practice guidelines already recommend multigene testing of both germline and somatic (tumor) DNA for many patients with prostate cancer. Mutations of homologous recombination repair genes, especially BRCA1/2, may secure clinical indications for PARP inhibitor drugs for metastatic, castration-resistant prostate cancer. The role of the practicing pathologist as steward of tissue and consultant integrating clinical, histologic, and molecular findings continues to grow. Few US Food and Drug Administration (FDA) drug approvals rise immediately to the consciousness of practicing surgical pathologists, but many may still recall the weeks and months following May 23, 2017. The FDA had just approved the anti–programmed cell death 1 protein monoclonal antibody pembrolizumab for a remarkable histology-agnostic indication: for unresectable or metastatic microsatellite instability-high (MSI-H) or mismatch repair deficient (dMMR) solid tumors that progress following prior therapy and for which no alternative therapy has been established.1 At the time, there was no approved companion diagnostic test for MSI-H or dMMR disease, and participation in the relevant clinical trials leading to the histology-agnostic approval was based on laboratory-developed tests performed in the community.2 (Contemporary prescribing information specifies use of an FDA-approved test for MSI-H or dMMR disease.3) In the weeks following the FDA approval, pathologists received numerous orders for MSI testing, especially dMMR immunohistochemical testing, in advanced solid tumors. The FDA’s first primary tumor type agnostic approved indication1 had thus led to perhaps the first time an FDA approval immediately changed surgical pathology practice. It heralded an era of an increased role for surgical pathologists as tissue stewards and re-reviewers vis-à-vis send-out testing requested by clinical teams to make treatment decisions. Genitourinary (GU) pathology has been a relative latecomer to routine molecular testing of archival materials. Certainly, observations such as defining chromosomal abnormalities of renal cell carcinoma subtypes4 or isochromosome 12p and 12p gains in the category of germ cell neoplasia in situ–derived testicular tumors5 validated our classification schema for these tumors.6,7 Unlike in hematopathology and other specialty areas, however, the practicing GU histopathologist generally did not devote the same level of effort to stewardship and review of archival diagnostic tissue for in-house and send-out molecular testing. Even commercial “black box” prognostic tests for prostate cancer8 have not, in our experience, reached the same level of saliency as in breast pathology. Outside of fluorescence in situ hybridization testing for TFE3/B-rearranged or amplified renal cell carcinoma,9,10 FGFR3 mutation testing for urothelial carcinoma,11 or ERG fusion testing to confirm prostatic histogenesis,12 molecular testing in GU pathology had not been a principal component of our practice or principal contributor to management of our patients. That era has passed, with a growing number of diagnostically and clinically useful testing options, as reviewed comprehensively recently.13,14 This special article is intended to provide a primer for the practicing surgical pathologist regarding the now standard-of-care use of molecular biomarker testing in prostate cancer, as touches on our role as stewards of tissue and consultants in interdisciplinary patient management. Simply put, we have noted underrecognition in the GU pathology community, perhaps even the urologic oncology community writ large,15 that current National Comprehensive Cancer Network (NCCN) guidelines (version 4.2024 at the time of writing)16 explicitly recommend molecular testing, both germline and somatic, for many patients with prostate cancer. In the case of somatic (tumor sample) testing, the recommendation is for multigene testing, including, at a minimum, BRCA1/2 and the related homologous recombination repair (HRR) pathway–related genes ATM, PALB2, FANCA, RAD51D, CHEK2, and CDK12 for any patient with metastatic prostate cancer. In addition, the recommendation includes testing for MSI-H and dMMR for patients with metastatic, castration-resistant prostate cancer (CRPC) for whom tumor mutation burden testing is also considered. (Similar to MSI-H/dMMR, tumor mutation burden–high, defined as ≥10 mutations/megabase, is also an FDA-approved immunotherapy biomarker for unresectable or metastatic solid tumors, including prostate cancer.17) The NCCN also recommends consideration of this testing for patients with regional-stage prostate cancer. Either way, such testing is often performed as part of large multigene or comprehensive genomic profiling studies using next-generation sequencing (NGS), either in institutional or national reference laboratories. From a practical standpoint and relevant to pathologists, the NCCN recommends directly sampling metastases for histologic and molecular evaluation.16 This sampling has the advantage of allowing confirmation of histology and prostatic origin, especially for metastases in older patients for whom other processes may clinically simulate prostate cancer metastasis. In addition, such sampling can