Alumkal Analyzes PARP Inhibition in Mutated Metastatic Castrate-Resistant Prostate Cancer

Joshi Alumkal, MD

Joshi Alumkal, MD, a leader, Prostate/Genitourinary, Medical Oncology Section, associate Division Chief for Basic Research, Hematology-Oncology Division, and professor, Department of Internal Medicine, Rogel Cancer Center at University of Michigan, assessed the available checkpoint inhibitors for the treatment of mutated metastatic castration-resistant prostate cancer.

Targeted Oncology^TM: How does the identification of mutations in homologous recombination repair (HRR) genes affect expected response to treatment?

ALUMKAL: In regard to checkpoint inhibition, the verdict remains out. We have clear data about PARP inhibitor therapy with patients who have CDK12 [mutations], although this is a small number [of patients in studies].

The PROfound study [NCT02987543] was published in the New England Journal of Medicine [last year], one in the spring, one in the fall. [The study featured] patients with mCRPC [metastatic castration-resistant prostate cancer] who progressed on a next-generation hormonal agent, [such as] abiraterone [Zytiga] or enzalutamide [Xtandi], and who had alterations in any of 15 [HRR] genes—BRCA1/2, ATM, CDK12, CHEK1, CHEK2, FANCL, PALB2, [BRIP1, BARD1, PPP2R2A, RAD51B, RAD51C, RAD51D, and RAD54L]. The mutations were ascertained by tumor mutational testing.

There are 2 cohorts: cohort A, which is focused on BRCA1/2, and ATM, and cohort B, [focused on] the other [12] alterations. This [grouping] is principally based upon smaller, earlier, single-arm studies demonstrating that patients with BRCA1/2, and ATM [alterations] appeared to have a greater benefit than the patients with those other HRR alterations. That’s why [investigators] lumped those patients together to be able to analyze the trial in cohort A alone versus the entire cohort as a whole to determine whether or not the drug was just effective in subgroups or the entire population.¹

Patients were randomized to either 300 mg twice daily of olaparib [Lynparza] or physician’s choice of the other androgen receptor [AR]–targeting agent that the patient hadn’t received. [After] the blinded independent central review of radiographics, patients were allowed to cross over to the olaparib [arm] if they were in the control arm.

The primary end point was radiographic progression-free survival [rPFS] in cohort A. This was based upon the RECIST [Response Evaluation Criteria in Solid Tumors] and Prostate Working Group 3 criteria with central review. The secondary end points included rPFS in both cohorts, looking at the group as a whole: response rate in cohort A, time to progression, and overall survival [OS].

An alteration in more than 1 of [the 15 HRR included] genes was found in 28% of samples; 778 samples of 2792 were tested. [PROfound] was a challenging study to pull off because [there] have [been] so many screening failures, but I think it was critical to understand more about how this drug may be working by prespecifying patients. I think even with these prespecifications, the results were modest in terms of [olaparib’s] benefit versus the control arm.

What was the efficacy for patients in the PROfound trial?

[For] the outcomes in cohort A, the median rPFS for patients treated with an AR inhibitor was about 3.6 months. It’s about double that for the PARP inhibitor olaparib [7.4 months]; hazard ratio, 0.34, significant P value [P < .001].

Looking at both cohorts A and B, the rPFS is lower when you factor in those 12 other alterations. So [the other alterations are] sort of diluting out the benefit; 5.8 months versus 3.5 months in the control group and a hazard ratio of 0.49 [P < .001]. As a whole, even with the cohort B patients, it’s still statistically significant.

The forest plot was not quite as dramatic as [other studies’]…but overall, the study [results] were positive in both cohorts A and B. For patients who had prior taxane use, there appears to be a greater benefit [with olaparib over AR inhibitors] than for those who had not had prior taxane use [HR, 0.39 vs 0.77]. If measurable disease was present, patients [with prior taxane use] tended to do better in terms of progression or death [HR, 0.41 vs 0.64], and several other criteria of interest. Visceral disease [also] tended to do better [HR, 0.42]. Patients with low performance status, particularly ECOG 1, tended to do better as well [HR, 0.45]. Age didn’t seem to matter. [As far as] regions, Europe and North America appeared to have a greater advantage, although in Asia, there was still a strong trend [HR, 0.67].

