Defining biology and recurrence risk in prostate cancers treated by neoadjuvant chemohormonal therapy
Year: 2023
Type: editorial
Abstract: The vast majority of prostate cancers are hormone dependent and respond durably to hormonal-based therapies in the recurrent and de novo metastatic settings (1), and androgen deprivation therapy thus remains the most common first-line systemic therapy. With androgen deprivation therapy as a backbone, the addition of more intense androgen receptor pathway inhibitors and/or chemotherapies in the newly metastatic, hormone-sensitive setting has demonstrated improved survival in numerous phase III studies including STAMPEDE (Systemic Therapy in Advancing or Metastatic Prostate Cancer: Evaluation of Drug Efficacy) arm C (2) and CHAARTED (ChemoHormonal Therapy Versus Androgen Ablation Randomized Trial for Extensive Disease in Prostate Cancer) (3). These findings, in part, have underscored support for a resurgence in identifying effective neoadjuvant therapies for patients with high-risk localized prostate cancer, moving treatments with efficacy in later settings to earlier in the disease course (4). Although many prostate cancer neoadjuvant therapies remain experimental with long-term survival benefits still uncertain, these studies also function as human model systems for evaluating treatments with tumor tissues due to be removed anyway, serving to elucidate the biology of drug responses and understanding mechanisms of resistance. In this issue of the Journal, Sumiyoshi etal. (5) perform an in-depth molecular analysis of surgical samples after neoadjuvant therapy to identify tissue-based correlatives with recurrence-free survival. Similar to breast and lung cancer, neoadjuvant therapy in prostate cancer seeks to intensify radical local therapy (in this case, prostatectomy) with presurgical systemic therapy to treat micrometastatic tumor subclones that would otherwise drive recurrence months or years after surgery. Completed in 2018, the Cancer and Leukemia Group B (CALGB) 90203 study tested whether the addition of 6 cycles of docetaxel plus androgen deprivation therapy for 18-24 weeks prior to surgery improved overall biochemical (prostate-specific antigen [PSA]) progression-free survival vs surgery alone (6). Importantly, this trial’s concept was designed before larger studies such as STAMPEDE reported their results, and the premise underlying the treatment paradigm was that the sensitivity of higher volumes of metastatic prostate cancer to docetaxel and androgen deprivation therapy should have similar effects on smaller micrometastatic foci. The rates of 3- and 5-year biochemical progression-free survival were not statistically different between the neoadjuvant arm and the surgery arm (0.89 vs 0.84; P = .11 for 3-year; 0.81 vs 0.74 for 5-year), but patients receiving neoadjuvant therapy still demonstrated a greater median event-free survival including freedom from additional treatment (4.53 years) vs those in the surgery arm (1.81 years) (6). Unlike many intense neoadjuvant androgen deprivation therapy trials lacking a taxane component, no patient experienced complete pathologic responses, and no correlations were observed between recurrence and tumor volumes on final pathology (7). Therefore, in the current study, the authors (5) examine molecular features of radical prostatectomy tissues to identify factors that are associated with event-free survival, using a capture-based DNA sequencing assay intended to identify point mutations in 73 genes associated with prostate cancer development, such as AR, TP53, and SPOP. Similar panels have demonstrated considerable utility in capturing pertinent targetable alterations and important driver events in other cohorts (8,9) while providing additional relevant estimates about tumor fraction and somatic copy number variation by virtue of aggregate analyses of each sample. A key finding is that patients in the neoadjuvant therapy arm (referred to as “arm A” in the article) harbor statistically lower mutant allele Sequencing-based tumor fractions (SeqTF) than the prostatectomy arm (arm B), and detecting any SeqTF in arm A is statistically associated with PSA progression (hazard ratio [HR] = 6.3, P < .005) and any-event progression (HR = 2.1, P < .005). The SeqTF difference between arms is not wholly unexpected, as samples in the prostatectomy arm should not harbor treatment effect, and thus the acquisition of tumor material for sequencing was likely more straightforward (Figure 