Abstract
The treatment of metastatic castration-sensitive prostate cancer (mCSPC) has seen remarkable breakthroughs over the last few years. Diagnostic and therapeutic advances have given rise to debates about risk stratification and optimal first-line treatment selection, as well as to concerns about potential overtreatment in a disease state with a highly heterogeneous clinical behavior. Here, we use case reports from our practice to review the clinical trials exploring intensified triplet regimens combining androgen deprivation therapy with second-generation androgen receptor signaling inhibitors and docetaxel, and we offer our recommendations on how to best select candidates for these novel combinations. Furthermore, the growing adoption of PET imaging with increasingly sensitive and prostate tissue-specific tracers replacing conventional staging technologies has led to the identification of a subset of low-volume mCSPC with nodal metastases which would otherwise not be considered abnormal by RECIST criteria. We describe our PSA-adapted approach to treatment in this unique population with non-measurable low-volume mCSPC which has not been specifically investigated in any phase III clinical trials. We also discuss ongoing clinical trials evaluating treatment de-escalation strategies. Finally, we review how local treatment modalities directed at the prostate or distant sites of disease in oligometastatic CSPC may benefit patients, and how we incorporate metastasis-directed therapy in the management of mCSPC.
Introduction
Prostate cancer (PC) is the second most common neoplasm in men worldwide and the most common in the United States. The number of new cases has been steadily increasing over the last decade and is estimated to impact 299,010 Americans in 2024, with the number of metastatic cases nearly doubling from 3.9% to 8.2%. More than 35,000 men are estimated to die of prostate cancer in 2024, representing the second most common cause of cancer death in the United States. 1
Metastatic PC represents an area of rapid diagnostic and therapeutic breakthroughs. As proof of this remarkable success a contemporary trial recorded a median overall survival (OS) surpassing historical estimates by greater than 16 months. 2 While several novel hormonal agents (abiraterone [Abi], Apalutamide [Apa], Enzalutamide [Enza], and Darolutamide [Daro]) and the chemotherapeutic Docetaxel (Doce) have proven highly effective in addition to androgen deprivation therapy (ADT) for the metastatic castration-sensitive PC (mCSPC) disease state, recent studies have leveraged distinct mechanisms of action and non-overlapping toxicities to combine these agents in triplet combinations.
Intensified regimens consisting of ADT, Doce and either Abi or Daro now have a proven OS advantage in mCSPC and current American Society of Clinical Oncology guidelines strongly recommend the use of such triplet combinations in all patients who are fit for chemotherapy.3–5 But mCSPC remains a highly heterogeneous disease state and subgroup analyses of phase III trials support that certain patients may achieve a similar OS benefit without upfront treatment intensification. Recognizing this aspect, investigators have more recently turned their attention to the study of risk stratification models in mCSPC and of treatment de-escalation strategies which may better balance efficacy and toxicity. As these efforts are ongoing, clinicians are confronted with individual patients for whom a more personalized approach can be devised using existing data. In addition to advances in therapy, diagnostic innovations such as the Prostate-Specific Membrane Antigen (PSMA)-positron emission tomography (PET) have increased our ability to detect a lower burden of neoplastic disease and, consequently, have identified a subgroup of patients with metastatic lesions detectable only by these sensitive methods but who would otherwise be categorized as having local or regional disease by conventional imaging. As this distinct population has not been directly studied in clinical trials, optimal management remains unclear and such patients are at risk for over- or undertreatment.
In this narrative review, we combine the best available data and our expert opinion to address still unanswered questions and ongoing debates pertaining to the best choice of first-line therapy, optimal risk stratification, and appropriate candidate selection for treatment in mCSPC. We illustrate our approach using common clinical scenarios to aid practicing oncologists in the increasingly complex decision-making process required for the care of mCSPC patients.
Risk Stratifying Metastatic Castration-Sensitive Prostate Cancer
Proposed Risk Stratification Criteria for Metastatic Castration-Sensitive Prostate Cancer.