Looking at the specific gene alteration, BRCA2 looked like the winner, with a hazard ratio of 0.21, in terms of the defect most associated with benefit. In BRCA1 [alterations],the confidence interval crossed 1, and the hazard ratio was 0.41. For many of these other alterations, includingCDK12, which we think about possibly for immunotherapy benefit, [although] that remains to be proved as the trials are ongoing, there wasn’t a significant benefit with a PARP inhibitor in [the CDK12] subgroup of patients versus with an AR inhibitor [HR, 0.74]. But I think the challenge of these studies is that the other alterations were rarer than the BRCA1/2 and ATM alterations. So we’re talking about fairly limited numbers of patients with these other 12 alterations.

Looking at the OS in patients from cohort A, [it was] 19.1 months with olaparib versus 14.7 months with the control in the intention-to-treat [population]. In the crossover- adjusted analysis, the hazard ratio is even better [0.42 vs 0.69] than in the intention-to-treat population.²

Looking at OS by the subgroups, again in cohort A, all the patients appeared to [have] a benefit in [reducing] risk of death, [with a] hazard ratio of 0.69. But many other clinical covariates weren’t helpful to really pick out who was the ideal candidate in cohort A and who’s likely to have a survival benefit with a PARP inhibitor up front versus a second AR inhibitor.

Looking at the improvement and outcome for patients in the overall population, [there was a] 17.3 months OS for those treated with olaparib versus 14 months in patients treated with a second AR inhibitor, with a hazard ratio of 0.79. Again, if one doesn’t include the patients who crossed over [86 of 131; 66%], the hazard ratio looks even better for olaparib up front versus using an AR inhibitor up front [HR, 0.55].

In terms of the response rate in cohort A, it was very rare to see [a response with] the second AR inhibitor [2.3% response rate], but 33.3% of patients treated with olaparib had an objective response [odds ratio, 20.86; P < .0001].

What is the safety profile like with olaparib? Were there any adverse events (AEs) of concern?

[As far as] the AE profile, for patients on physician’s choice versus olaparib, the number of AEs grade 3 or higher was 50.8% in olaparib, 37.7% in physician’s choice, and dose reductions [were] more common to occur with olaparib versus the physician’s choice arm of the AR inhibitor [22.3% vs 3.8%].

Some of the most common AEs with olaparib were hematologic toxicities, particularly anemia [46.5%], nausea [41.4%], fatigue [41.0%], appetite changes [30.1%], and GI [gastrointestinal] AEs—diarrhea [21.1%], and vomiting [18.4%]. [Among AEs for] the physician’s choice group, there was some percentage of patients with anemia [15.4%], nausea [19.2%], and fatigue [32.3%]. There was a greater risk of pulmonary embolisms in patients treated with olaparib versus physician’s choice [4.3% vs 0.8%, respectively], but none of these proved to be fatal, fortunately.

We worry about the risks of DNA damaging agents like PARP inhibitors [in the] long term. For patients with advanced mCRPC who had already received an AR-targeting agent, it’s possible they weren’t treated long enough to see the effects of long-lasting treatment with a PARP inhibitor on bone marrow toxicity, myelodysplasia, and\ leukemia. But that’s something that we’ll, hopefully, learn more about, particularly as clinical trials are moving these agents to earlier phases of the disease. That will be something we’ll really have to look out for as a field.

How is olaparib used in the clinic?

The FDA approved olaparib for patients who have homologous recombination DNA repair mutations—BRCA1/2, ATM, or any of the 12 other genes.³

There’s a companion diagnostic with Foundation Medicine, based upon tumor tissue testing. The FDA also approved a blood-based FoundationOne assay that allows one to ascertain BRCA1/2 or ATM alterations based upon cell-free DNA sequencing. And so that’s another test that’s available to determine these mutations.

What other studies have looked at the use of HRR genes in patients with mCRPC?