1). Despite greater tumor cellularity in untreated cancer samples in the prostatectomy arm, a substantial number of tumors are reported as SeqTF of zero in both arms and are therefore considered not to harbor any molecular residual disease even though pathological cellularities are greater than zero. Although challenges associated with DNA sequencing assay performance in formalin-fixed material are a reasonable explanation for this finding, another important possibility is the contamination of neoadjuvant androgen deprivation therapy in the prostatectomy arm, as patients remained eligible for both arms of the study even after having received 3 months of neoadjuvant androgen deprivation therapy. Indeed, approximately 20% of the prostatectomy arm received androgen deprivation therapy, and a substantial difficulty in designing and executing correlative studies such as the current study is the confounding effect of treatment in visualizing tumor, especially when slides from this study were particularly limiting and may have precluded confirmatory stains to distinguish residual tumor from atrophic normal glands (Figure 1). This may have impacted sample acquisition from both study arms. Neoadjuvant therapy may confound distinction of tumor and benign prostate glands by routine hematoxylin and eosin (H&E) staining. A) H&E staining of an untreated prostate tumor, with tumor and benign glands readily distinguishable. B: left) H&E staining and (right) triple stain of a prostate specimen after neoadjuvant intense androgen deprivation therapy in which residual tumor foci and atrophic benign glands are challenging to distinguish without a confirmatory immunostain. Bar = 100 µm. B = benign; T = tumor. By interpreting the results of this study through the lens of these potential limitations, we can still identify additional insightful findings. For example, the authors report similar levels of TP53 point mutations in the neoadjuvant vs prostatectomy arms (13.8% vs 13.0%, respectively). This is greater than what was reported for prostate cancer in the Cancer Genome Atlas (TCGA) [7% (10)] and the Canadian Prostate Cancer Genome Network (CPC-GENE) [3% (11)], 2 cohorts comprised of nearly 800 untreated localized prostate tumors, suggesting that neoadjuvant androgen deprivation therapy (including at least a subset of approximately 20% of the prostatectomy arm receiving androgen deprivation therapy) with or without docetaxel may enrich for TP53 point mutations. Deleterious alterations to TP53 (often in conjunction with PTEN) contribute to castration resistant phenotypes and have been previously reported as a predictive feature of resistance to neoadjuvant androgen deprivation therapy (12,13). By contrast, the authors report SPOP point mutations enriched in the prostatectomy arm. SPOP mutations have been associated with greater androgen receptor activity and neoadjuvant androgen deprivation therapy sensitivity (12-14), and thus their less relative abundance in residual posttreatment tumor in the neoadjuvant arm (5%) vs the untreated prostatectomy arm (13%) or TCGA and CPC-GENE (9%) is consistent with prior findings. Importantly, leveraging these data against the long-term follow-up afforded by this cohort allows one to ask whether earlier introduction of these agents can influence the evolutionary trajectory of the tumor and potentially render future therapies less effective. Patients in the neoadjuvant arm who harbor TP53 mutations posttreatment have shorter durations of event-free survival (HR = 1.8, P = .05) and overall survival (HR = 4.5, P = .05), but stronger trends are observed in the prostatectomy cohort (event-free survival: HR = 2.5, P = .01; overall survival: HR = 5.0). Notwithstanding the proportion of prostatectomy patients who also received androgen deprivation therapy, this collectively suggests that neoadjuvant therapies may prolong event-free survival but do not seem to influence resistance to future therapies. Another common question arising from neoadjuvant therapy clinical trials is whether residual tumor adapts to therapy with different phenotypes that could be targeted by a precision adjuvant approach. In this current study, the authors also performed gene expression analyses using a targeted NanoString panel covering 155 prostate cancer-related genes. Consistent with the SPOP mutation trend, the PSA gene KLK3 as a measure of androgen receptor activity is more expressed in tumors from the prostatectomy arm, while tumors from the neoadjuvant arm