In the CHAARTED trial investigating the addition of Doce to ADT with luteinizing hormone-releasing hormone (LHRH) agonists, patients with high-volume disease had superior outcomes with the addition of chemotherapy (an absolute prolongation of 17 months in median OS at the time of first analysis from 32.2 months to 49.2 months; hazard ratio for death, 0.60; 95% CI, 0.45 to 0.81; P < 0.001), but those with low-volume disease did not and had equivalent survival irrespective of the presence of synchronous or metachronous metastases or prior local therapy.8,10 At long-term follow-up of 54 months, all low-volume disease patients had an excellent median OS of nearly 60 months, highlighting the importance of weighing the 6% risk of grade 3/4 febrile neutropenia and 2% risk of infection with neutropenia associated with Doce. 10 Subsequent subgroup analyses of modern triplet trials have reached similar conclusions for low-volume disease patients. The addition of Daro to ADT + Doce in ARASENS (hazard ratio [HR], 0.68 [95% CI 0.41-1.13]) or of Abi to ADT + Doce in PEACE-1 (HR, 0.83 [95% CI 0.50-1.39); P = 0.66) did not lead to an OS benefit among low-volume disease patients and the median OS was not achieved by 5 years in this population.4,5
At the same time, observations from multiple trials call into question whether a favorable risk subgroup of mCSPC patients without an indication for chemotherapy derives significant survival benefit from second-generation ARSIs compared to older antiandrogens such bicalutamide (Bica). Within the limitations of cross-trial comparisons, studies exploring the addition of Abi, 11 Apa, 12 and Enza 13 to ADT have resulted in an OS of 82%–85% at 24-month with addition of ARSI. These OS rates are comparable to those observed in SWOG9346 which investigated continuous (cADT) vs intermittent (iADT) in mCSPC after an initial 7-month induction of cADT + Bica and reported that, in patients with 7-mo PSA ≤0.2 ng/mL, 24- and 36-month OS were 90% and 85%, respectively. 14 Similarly, in the control arm of the contemporary SWOG1216 trial consisting of ADT + Bica, an excellent 24-month OS of 83% was observed irrespective of PSA response. 2
The LATITUDE trial explored the impact of Abi in a predefined high-risk mCSPC population (Table 1) postulated to benefit most from the use of second generation ARSIs. 9 As defined by LATITUDE, this high-risk subgroup had a median OS of <36 months with ADT alone, whereas the addition of Abi increased 36-month survival to 66%. The added benefit of Abi in low-risk patients was subsequently investigated by a post-hoc analysis of patients with mCSPC from STAMPEDE who had been stratified using LATITUDE criteria. 15 In the low-risk subgroup, Abi demonstrated a significant, but modest OS advantage with a number needed to treat of 20 to prevent one death, as compared to 5 in the high-risk group. A subgroup analysis of low-risk patients included in the ARASENS trial found a statistically significant benefit for the addition of Daro to ADT + Doce (HR for death, 0.62 [95% CI 0.42-0.90]), but which translated into only a small absolute benefit as neither group had reached median OS. 5
The PSA response to therapy appears to be a better prognostic factor compared to pre-treatment variables and is likely the most robust such marker available at present. A PSA decline to ≤0.2 ng/mL at 7 months whether patients received ADT alone or combined with Doce was predictive of superior median OS compared to patients with PSA >0.2 ng/mL at this time point (60.4-72.8 months vs 21.6-25.2 months). 14 Notably, however, the addition of chemotherapy did increase the odds of achieving a PSA <0.2 ng/mL in the CHAARTED trial. More recently, a pre-planned analysis of the PEACE-1 trial evaluating doublet (ADT ± Doce) and triplet (ADT + Abi ± Doce) demonstrated that the 8-month PSA <0.2 ng/mL subgroup had not reached median OS after a median 4.4-year follow-up regardless of the administered treatment. 16 By contrast, patients who had a 7- or 8-month PSA >4 ng/mL had markedly inferior median OS irrespective of prior treatments.14,16 Analysis of contemporary data of patients treated with ADT with Bica or orteronel in the SWOG S1216 trial using the same PSA response categories both confirmed the value of 7-month PSA in predicting 5-year OS and found a similar predictive ability for PSA measured after 3 months of treatment. 17
The main barrier in implementing PSA response into clinical decision-making, however, is that it represents a posteriori knowledge, only available months after a treatment choice has been made. While it may inform treatment escalation or de-escalation decisions, it cannot aid in identifying the optimal initial treatment to balance toxicity and benefit for a given patient.
CASE 1: A 77-year-old practicing physician was admitted for severe low back pain. MRI of the spine showed a T12 pathological fracture and mixed lytic and blastic bone metastases. PSA was found to be >5000 ng/mL and staging CT of the thorax, abdomen, and pelvis, as well as bone scintigraphy showed bulky lymphadenopathy in the pelvis and diffuse bone metastasis involving the spine and the right femur. He underwent T11-L1 internal fixation and pathologic review of the T12 corpectomy specimen confirmed metastatic prostate adenocarcinoma. Next generation sequencing resulted as microsatellite stable, low tumor mutational burden, wild-type homologous recombination repair genes, wild-type TP53, PTEN and RB1. He started ADT with degarelix and Bica inpatient. 2 weeks later he presented to establish care in the oncology clinic with a plan to complete palliative radiation therapy (RT) to the T10-L1 vertebrae. What risk stratification criteria can guide the choice of most appropriate initial treatment for this patient?