There was another trial published last year, the TRITON2 study [NCT02952534]. There were 2 publications [as with the PROfound study]. This study focused on patients who had [any of among 15] HRR gene [alterations]. It was a singlearm study, [with key criteria of] mCRPC, and deleterious somatic or germline alteration in any of those genes, and they progressed despite an AR-targeting agent and 1 prior taxane chemotherapy for CRPC. The vast majority of [these patients] received docetaxel and an AR-targeting agent. [The trial enrolled] patients with good performance status who had not received platinum chemotherapy or a PARP inhibitor and treated [them] with rucaparib [Rubraca] at 600 mg twice daily. The primary end point was confirmed objective response rate [ORR] by RECIST/Prostate Cancer Working Group 3 [criteria], and [the investigators] also evaluated PSA response rates for those that didn’t have measurable disease.⁴

Patients were enrolled, 1700 patients were screened to get 209 [eligible patients], 115 patients had deleterious BRCA1/2 alterations, with BRCA1 being much less common than BRCA2 [11% vs 89%]. For both BRCA1 and BRCA2, the majority were somatic rather than germline. In many cases, it wasn’t known whether both alleles were lost, particularly in those patients who might have just had germline testing.

The results [of the study] were a 43.5% ORR, and the PSA response rate was 54.8%. The CRs [complete responses were rare, 6.2% in the investigator-evaluable population, and 44.6% were PRs [partial responses]; in the central review evaluable population, 11.3% CRs versus 32.3% PRs. The majority of patients had stable disease in both assessments.⁵

I wouldn’t be surprised if [patients with a CR] were those with biallelic BRCA1/2 [mutations] that are really susceptible to PARP inhibition. When you add up the numbers, particularly in this cohort where [patients] received both chemotherapy and an AR-signaling inhibitor, as there really aren’t a lot of good options for these patients, [CRs are]= quite remarkable. This is limited to the BRCA1/2 [population of this trial], and those appear to be the patients who are most likely to respond to these drugs, and many of these patients are germline, so a lot of them are probably biallelic.

The rPFS was 9 months by the central review, and the time to PSA progression was 6.5 months. The OS data were not mature. The 12-month OS rate was estimated to be 73% in patients who were fairly far along who received several life-extending therapies. Again, this is the BRCA1/2 alteration group.

Looking at the non-BRCA1/2 alteration group of patients is much less dramatic than the BRCA1/2 [group], which I think had an impact on how this drug was ultimately approved by the FDA. [In this group], you’re not really seeing any CRs in the ATM, CDK12, or CHEK2[-mutated population]. In the “other mutation” group [FANCA, NBN, BRIP1, PALB2, RAD51, RAD51B, and RAD54L mutations], 7.1% did have CRs, and the PR rate was 10.5% in patients with ATM [mutations].⁶ [Responses weren’t] really seen in the CDK12[-mutated patients], 11.1% in the CHEK2[- mutated patients], and 21.4% in the “other” group.

Then stable disease, across the board, [was in the] 50% to 60% range in patients in these various mutational groups. In terms of the clinical benefit rate, it ranged from 20% on the low end with the CDK12 mutations to 54% with the “other” group. At 12 months, the clinical benefit rate, particularly in this “other” group, which is a hodgepodge of mutations, a small number of patients, only 14 [(37.5%) saw a] clinical benefit rate. PSA responses appeared to be highest in this “other” group versus these alterations—ATM, CDK12, CHEK2; similar story with the PSA progression timeframe, which was best in the other group.

The median treatment duration across all comers was 3.7 months. [For] patients with BRCA1/2 alterations, the median duration of treatment was 4.4 months. [As far as] treatment-related AEs, at least 1 [event was seen] in about 95% of patients; at least 1 AE greater than 3 [in severity] was seen in about 53% of patients. A good number of patients had to have treatment interruptions due to the therapy.

The most common AEs, similar to olaparib, were fatigue/asthenia [44.7%], nausea [42.4%], change in appetite [28.2%], and anemia [28.2%]. GI AEs were the principal ones that we see with these PARP inhibitors, so it appears to be a class effect.

How does the indication for rucaparib differ from that of olaparib?

Based upon the results of [the TRITON2] study, particularly the results in the BRCA1/2 alteration population, rucaparib was FDA approved [in 2020] for the treatment of patients who have [disease progression] on an AR inhibitor and taxane-based chemotherapy.⁷ The olaparib approval did not require taxane chemotherapy and allowed any of the homologous recombination DNA repair alterations, based upon the overall study [results] being positive. So, in addition to determining this by germline or mutational testing, the FDA also approved the FoundationOne liquid assay to detect BRCA1/2 mutations for rucaparib use, in addition to olaparib use.

