In the United States, prostate cancer (PC) is the most common cancer in men, with an estimated 288,300 new cases projected in 2023 representing roughly 29% of all new cancer cases in men.1 Men have a 12.6% chance of developing prostate cancer during their lifetime.1 Prostate cancer is also the second most common cause of cancer death in men with an estimated 34,700 deaths due to prostate cancer in 2023, representing 11% of cancer-related deaths in men.1 Prostate cancer incidence and deaths have increased since 2014 after two decades of decline; now, more than 30% of new cases are advanced cancers at diagnosis.1 Reports of a shift toward higher grade and stage at initial diagnosis have been attributed in large part to changes in prostate-specific antigen (PSA) screening practices.1 Prostate cancer incidence in African American men is 70% higher than in White men and prostate cancer mortality rates in African American men are approximately two to four times higher than those in every other racial and ethnic group.1 African American men may benefit more from screening and from testing for genomic biomarkers because they are more likely to harbor genomically aggressive cancer, even in categorically lower risk tumors.2,3 Most cases of prostate cancer are localized with excellent survival (>99%); however, the 5- year survival rate is substantially lower (~30%) for patients with distant metastasis.1,4
A mainstay of therapy in prostate cancer is androgen deprivation therapy (ADT). However, most castration sensitive prostate cancer (CSPC), also known as hormone sensitive prostate cancer (HSPC) eventually becomes resistant to ADT, a condition known as castration-resistant prostate cancer (CRPC).5 In recent years, the evidence has shown that combining additional therapies with ADT early during treatment for advanced and metastatic HSPC (mHSPC) prolongs time to CRPC and improves overall survival (OS).6-10 For this reason, combination therapy has become the standard of care in treating mHSPC.11-13 However, there are a number of options to use for combined therapy with ADT, including androgen receptor signaling inhibitors (ARSIs), chemotherapy, targeted therapy, immunotherapy, and combinations of these. Disease burden (i.e. volume of disease) plays an important differential role in response to therapy and is included in treatment guidelines.11,12 However, with the exception of specific recommended triplet therapies currently recommended for high volume mHSPC, the nationally recognized guidelines do not provide a preferred approach for choosing among the various options when dual combination therapy is appropriate.11-13 Further, some of the therapeutic options include agents with significant toxicities and side effects. In addition, prostate cancer is a heterogeneous disease with different underlying genomic signatures and phenotypes that respond differentially to available treatment options. As such, the use of molecular biomarkers to help guide the choice of therapy is a welcome addition to the treatment arsenal for advanced and metastatic prostate cancer.
Somatic and germline mutations in deoxyribonucleic acid (DNA) repair pathway genes occur in approximately 20% of prostate tumors.14,15 Hereditary prostate cancers are associated with hereditary breast and ovarian cancer (HBOC) and Lynch syndromes, resulting from germline mutations in homologous recombination repair (HRR) genes and DNA mismatch repair (MMR) genes, respectively.16,17 Patients with prostate cancer who have BRCA1/2 germline mutations in particular have increased risk of progression and decreased overall survival (OS).17,18 Patients with such HRR gene mutations respond to poly(adenosine diphosphate–ribose) polymerase (PARP) inhibition.19 Similarly, patients with MMR gene mutations, microsatellite instability and high tumor mutational burden may respond to pembrolizumab.20 An estimated 89% of mCRPC tumors contain a potentially actionable mutation; many of these are in the androgen receptor (AR) gene and AR-signaling pathways with germline mutations reported in approximately 12% of patients.15,21 For this reason, multigene tumor testing for somatic and/or germline mutations is recommended for patients with prostate cancer and is a covered service. Additionally, gene expression profile tests (GEPs) also take advantage of prostate cancer heterogeneity to further refine the risk for recurrence and metastasis and guide therapy selection for patients with prostate cancer.