The patient had both poor prognostic factors, such as de novo bone metastases at diagnosis, symptomatic metastases, high-volume disease by CHAARTED criteria, as well as favorable prognostic factors (absence of visceral metastases and preservation of functional TP53, RB1, and PTEN). While he would be a candidate for triplet therapy, the incorporation of chemotherapy could prove risky given his recent surgery and would have to overcome logistical challenges during planned RT to the spine.
CASE 1 continued: The patient completed palliative RT to the spine over the ensuing month and his PSA decreased from >5000 ng/mL to 713 ng/mL over the span of 5 weeks from initiation of ADT + Bica. The first-generation antiandrogen was then changed to Abi and 7 months into combined ADT + Abi, his PSA decreased to 0.97 ng/mL. After 19 months of treatment, PSA became undetectable at < 0.1 ng/mL and, after 25 months, he elected to take a break from Abi due to progressive fatigue. The PSA remained <0.1 ng/mL for an additional 18 months before it began to rise slowly to 0.61 ng/mL 30 months after discontinuation of Abi while on ADT monotherapy. At the time, computed tomography and bone scans showed no new lesions, stable known bone metastases and no measurable lymph nodes. What is the next best course of action?
In this case, ADT + Abi resulted in a rapid, deep, and durable response and the current PSA while off Abi, although slowly rising, remains under 1 ng/mL. For patients with undetectable PSA while on combined ADT + ARSI, it is reasonable to stop ARSI and continue ADT alone as maintenance. “Treat until progression” is the strategy used in all the phase III trials comparing combined ADT + ARSI with ADT monotherapy in mCSPC. Although this is a conventional way to investigate the efficacy and safety of an experimental drug for FDA approval, we have shown in early phase clinical trials and using mathematical modeling that this strategy tends to select for treatment resistant clones and that it is probably not the best approach to maximize benefit.18–23 Stopping ARSI after adequate cancer control has several advantages such as reducing long-term toxicities (including financial toxicities) and would preserve ARSI rechallenge as an option when the cancer progresses to the castration-resistant stage. Finally, prolonged combined ADT can increase the risk of transformation of prostate adenocarcinoma to the aggressive androgen-independent neuroendocrine or small cell prostate cancer. 24 In our current practice, we offer discontinuing ARSI and maintenance treatment with ADT monotherapy as an option for patients with mCSPC who have achieved <0.2 ng/mL PSA within 7 months of combined ADT.
CASE 1 continued: The patient was intermediate risk based on his 7-month PSA response according to SWOG 9346 criteria. He elected to continue ADT monotherapy and his PSA remains <1 ng/mL 5.4 years after initiation of systemic treatment.
For decisions regarding initial treatment, mCSPC can broadly be stratified into high volume based on CHAARTED criteria and high risk based in LATITUDE criteria, and low volume and low risk if patients do not meet these respective criteria. Within the high-volume and high-risk categories, patients with liver metastases have an exceptionally poor prognosis and merit special consideration. Among low-volume, low-risk stage IV patients, treatment can be augmented with incorporation of local modalities such as RT to the prostate in the presence of locoregional nodal metastases only (TxN1M0) or both the prostate and distant metastases if few (oligometastatic disease). For many patients risk-adapted strategies based on PSA may balance both the benefits and long-term risks and burden of treatment. Our proposed approach is outlined in Figure 1. In brief, we continue to divide mCSPC into high-volume and low-volume disease as defined by the CHAARTED study given the significant and meaningful OS benefit derived by high-volume disease patients receiving Doce across multiple trials.8,10,25,26 We prefer triplet regimens for patients who are fit for chemotherapy and have high-risk disease in excess of that defined by high-volume of metastatic burden, such as those with synchronous, symptomatic, or hepatic metastases. Given the excellent prognosis of low-volume disease patients regardless of exposure to Doce, we individualize management with escalation or de-escalation of therapy based on the 7-month PSA response and we aim to incorporate local treatment modalities to either the primary tumor or oligometastatic bone sites. We also account for the unique clinical subset of non-measurable, low-volume mCSPC which we consider a lower risk state, previously categorized as biochemical recurrence, and for whom we frequently recommend ADT + Bica with a plan for early escalation to a second generation ARSI if the 6-week PSA response is minimal. Proposed treatment schema for metastatic castration-sensitive prostate cancer. Abi, abiraterone; ADT, androgen deprivation therapy; ARSI, androgen receptor signaling inhibitor; Bica, bicalutamide; Daro, darolutamide; MDT, metastasis-directed therapy; RT, radiation therapy.