exclude high-grade neuroendocrine transformation of prostate cancer, which often triggers different treatment options.18 With bony samples, it is exceptionally important to employ EDTA-based decalcification rather than most traditional protocols using strong inorganic acids, such as hydrochloric and nitric acids,19 known to degrade DNA and RNA.20,21 If it is not feasible to biopsy a metastasis, NGS testing of archival (formalin-fixed, paraffin-embedded) prostate cancer samples has been successful with the use of newer samples (<5 years of age), though approximately half of samples older than 10 years of age may yield results.22 For this reason, although not currently required in synoptic reports in the United States, the prospective addition of comments in pathology reports regarding preselection of appropriate tissue blocks with ample tumor content for molecular testing facilitates facile retrieval of archival material for future assays. In contrast, also commenting on cases where insufficient diagnostic tissue is present for future studies may also have value in this area of interinstitutional patient consultation for second opinions. We would emphasize the importance of the pathologist giving some consideration to the tissue-selection guidelines of the individual molecular biomarker assays to be used. Most often, the instructions provided by the HRR testing lab include selecting a block with the highest tumor grade, content, and purity, but these parameters are not necessarily a foregone conclusion across all methods and treatment planning needs, especially for clinical trials, which may have specific biomarker needs. In addition, careful consideration of the clinical contextual appropriateness of the particular sample on which additional testing is requested is warranted. On occasion, we have received testing requests designating use of limited prostate biopsy samples when subsequent prostatectomy blocks (likely yielding greater nucleic acid input and potentially more valid representation of the index cancer) were available, and we have intervened to facilitate changing the request to reference the more appropriate sample. Also, we have received requests for testing primary tumor tissues when subsequent metastatic samples were also available and might be more appropriate to the disease management question. More alarmingly, there have been occasions when testing of a patient’s prior tumor of another type has been requested mistakenly and needed to be corrected. Each of these scenarios highlights the pathologist’s tissue steward role23,24 and begs for close intradisciplinary coordination for safe, effective diagnostic care. Although in prostate cancer HRR mutations are thought to be mostly “truncal,” arising early during oncogenesis and therefore mostly presumed to be highly concordant between original and later metastatic or noninvasive circulating tumor DNA (ctDNA) samples,25,26 there is also some evidence of the emergence of these mutations in a subset of cases during the course of treatment.27,28 This finding supports the recommendation of testing a new sample if a long interval has elapsed. Sampling a metastatic lesion also has the advantage of avoiding concerns regarding many ctDNA-based assays, where low metastatic disease burden could result in assay failure or false-negative results, especially for genomic loss (deletion) of HRR genes in a tumor.29 Conversely, with ctDNA-based testing, a phenomenon known as clonal hematopoiesis of indeterminate potential (CHIP) can also produce false-positive results for HRR genes (in up to 10% in 1 study).30 For this reason, a variety of approaches31-33 are being explored and implemented to address concerns related to CHIP artifacts in this rapidly developing area. A similar technical issue to consider, germane to testing for dMMR and MSI-H phenotypes in prostate cancer, is that some experience suggests that the widely used “5-marker Bethesda panel” for multiplexed polymerase chain reaction–based MSI testing may be less sensitive for prostate cancer than NGS-based strategies,34 which seem to be relatively concordant with immunohistochemistry-based findings.35 To our minds, these concerns underscore the aforementioned NCCN recommendation of directly testing metastatic samples if clinically safe and feasible and support the growing use of comprehensive approaches to biomarker testing. It is acknowledged that pathogenic variants in HRR or MMR pathway genes detected by tumor somatic sequencing studies may reflect germline variants, and their detection may engender NCCN-recommended and American Society of Clinical Oncology–recommended genetic counseling and germline testing when encountered.16,36 Indeed, the NCCN separately and additionally recommends germline genetic testing in individuals where it is likely to affect disease management or inform risk of other cancers for the affected patient or family members. Independent of issues of involving trial eligibility, this is certainly true in the case of metastatic CRPC, where mutation of any of several of these genes may secure specific therapeutic indications. If so, multigene testing, including at least the MMR genes and a subset of the better characterized HRR genes, is