Gene Expression Profile Tests
NCCN guidelines recommend the Decipher® Prostate (Veracyte) genomic risk classifier, a GEP, to inform adjuvant treatment if adverse features are found post-radical prostatectomy (RP) and to risk stratify patients with PSA resistance/recurrence after RP.11 Decipher® Prostate is a 22 RNA biomarker test that was developed using a whole-transcriptome based approach.22-24 The assay is performed on FFPE prostate cancer tumor tissue from the diagnostic biopsy or prostate resection tissue. Results are reported as a genomic classifier score (GC) between 0 and 1 (with higher scores indicating greater risk of metastasis) based on gene expression using a machine-learning algorithm. The molecular pathways represented include cell proliferation, cell death, invasion and metastasis, androgen signaling, immune activity and response, growth and differentiation, angiogenesis and metabolism functions. It has been clinically validated to predict the risk of prostate cancer metastasis at initial diagnosis or after radical prostatectomy.23,24
The Prediction Analysis of Microarray 50 gene expression profile (PAM50) is a well-studied primary tumor classification system initially developed from large-scale transcriptomic analyses of breast carcinoma.25-27 It is used to partition breast carcinomas into multiple intrinsic subtypes, including luminal A, luminal B, basal, HER2, and Normal-types. The classifier has subsequently demonstrated the ability to subtype other solid tumors, including urothelial and prostate cancers.28-30 In localized prostate cancer, a PAM50-based analysis of samples from seven cohort studies found that luminal B cancers displayed poor clinical outcomes following primary therapy but outcomes were improved with the use of ADT; further, luminal B tumors responded better than non–luminal B tumors to postoperative ADT.29 Additionally, prostate cancers classified by PAM50 as basal-like have been found to be highly enriched for low AR signaling, with features resembling mCRPC and high recurrence rates following primary treatment.31
More recently, the Decipher GC, PAM50 classifier and other gene expression profile (transcriptomic) tests have been used to prognosticate and predict response to therapy in advanced and metastatic prostate cancers. The Chemohormonal Therapy Versus Androgen Ablation Randomized Trial for Extensive Disease in Prostate Cancer (CHAARTED) and the Systemic Therapy in Advancing or Metastatic Prostate Cancer: Evaluation of Drug Efficacy (STAMPEDE) trials demonstrated a survival benefit of ADT in combination with docetaxel compared to ADT monotherapy, particularly in the subgroup with high-volume disease.6,7,32 A subanalysis evaluated the association of three different transcriptomic signatures (Decipher GC, PAM50, and Androgen Receptor Activity (AR-A)) with the response to therapy in men with mHSPC from the CHAARTED trial.33 Expression profiling was performed on primary prostate cancer tissue. The study found a nearly even split between luminal B and basal tumor signatures; luminal A tumors were lacking, the rationale being that these tend to be enriched in localized prostate cancer. Additionally, prognostic information was found to vary according to transcriptomic subtype; for example, higher GCs were associated with lower OS and ttCRPC. Moreover, a positive effect of chemohormonal therapy on OS was observed across all GC (risk score) groups; however the relative benefit varied by GC group and was significantly greater in those with the highest GCs, termed Quartile 4 (Q4) [Q4: hazard ratio (HR) 0.41 [95% confidence interval (CI) 0.19–0.84], p=0.01], compared to those with the lowest GCs, termed Quartile 1 (Q1) [Q1: HR 0.72 [95% CI 0.29–1.73], p=0.46].33 Similarly, OS significantly improved in the PAM50 Luminal B subgroup that was treated with docetaxel (median OS: 29.8 vs 52.1 months, HR 0.45 (95%CI 0.25–0.81), p=0.007).33 Though this effect was not seen in the basal subgroup (even in those with high volume disease), patients with both luminal B and basal subtypes showed an improved time to CRPC (ttCRPC) with the addition of docetaxel. Finally, the addition of docetaxel improved both OS and ttCRPC in all AR-A subtypes.33 Therefore, though the benefit of combination therapy was seen across all subgroups, the magnitude of benefit was greatest in patients with specific transcriptomic signatures, particularly those with high GC score and luminal B subtype. This information can serve as a guide when counseling patients regarding choice of one specific therapeutic option over another.