High-Volume M1b Castration-Sensitive Prostate Cancer
Patients who present with high-volume disease have proven OS benefit from the addition of cytotoxic chemotherapy to ADT.8,25,26 Despite increasingly robust data showing greater benefit for triplet combinations of ADT + Doce + either Daro 27 or Abi 4 compared to ADT + Doce, it remains unclear if there is an advantage to adding Doce in the setting of upfront ARSI and whether all mCSPC with high-volume disease would benefit from such intensified treatment strategies. Several recent meta-analyses based on the published phase 3 trials in mCSPC did not show a statistically significant OS benefit with triplet therapy when compared to combined ADT + ARSI. 28 As such, ADT + ARSI remains a reasonable option and subsequent treatment can be de-escalated for patients who achieve undetectable PSA after 7-8 months of treatment based on the excellent prognosis for this subgroup observed in the PEACE-1 trial. 16
To the extent that subgroup analyses can be informative, a pattern has emerged from all triplet therapy trials suggesting that patients most likely to benefit from intensified treatment may be those with de novo or synchronous high-volume metastatic disease. Both ARASENS and ENZAMET which enrolled primarily synchronous mCSPC patients found no benefit in the subgroups with metachronous disease (HR for death 0.61 [95% CI, 0.35-1.05] for ARASENS and HR for death 1.10 [95% CI, 0.65-1.86] for ENZAMET). By contrast, PEACE-1 selected only for patients with synchronous mCSPC. In light of available data, it is reasonable to treat most patients with metachronous high-volume mCSPC with doublet therapy (ADT + Abi/Enza/Apa or ADT + Doce). In our practice, we reserve triplet combinations for those with symptomatic (bone pain) de novo synchronous high-volume mCSPC. It is worth noting that, as with efficacy, the additive toxicity of Doce to ADT + ARSI doublets is unknown, as these combinations have not been directly compared in clinical trials. In PEACE-1, the addition of Abi to ADT + Doce let to an 11% increase in the incidence of severe (grade ≥3) adverse events, primarily in the form of hypertension (22% vs 13%) and hepatotoxicity (6% vs 1%). In the PEACE-1 subgroup not receiving Doce, Abi resulted in a 25% increase in severe adverse events, again primarily hypertension (29% vs 16%) and hepatotoxicity (6% vs 1%). 4 By contrast, Daro added to ADT + Doce did not result in excess toxicity in the ARASENS trial. 5
M1c Castration-Sensitive Prostate Cancer with Liver Metastases
CASE 2: A 51-year-old man presented with rectal bleeding, hematuria and pelvic pain. Digital rectal examination was positive for dark stool. An enlarged prostate with a lateral nodule was also appreciated. PSA was elevated at 15.9 ng/mL. Colonoscopy identified a low-lying rectal mass as the likely source of the bleeding and pathologic examination was consistent with an invasive adenocarcinoma of presumed colorectal origin. Immunohistochemistry revealed loss of MSH2 and equivocal staining for MSH6. Molecular profiling showed high microsatellite instability, high tumor mutational burden (10 mutations/Mb), MSH2 loss of exons 7-16, mutations in STK11 and TP53, as well as RB1 loss of exons 3-17. Despite a concerning family history (mother with ovarian cancer, sister with BRCA1-associated breast cancer), germline genetic testing was unrevealing for pathogenic variants. 18Fluorodeoxyglucose PET/CT scan showed a metabolic focus in the lower rectal region (SUV 6.8) suggesting a primary rectal neoplasm and noted an infiltrative process inferior within the prostate gland (SUV 5.3), as well as diffuse hepatic metastases up to 4 cm in size (maximum SUV 6) and osseous metastases in the left sacrum, right inferior pubic ramus, left acetabulum. Pelvic MRI confirmed the presence of a sessile rectal wall mass consistent with neoplasm with an adjacent irregular enhancing mass in the perirectal area involving the prostate gland and bladder wall likely to represent regional neoplastic metastatic disease from rectal carcinoma or prostate carcinoma. Distal bilateral ureter dilation raised concern for obstruction at the level of the bladder. A subsequent diagnostic and therapeutic transurethral resection of bladder and prostate biopsy was consistent with prostate adenocarcinoma, Gleason 8 (4 + 4). In the interim, the patient began combination chemotherapy with 5-fluorouracil, leucovorin and oxaliplatin for presumed metastatic colorectal cancer. What is the next best step in management?