recommended. In addition, the gene HOXB13, which is associated with substantially increased risk of prostate cancer,37,38 is included in the list. Unlike somatic tumor profiling, for which we select and release tissue, the surgical pathologist is usually unaware of these germline tests, often ordered on an outpatient basis from reference labs by oncologists and genetics specialists, using peripheral blood or mucosal swab samples. Yet, findings from these tests may arise in our interdisciplinary tumor board conferences, including for comparison with these patients’ somatic tumor profiles. In the past, prostate cancer had not been thought strongly related to germline or somatic deleterious variants of BRCA1 and BRCA2 or to HRR pathway genes (nor, for that matter, MMR genes); mutations in these pathways were detected only in a small subset of the earliest large comprehensive tumor sequencing studies.39,40 With experience with larger cohorts, however, expansion of multigene panels and increased emphasis on late-stage disease,25,41-43 DNA damage repair genes, principally HRR and MMR mutations, have been characterized in approximately 1 in 10 early-stage prostate cancers and approaching 1 in 4 metastatic cases, with BRCA2 most prevalent in most studies.44-49 When detected in tumors, deleterious DNA damage repair gene variants in sequencing assays are present in the affected patient’s germline approximately half the time upon dual comparative testing.50,51 Some available sequencing assays test both tumor and normal tissue for inference of variants’ germline vs somatic status; however, many assays do not, including some with FDA companion diagnostic assay status for testing HRR and MMR genes in prostate cancer. Thus, in many cases, these assays are ordered, run separately, and compared. The surgical pathologist might understandably wonder why all this testing is necessary. A substantial body of data indicates that testing for deleterious variants of these pathways may be prognostic and predictive. At a minimum, for an affected individual, germline status for BRCA2 has shown an association with numerous negative prostate cancer disease-specific prognostic and even standard treatment outcomes44,52 in addition to the opportunity to consider testing and surveillance recommendations for at-risk family members.36,53-56 Although this approach has not been as commonplace in urologic oncology practice, it bears consideration that uses of genetic testing and counseling for identification of hereditary disease are well established for breast and ovarian tumors. Yet, the reason for suddenly increased interest in this area primarily regards the predictive value of these mutations—specifically, HRR mutations in prostate cancer. As with cancers of ovarian, breast, and other organ sites, these mutations confer relatively distinct sensitivity to a class of drugs known as poly(ADP-ribose) polymerase (PARP) inhibitors, for which there are a growing number of FDA-approved indications for metastatic CRPC.57 From the standpoint of activity, these drugs exploit a phenomenon, originally observed in genetics, known as synthetic lethality.58 Simply put, synthetic lethality describes a genetic relationship where loss of function of both gene copies (but not either alone) is lethal to a cell. Applied to cancer drug development, synthetic lethality implies a scenario in which a drug causes cell death only in cells where both gene copies of a particular gene (eg, both the maternally and paternally inherited alleles) are mutated, sparing toxicity in normal cells. The PARP inhibitors inhibit DNA base excision repair function, which causes accumulation of single-strand breaks in DNA that are converted to double-strand breaks during cell replication. Tumors with loss of function of the HRR pathway (as with biallelic loss of function of BRCA2 in a prostate cancer) are sensitized to PARP inhibitors, because it is the HRR pathway that would have facilitated high-fidelity homologous recombination repair of double-strand breaks, rescuing the cell from more error-prone repair pathways that accumulate DNA damage and trigger cell death. This property, taken with observation of prevalent HRR mutations in metastatic CRPC that have limited treatment options, has opened the door to an exciting new direction for therapy, including monotherapy and in combination with high-potency androgen receptor signaling inhibitors (ARSIs).59 In addition, HRR deficiency is believed to confer relative sensitivity to platinum-based chemotherapy for prostate cancer,60 similar to other tumor types, for which experience is growing. The PARP inhibitors can be administered alone (olaparib or rucaparib) for patients with prior androgen-deprivation therapy and exposure to an ARSI (eg, enzalutamide, abiraterone) or concurrently with an ARSI (olaparib with abiraterone, niraparib with abiraterone, or talazoparib with enzalutamide) for patients with prior exposure to androgen-deprivation therapy only. Five phase 3 clinical trials (PROfound, TRITON3, PROpel, MAGNITUDE, and TALAPRO-2) have demonstrated that these PARP inhibitor regimens can prolong time to disease progression or death for patients with CRPC and germline or somatic HRR mutations.43,61-64 