Similar findings have been reported in men with CRPC. The Selective Prostate Androgen Receptor Targeting with ARN-509 (SPARTAN) trial found that the addition of apalutamide (APA), a second-generation ARSI, to ADT significantly improved metastasis-free survival (MFS), time to second progression (also referred to as time to progression-free survival 2 (PFS2), and OS in men with nonmetastatic CRPC (nmCRPC).34 A subanalysis evaluated the association of the Decipher GC and the PAM50 signatures with the response to therapy in men from the SPARTAN trial.35 The study found that although all patients had improved outcomes with the addition of APA to ADT vs. placebo + ADT, those with higher GC scores showed the greatest improvements in MFS (HR 0.21; 95% CI, 0.11-0.40; P < .001), OS (HR, 0.52; 95% CI, 0.29-0.94; P = .03), and PFS2 (HR, 0.39; 95% CI, 0.23-0.67; P = .001).35 In the placebo + ADT arm, patients with higher-risk GC scores showed significantly shorter MFS than those with lower-risk GC scores. However, when patients received combination therapy with APA + ADT, there was a substantial improvement in MFS outcomes for the GC-high patients such that the GC high- and low-risk scores showed similar and overlapping MFS, suggesting that the addition of APA overcame the poor prognosis associated with a high-risk GC score.35 Finally, patients with the luminal subtype in the APA + ADT arm had a significantly longer MFS (APA + ADT: HR, 0.40; 95% CI, 0.18-0.91; P = .03; placebo + ADT: HR, 0.66; 95% CI, 0.33-1.31; P = .23) compared with patients with basal subtype; similar trends were observed for OS and PFS2.35
The gene expression association studies described above were conducted in primary tissue from prospective, double-blind, randomized trials, though the transcriptomic analysis was performed retrospectively. Both of the studies (one evaluating mHSPC and the other evaluating nmCRPC) found that although all patients seemed to benefit from combination therapy with ADT, the greatest benefit of combination therapy (with docetaxel or APA) was seen in patients with high Decipher scores and in those with the luminal B subtype of prostate cancer.33,35 Similarly, a retrospective analysis of 634 tissue biopsy samples from four cohorts of men with mCRPC found that patients with luminal tumors (found to be enriched with genes associated with androgen signaling) had significantly better survival (HR, 0.27; 95% CI, 0.14-0.53; P < .001) than patients with basal tumors (HR, 0.62; 95% CI, 0.36-1.04, P = .07) when treated with ARSI combination therapy; this study also noted that the basal signature was also seen almost exclusively in the small cell/ neuroendocrine subtype tumors.36 Importantly, in this study, when they examined patients who received prior ARSI therapy, the difference between luminal and basal tumors disappeared.36
Studies have also evaluated the transcriptome of metastatic prostate tissue samples. A recent study of men with mCRPC who progressed on ADT evaluated whether genomic and transcriptomic features from metastatic biopsies prior to treatment (with enzalutamide, abiraterone, or docetaxel) would be predictive of de novo treatment resistance to enzalutamide, which occurs in approximately one-third of men.37 The transcriptional findings demonstrated that gene sets linked to low AR expression and a basal/stem cell signature were activated in nonresponders and were associated with a more aggressive phenotype and worse outcomes, as has been reported by others.36-38 Lower AR transcriptional activity was associated with enzalutamide resistance; however, levels of AR mRNA and protein expression were similar between nonresponders and responders, suggesting that AR changes may not be good predictors of enzalutamide response.37 Notably, only a single metastatic site was biopsied prior to treatment; despite this, these biopsies did show features strongly associated with de novo enzalutamide resistance, suggesting that single-site biopsy may be adequate.37 Another retrospective study evaluated archived biopsies of mCRPC samples from men who had died of prostate cancer to look for whether the PAM50 classification can partition mCRPC into subtypes reflecting cell of origin and whether these correlate with specific genomic aberrations or cellular phenotypes.39 Basal tumors were found to be largely resistant to ADT plus docetaxel and, as noted in prior studies,37,38 the neuroendocrine (NE) tumors in the cohort were found to almost exclusively classify as basal.39 Importantly, these cells with small cell NE differentiation are generally AR-null.40,41 This is unlike the mCRPC adenocarcinomas that have retained AR signaling but still express luminal or basal signatures in near-equal proportions, indicating that there are multiple drivers of the mCRPC phenotype, an observation that also been documented in other reports.39,40 Finally, though phenotypes of most primary tumors and their metastases remained intact despite therapy with ADT, docetaxel and ARSIs, there were observed instances of discordance in cases where multiple metastatic tumors were acquired. This occurred in 40% of men (n = 23) for who at least one tumor received a discordant classification from other tumors; for a subset of these, the assignment to a particular phenotype lacked confidence.39 Excluding tumors with lower confidence PAM50 classification (<0.75) resulted in a 78% concordance across tumors within an individual.39 Multiple additional studies have also found (using a number of different genomic test methodologies) a high degree of homogeneity between metastatic lesions within an individual.42-44 Further research is needed to better define intra-individual genomic tumor heterogeneity and divergence as well as phenotypic plasticity that may occur as a result of pressures from systemic therapies.