The patient in CASE 2 had an equivocal presentation with a pelvic mass involving the rectum and the prostate, as well as osseous and hepatic metastatic disease. Although the PSA is abnormal suggestive of prostate adenocarcinoma, the modest elevation is incongruous with the metastatic burden, as is the presence of metastases in the liver which would be a more typical presentation for a primary colorectal tumor. De novo aggressive variant prostate cancer, a rare scenario referring to androgen-insensitive PC, could present similarly. Of note, not all aggressive variant PCs exhibit distinct histologic features such as small cell neuroendocrine phenotype or express neuroendocrine markers (synaptophysin or chromogranin) on immunohistochemistry, but mounting evidence describes a specific molecular signature for these tumors irrespective of histology.29,30 The presence of combined aberrations in TP53, RB1 and/or PTEN are commonly seen in aggressive variant prostate cancer. 31 Although less common compared to colon cancer, DNA mismatch repair deficiency is seen in a small subset of prostate cancer. Based on the above clinical and genomic data, we recommend obtaining additional tissue, preferably from an accessible metastatic site most likely to yield a good specimen such as a liver metastasis.
CASE 2 continued: Considering the new information obtained from the bladder mass pathology, a liver biopsy was pursued which confirmed the diagnosis of metastatic prostate adenocarcinoma. Upon review of the original rectal mass specimen, the diagnosis was changed to adenocarcinoma of prostate origin based on positive staining for NKX3.1. PSA and synaptophysin were not expressed. Chemotherapy was stopped and ADT with degarelix was initiated. The patient subsequently enrolled on a clinical trial exploring sequential therapy with combined ADT (LHRH agonist and Apalutamide) followed by chemohormonal therapy (ADT + docetaxel x 4 cycles) [NCT05189457]. He achieved a partial radiographic response and PSA became undetectable. Per trial protocol, he had a repeat prostate biopsy which revealed treatment-related neuroendocrine prostate carcinoma (synaptophysin positive, NKX3.1 negative). Per protocol, he underwent 2 additional cycles of docetaxel plus the anti–PD-1 monoclonal antibody tislelizumab at 200 mg intravenous infusion once every 3 weeks for 6 doses along with continuous LHRH analog. At his latest follow up 22 months after initiation of prostate cancer-directed therapy, his PSA remains undetectable with complete response of the liver metastases.
High-burden or high-risk disease patients who present with liver metastasis are a distinct mPC subset with a particularly poor prognosis. None of the modern trials investigating ADT + ARSI doublets in mCSPC have found a significant OS benefit in patients with liver metastases (in contrast to those with pulmonary metastases) for Abi (HR 0.82 [95% CI 0.41-1.66]), 4 Enza (HR 1.16 [95% CI, 0.67-2.00]),13,32 or Apa (HR 0.76 [95% CI, 0.47-1.23]). 33 In this unique population with aggressive PC targeted exome sequencing with a CLIA-certified assay based on either prostate or liver biopsies should be performed as soon as possible. Due to the heterogeneity of Gleason pattern and tumor contents among different biopsy cores, sequencing based on liver biopsy is preferred. Enrollment in a clinical trial is strongly recommended for mCSPC with liver metastases. Outside clinical trial, triplet therapy with ADT + Doce + a second generation ARSI should be considered.
Low-Volume Metastatic Castration-Sensitive Prostate Cancer
CASE 3: An 83-year-old male who would like to avoid ADT established care for his recurrent PC. He was originally diagnosed with high-risk localized prostate cancer (pT3 pN0, Gleason score 9 [4 + 5]) eight years prior and treated with radical prostatectomy. His PSA decreased from the pretreatment value of 8.6 ng/mL to a postoperative nadir of 0.015 ng/mL, before increasing slowly over the next 3 years to a level of 0.22 ng/mL. He underwent salvage radiation therapy to the prostate bed with only a transient PSA response. At the time of initial evaluation eight years following his original surgery and 5 years after salvage radiation, his PSA was 2.88 ng/mL with a doubling time of >9 months. 18F-Fluciclovine PET/CT scan reported sub-centimeter right external iliac lymph nodes with a maximum SUV uptake of 1.9. A follow-up pelvic MRI did not reveal any enlarged or morphologically abnormal lymph nodes to suggest metastases. An 18F-sodium fluoride PET/CT was negative for osseous metastasis.
At the time of evaluation, despite a rising PSA, the patient has no clear evidence of metastatic disease on conventional imaging. As such, his recurrent prostate cancer is best defined as a biochemical recurrence. Although this is not a lethal stage of prostate cancer, his biochemically recurrent prostate cancer after salvage RT would not be considered curable. The patient was offered iADT with LHRH agonist which had been shown to be non-inferior to continuous ADT in terms of overall survival in the phase 3 PR7 trial. 34 Given he was against ADT, peripheral androgen blockade with Bica was also offered. Although not an option at the time of the discussion, the recent EMBARK trial demonstrated an improvement in metastasis-free survival with the use of combined ADT (LHRH + Enza) or Enza monotherapy compared to LHRH analog monotherapy in patients with high-risk biochemical recurrence defined as a PSA doubling time <9 months and a nadir post-radical prostatectomy of ≥1 ng/mL or post-radiation therapy of ≥2 ng/mL. 35 Of note, the patient in CASE 3 would not have met inclusion criteria for the EMBARK trial.