In the PROfound study, median overall survival was also longer with olaparib than an alternative ARSI, at 17.3 vs 14.0 months.65 Notably, the design of these studies varied with regard to the specific HRR gene mutations included and permitted chemotherapy exposure (whether prior taxane chemotherapy was allowed and whether docetaxel was allowed in the comparator arm), and appropriate patient selection for each regimen is critical. Pooled analysis of these and similar clinical trials suggests that PARP inhibitor efficacy differs by the mutated HRR gene, with the greatest benefit observed among patients with BRCA2 mutations: Among the trials assessing combination regimens, median overall survival was 33 months with PARP inhibitor therapy vs 24 months in the control arms for patients with BRCA2 variants.66 Conversely, improvement in progression-free survival or overall survival could not be clearly demonstrated for patients with ATM or CHEK2 variants.66 Among patients with BRCA1 or BRCA2 variants, benefit has been demonstrated for patients with alterations of both germline and somatic origin, reinforcing the importance of testing for each to determine PARP inhibitor candidacy.67 Homologous recombination repair gene mutation status may also ultimately inform treatment decisions in both earlier and later clinical settings. Two large phase 3 clinical trials are assessing the addition of niraparib or talazoparib to standard first-line treatment for patients with metastatic castration-sensitive prostate cancer and HRR gene mutations have closed to accrual, with reporting of results expected in the near future.68,69 For later-line therapy, patients with HRR gene-mutated prostate cancer are more likely to respond to platinum-based chemotherapy regimens than are patients without HRR gene mutations, and treatment responses to platinum-based chemotherapy regimens have been observed after progression on PARP inhibitor therapy.60 Prospective studies to better characterize clinical responses to platinum-based chemotherapy are ongoing (ClinicalTrials.gov identifiers NCT02955082 and NCT04038502). These findings, at the intersection of HRR deficiency, PARP inhibitor therapy, and platinum therapy, will be of particular interest because secondary “reversion mutations” restoring wild-type function of the relevant HRR gene (best characterized for BRCA1/2) have been identified as a mechanism of resistance to PARP inhibitor therapy.70 Indeed, similar types of reversion mutations have been identified arising dynamically at different time points during treatment with PARP inhibitors and with platinum-based chemotherapy.71,72 Establishing this upcoming standard of care for prostate cancer offers prognostic and predictive opportunities for patient care that were not available before. It bears noting, however, that key follow-on questions remain. Certainly, one principal question regards the relative importance of individual HRR pathway genes or even mutation types as a marker for prognosis or PARP inhibitor use. Although evidence for BRCA2 variants is well established, several of the other pathway genes included in PARP inhibitor clinical trials (and therefore listed as qualifying for the drug indications) may not confer the same sensitivity and drug benefit.44 Indeed, recently published experience suggests that in particular homozygous loss of BRCA1/2 (at an estimated 3% frequency in metastatic prostate cancer) may confer a more durable response to PARP inhibitors than BRCA1/2 variants because of the inability to form reversion mutations (given that the gene is deleted).29 There is even some evidence that the of of additional genes, with HRR gene mutations, may be regarding PARP inhibitor signaling the opportunity for in predictive For that matter, even for a HRR gene such as it can be for the pathologist or using many standard NGS to the status of deleterious variants and to whether the mutations mutations a vs mutations such as may be related to It is not a foregone conclusion that testing for individual genes or the predictive There is growing to that a from comprehensive NGS loss of and other parameters may be even more than individual an area of clinical regards whether high-potency now included with PARP inhibitors in approvals of drug may be of an relative sensitivity even to prostate cancers without Although of testing and is it is already that for prostate cancer, a new standard for care has been this is an area that has the potential to for growing of patients with prostate cancer, including many individuals with limited treatment Thus, GU pathology continues to other as one of diagnostic with assays, and comprehensive molecular We that in the years this role will only in but we that the gains in value for patient care will be This was to the interdisciplinary of the content by some of the and for the American Society for Clinical and Clinical on biomarker testing in prostate cancer, in part by an from for and from as well as a relationship with is an of a of in which also is a consultant current and in is an in has received institutional support the of and has for is a on a to the of on gene in prostate the diagnostic of use has been to to and is included in the as by the of