show increased expression for markers of neuroendocrine differentiation such as NEUROD1 and CHGB. However, with a median event-free survival of 4.53 years, tumors in patients receiving neoadjuvant therapy had long durations with testosterone recovery before being exposed to further treatments, and 10-year overall survival probability in the neoadjuvant arm was 80%. Given the rapid decline of patients with advanced prostate cancer who progress to treatment-emergent neuroendocrine disease (15), it is unlikely that the tumor clones driving initial relapse years after the completion of neoadjuvant therapy are truly neuroendocrine in most cases. Opportunity for combination therapy or precision adjuvant approaches may have greater efficacy immediately after surgery for targeting remaining subclones rather than waiting until biochemical or metastatic progression. A lingering uncertainty that is not fully addressed in the current study is to what extent the observed features of residual tumor after neoadjuvant therapy arise as a consequence of selection by treatment or are representative of adaptive but preexisting subclones. Prior work by the authors on a subset of samples attempted to answer this question by longitudinal profiling of baseline biopsies with prostatectomy specimens (16). Considering TP53 point mutations for example, they were observed in samples from 5 patients at baseline as well as after therapy; in 2 patients, they were detected in baseline but not posttreatment; and in 4 patients, TP53 mutations were only observed posttreatment. However, 18-24 weeks of therapy is a relatively short time for prostate cancers to evolve appreciably to select for completely new mutations, which raises concerns about sufficient tumor sampling prior to molecular analysis. A likely model is that in at least a subset of patient samples, mutations such as TP53 arise prior to treatment and condition the tumor into a more plastic state that influences treatment outcome. By contrast, expression of neuroendocrine markers in prostate cancer prior to treatment exposure is rare. Increased expression of plasticity genes posttreatment likely reflects an adaptive state of residual tumor that is less dependent on androgen receptor and potentially more proliferative, driven in part by preexisting genomic alterations setting the fate of the tumor to progress. Indeed, tumors expressing greater levels of EZH2, a marker of phenotypic plasticity and emergence of castration resistance, were associated with greater SeqTF and earlier any-event progression (HR = 2.6, P = .01). An overall conclusion of the CALGB 90203 study was that neoadjuvant chemohormonal therapy did not statistically improve overall survival over prostatectomy alone and was similar to CHAARTED, which showed benefit in high-burden but not low-burden disease (3,6). Although these findings may cast neoadjuvant therapy clinical trials in a negative light, neoadjuvant studies are valuable for studying mechanisms of resistance, functioning as a tool to assess human tumor biology in direct response to drug exposures. The knowledge gained by the current study, and others like it, in evaluating the molecular features of treatment response, are readily applicable to advanced settings where overall survival benefits have been proven. Thus, even if the results of CALGB 90203 are not practice changing, the current study presents an array of DNA- and RNA-based markers that may aid to stratify patients with metastatic hormone-sensitive prostate cancer. No new data were generated or analyzed for this editorial. Scott Wilkinson, PhD (Writing—original draft; Writing—review & editing) and Adam G. Sowalsky, PhD (Conceptualization; Supervision; Writing—original draft; Writing—review & editing). This work was supported by the Intramural Research Program of the National Cancer Institute, NIH. Adam G. Sowalsky (AGS) reports that the National Cancer Institute (NCI) has a Cooperative Research and Development Agreement (CRADA) with Astellas. Resources are provided by this CRADA to the NCI. AGS receives no personal funding from this CRADA but is the primary investigator of the CRADA. Scott Wilkinson has no disclosures. The funder had no role in the writing of this editorial.
DOI: 10.1093/jnci/djad192
Authors: Scott Wilkinson, Adam G Sowalsky
Institution: National Cancer Institute
Language: English
Subfield: Pulmonary and Respiratory Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
Estimated APC paid: $4,978
OpenAlex ID: W4387475716