CASE 3 continued: In 2019 the patient pursued PSMA-radioguided lymphadenectomy in Germany with removal of 13 lymph nodes (6/13 were involved with metastatic prostate adenocarcinoma). The PSA decreased minimally to 2.88 ng/mL. A CT scan of the chest, abdomen, and pelvis obtained after surgery did not reveal any metastatic lesions. Repeat 18F-Fluciclovine PET/CT scan 2 months later in the setting of a rising PSA to 3.5 ng/mL identified an avid retroperitoneal node (SUV max 4.03). The patient continued to decline systemic therapy and opted for stereotactic body radiation (SBRT) to the retroperitoneal node at a dose of 35 Gy in 5 fractions in 2020 which resulted in no significant change in the PSA. Months later he underwent a second PSMA-radioguided excision of 4 cervical lymph nodes (4/4 involved by prostate adenocarcinoma), again with no PSA response. Next generation sequencing testing of the cervical lymph nodes yielded no actionable aberrations but revealed an SPOP F102 V mutation.
The PSMA PET scan has demonstrated superior sensitivity (85% [95% CI, 74-96]) vs 38% [95% CI, 24-52]) and specificity (98% [95% CI, 95-100] vs 91% [95% CI, 85-97]) in detecting metastatic prostate cancer compared to conventional imaging modalities such as computed tomography and bone scans. Its sensitivity to detect metastatic disease seems to be most improved for pelvic nodal sites (91% vs 59% [32% absolute difference; 95% CI, 28-35]). 36 It was not approved in the US for PC until December 2021. In the case described above, the patient’s PC was upstaged to stage IVA (pN1) after PSMA-radioguided lymphadenectomy in Germany.
The wide adoption of PSMA PET in the US in the past 2 years has led to the emergence of a unique subset of low-volume N1 or M1a mCSPC with demonstrable small nodal metastases which exhibit uptake of the PET tracer, but are not considered measurable by RECIST criteria given the diameters of their short axis are <1.5 cm on computed tomography. One of the FDA approved indications of PSMA PET is to stage biochemically recurrent prostate cancer, a state in which rising PSA levels point to the existence of disease, but which is radiographically invisible on conventional imaging. In this setting, PSMA PET reveals distant metastases in over half of such patients (55%), yet all phase III trials investigating systemic therapy in mCSPC predate this technology and how to best manage this unique subset has not been formally investigated.37,38 We would call these cases “non-measurable low-volume” metastatic PC to differentiate them from low-volume metastatic PC in phase III trials. Given that patients with low-volume metastatic disease derive an OS benefit from combined ADT in phase III trials, it would be reasonable to initiate ADT + ARSI in this subgroup with the option of treatment de-escalation based on 7-month PSA response and this is our preferred strategy in this subgroup. The Speckle-type POZ protein (SPOP) is a substrate-binding adaptor of the CULLIN3/RING-box1 E3 ubiquitin ligase complex which can target several other proteins for degradation. Frequently abnormal in PC, mutations lead to impaired binding and degradation of several target protein, including the androgen receptor, making SPOP-mutated cancers more sensitive to ADT + ARSI than to ADT + Doce.39,40
CASE 3 continued: The patient remained on surveillance for nearly 18 months in the setting of slow PSA progression and emergence of PSMA PET avid supraclavicular and intraabdominal lymph nodes on follow-up imaging. He opted to initiate peripheral androgen blockade with Bica when PSA increased to 13.6 ng/mL. After 3 months, the PSA had risen further to 17.2 ng/mL and treatment was switched to relugolix. Five months into relugolix treatment his PSA was 1.57 ng/mL and his latest PSA was 0.42 ng/mL 12 months into treatment with good tolerance.
Despite the SPOP mutation, he had suboptimal PSA response to the LHRH antagonist relugolix. Due to his concerns for side effects, he remains on relugolix monotherapy.
Low-volume mCSPC also includes oligometastatic CSPC which is commonly defined as 5 or fewer metastases. There are no proven upfront risk stratification strategies for low-volume mCSPC. Although subgroup analysis of the STAMPEDE trial reported prolonged OS by adding prostate RT to ADT for low-volume mCSPC, no significant benefit was demonstrated in the PEACE-1 trial with prostate irradiation.41,42 Of note, low-volume mCSPC with >3 bone metastases and a Gleason score of 8 or above would have met high-risk criteria in the LATITUDE trial and we would not recommend adding RT to the prostate or to asymptomatic metastatic lesions for high-risk low-volume mCSPC. One strategy in this scenario is initiation of ADT + ARSI with the option of refining risk assessment based on the PSA response with potential de-escalation to ADT monotherapy if PSA becomes ≤0.2 ng/mL within 7 months of combined ADT. Extrapolating from the STAMPEDE trial investigating the addition of 2 years of combined ADT with Abi and LHRH analog to RT of the prostate for N1 prostate cancer, a treatment break is sometimes offered to patients with oligometastatic prostate cancer. 43 As CASE 3 highlights, even with metastasis-directed therapy (MDT), oligometastatic CSPC remains an incurable disease. Intermittent ADT or more sophisticated adaptive therapy based on PSA and testosterone kinetics could offer noninferior disease control with less treatment related toxicities.
RT directed at distant disease sites and metastasectomy are also increasingly being explored in the oligometastatic setting. One phase II trial randomized asymptomatic metachronous oligometastatic CSPC (≤3 extracranial metastases, either N1 or M1, diagnosed on choline positron emission tomography–computed tomography) to either observation or MDT (SBRT or metastasectomy) and found superior median time to ADT initiation of 21 months in the MDT group (vs 13 months in the observation group). 44 ORIOLE, another phase II trial, randomized a similarly defined population of oligometastatic CSPC (≤3 extracranial metastases but diagnosed using the more sensitive PSMA-based PET radiotracer 18F-DCFPyL) to either SBRT or observation. At approximately 19-month median follow-up, the SBRT group did not reach median progression-free survival (PFS) based on a composite endpoint of clinical/biochemical/radiographic events, initiation of ADT, or death, compared to a median PFS of 5.8 months in the comparator arm. 45 The more recent EXTEND trial randomized patients with ≤5 metastases (mostly M1, CSPC and metachronous, although small numbers of N1, CRPC, and synchronous patients were included) after at least 2 months of ADT to either MDT + iADT or iADT alone. All patients underwent an induction period of 6 months of ADT before going on a mandatory break. 46 At a median follow-up of 22 months, median PFS had not been reached in those who received MDT + iADT compared to a median PFS of 15.8 months in those who received ADT alone. Furthermore, patients who received MDT were able to remain off ADT for significantly longer (median not reached) compared to those who received only the induction period of ADT (6.1 months; 95% CI, 3.7 months to not estimable).
Oligometastatic disease is similarly amenable to de-escalation strategies based on PSA response at 7 months. The EXTEND trial also supports the option of ADT break for patients who received both RT to the prostate bed and MDT with close surveillance for reinitiation of ADT when PSA increases above 1 ng/mL. 47
TxN1M0
For patients who qualify as stage IV CSPC based on metastatic spread to locoregional lymph nodes but who otherwise have no evidence of distant metastatic disease (TxN1M0) we recommend multimodality therapy including RT to the prostate and seminal vesicles and ADT for 3 years + Abi for 2 years. In the STAMPEDE trials, this time-limited approach improved both metastasis-free survival, as well as OS and PC-specific survival compared to ADT alone. 43
Ongoing de-escalation Trials in Metastatic Castration-Sensitive Prostate Cancer
Several trials evaluating treatment de-escalation based on landmark PSA response in mCSPC are currently ongoing. The A-DREAM trial (NCT05241860) is a phase II adaptive study employing stopping and starting rules based on PSA response in mCSPC treated with ADT + ARSI. Specifically, patients who achieve PSA <0.2 ng/mL after 18-24 months of ADT + ARSI undergo a treatment break with reinitiation of treatment in the event of PSA rise ≥5 ng/mL, radiographic, or clinical progression. The primary end point of A-DREAM is 18-month eugonadal treatment-free survival following treatment interruption.
EORTC-2238 GUCG (De-Escalate) is a randomized pragmatic trial investigating iADT in patients achieving a PSA ≤0.2 ng/mL after 6-12 months of ADT + ARSI. The study endpoints include OS at 3 years and the proportion of patients on iADT at 1 year, as well as health-related quality of life and health economics parameters.
The phase III LIBERTAS trial (NCT05884398) is exploring an intermediate de-escalation strategy in which patients with mCSPC having achieved a PSA nadir of ≤0.2 ng/mL within the first 7 months of combined ADT + Apa are randomized to iADT or cADT while continuing Apa monotherapy. The co-primary endpoints are radiographic PFS and the severity of the Adjusted Hot Flash Score at 18 months (OS is a key secondary endpoint).
One caveat to the above trial design is the long induction period which is not optimal to preserve prostate cancer cells sensitivity to ADT based on our mathematical modeling of tumor evolution under ADT.18,20,22,48 Supporting this observation, recent data from the SWOG 1216 trial indicate that mCSPC can be risk stratified based on the PSA response to combined ADT as early as 3 months into treatment. 49 The ideal way to de-escalate treatment would be designing a treatment schema based on individual’s PSA response over time. We are planning our second adaptive therapy trial in mCSPC with real time modeling and treatment decision making.
De-escalation trials described above will provide much needed evidence to support both the use of landmark PSA as a predictive biomarker in guiding therapy and for the efficacy of an adaptive approach that can decrease ADT utilization in PC patients with disease exquisitely responsive to hormonal treatment and, consequently, improve their quality of life and toxicity burden. Results will likely prove complementary to those derived from treatment intensification trials, as such high-risk patients are bound to continue to receive triplet regimens. Unanswered questions pertaining to the optimal selection of first line therapy (second generation ARSI + ADT and Doce + ADT) and of when to escalate treatment to include agents not previously used remain.
Limitations
Our narrative review draws on our focused practice of treating PC in addition to published data. By its nature, our review is subjective and must be viewed as complementary to systematic reviews which use meta-analytic methods to integrate available evidence in structured formats to objectively estimate comparative efficacy of treatments under study. 50 Both review approaches-narrative and systematic-have their limitations and merits and, when available, we present existing meta-analyses on discussed topics.51,52 We acknowledge the uncertainty in available data and make use of extensive experience, evidence, and extrapolation to aid clinicians in the shared decision making process with their mCSPC patients.
In analyzing PC trials, we have purposefully emphasized contemporary studies and have been selective about older ones. This choice was deliberate in an attempt to account for the ever-evolving landscape of highly effective therapies that have emerged in the last decade but may have also predisposed our efforts to selection bias. Furthermore, to provide a comprehensive picture of the data supporting the management of mCSPC, we have necessarily had to rely on cross-trial comparisons, an approach which has its own evident limitations such as heterogeneity among patient populations or definitions of endpoints.50,53
As most PCs have a prolonged natural history and generally affect older men with multiple comorbid conditions, many patients will never experience symptoms or die from their metastases. For this reason, defining benefit in mCSPC patients is often the greatest challenge. In focusing on 24-month OS we aimed to explore an endpoint that is objective, meaningful, and less likely to be affected by the competing risks of other causes of death or subsequent therapies. Additionally, 24-month OS has the added advantage of being presently mature and reported in most mCSPC trials which notoriously require long follow-up. Metrics such as PFS (clinical, biochemical or radiographic) are not valid surrogate end points for OS. 54 OS is, however, also an imperfect end point insofar as these other metrics have meaning for patients (PFS and quality of life among others) and has the inherent disadvantage of mapping out only short-term benefit. Our emphasis on risk-adapted therapy, both emerging paradigms and how we have incorporated available evidence into our practice to tailor treatment, is a response to the recognized uncertainty surrounding the concept of benefit in mPC.
Conclusion
The ever-expanding array of effective therapeutic strategies in metastatic castration-sensitive prostate cancer has markedly improved patient outcomes but has also increased the complexity of clinical decision making. Recent years have brought forth novel combinations of drugs with diverse and potentially synergistic mechanisms of action and non-overlapping toxicities, but concerns about overtreatment persist in the absence of precise risk stratification tools in a disease with a highly heterogenous course. Although clinicians are confronted with increasingly challenging questions of patient selection and treatment sequencing, it is worth recognizing that such debates are an expected by-product of progress. We describe our interpretation of current data to better risk stratify mCSPC while recognizing the need for refining existing paradigms. Future efforts are bound to focus on formally identifying patients who benefit most from treatment intensification or de-escalation using validated clinical variables or emerging molecular data.
Abbreviations
Androgen Deprivation Therapy
Abiraterone
apalutamide
bicalutamide
Continuous Androgen Deprivation Therapy
Darolutamide
Docetaxel
Enzalutamide
Intermittent Androgen Deprivation Therapy
luteinizing hormone-releasing hormone
Metastatic Castration-Sensitive Prostate Cancer
Metastasis-Directed Therapy
Overall Survival
Positron Emission Tomography
Prostate Cancer
Progression-Free Survival
Prostate-Specific Antigen
Prostate-Specific Membrane Antigen
Radiation Therapy
Footnotes
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article:The authors disclose honoraria from AstraZeneca, Bayer, Dendreon, and Sanofi for speaker fees or serving on advisory boards.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