One thing to be aware of is that not everything that one sequences from blood is from the tumor. Recently, [a study] from the University of Washington has demonstrated that many patients, particularly older patients with prostate cancer, will have alterations in their cell-free DNA driven by CHIP [clonal hematopoiesis of indeterminate potential] clones or clonal hematopoiesis.⁸ So [the investigators] ascertained that by sequencing both the [paired cell-free DNA] from these patients and normal white blood cell [control samples] in addition to the germline, some of the mutations present in the cell-free DNA were also present in the normal white blood cells but not the germline. So that signal in these liquid biopsy assays—and this was a homegrown assay at the University of Washington—was being confounded by CHIP clones that may be a precursor for leukemias and may be associated with the risk of cardiovascular disease and other cancers. It’s not routine for companies like Guardant Health or FoundationOne to do sequencing of the white blood cells in germline, so you can be confident that the signal that’s present in the cell-free DNA testing is from a tumor.

One of the most frequently mutated genes in the University of Washington study published in JAMA Oncology in October— and there’s a nice commentary that Zachery R. Reichert, MD, PhD, and I wrote in the same issue—was ATM, which was 1 of the genes on the approval [list] for olaparib.⁹ Those patients appear to have less of a benefit than the patients with BRCA1/2 [alterations] with olaparib. But the FoundationOne liquid biopsy testing allows one to look for [gene] mutations, and some of [them], particularly ATM, may be driven by CHIP clones rather than the tumor DNA. [That’s] just something to be aware of if using liquid biopsies to ascertain these mutations. It may be somewhat of a moot point. We know the [patients with] ATM [alterations] don’t appear to have the bang-up responses with either rucaparib or olaparib like the [patients with] BRCA1/2 [alterations do]. Even if I do somatic or germline testing and find an ATM alteration, I’m still a little bit perplexed about whether I should offer that therapy to patients, particularly if there’s another life-extending option that they could have, based upon the subgroup analysis we’ve seen from both the PROfound and TRITON2 studies.

What are your thoughts on expanded biomarker testing of HRR gene testing?

I think it will be really interesting to see....Certainly, I think these data would suggest that in patients with BRCA1/2 mutations, particularly [the subgroup] with a germline alteration where there’s a good likelihood that they may have biallelic inactivation, these drugs look most promising, whether it’s rucaparib or olaparib.

_References:

_{1. de Bono J, Mateo J, Fizazi K, et al. Olaparib for metastatic castration-resistant prostate cancer. N Engl J Med. 2020;382(22):2091-2102. doi:10.1056/NEJMoa1911440}

_{2. Hussain M, Mateo J, Fizazi K, et al; PROfound trial investigators. Survival with olaparib in metastatic castration-resistant prostate cancer. N Engl J Med. 2020;383(24):2345-2357. doi:10.1056/NEJMoa2022485}

_{3. FDA approves olaparib for HRR gene-mutated metastatic castration-resistant prostate cancer. News release. FDA. May 19, 2020. Accessed March 5, 2021. https://bit.ly/3jIugw6}

_{4. Abida W, Bryce AH, Vogelzang NJ, et al. Preliminary results from TRITON2: a phase II study of rucaparib in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC) associated with homologous recombination repair (HRR) gene alterations. Ann Oncol. 2018;29(suppl 8):viii271-viii302. doi:10.1093/annonc/mdy284.002}

_{5. Abida W, Patnaik A, Campbell D, et al; TRITON2 investigators. Rucaparib in men with metastatic castration-resistant prostate cancer harboring a BRCA1 or BRCA2 gene alteration. J Clin Oncol. 2020;38(32):3763-3772. doi:10.1200/JCO.20.01035}

_{6. Abida W, Campbell D, Patnaik A, et al. Non-BRCA DNA damage repair gene alterations and response to the PARP inhibitor rucaparib in metastatic castrationresistant prostate cancer: analysis from the phase II TRITON2 study. Clin Cancer Res. 2020;26(11):2487-2496. doi:10.1158/1078-0432.CCR-20-0394}

_{7. FDA grants accelerated approval to rucaparib for BRCA-mutated metastatic castration-resistant prostate cancer. FDA. May 15, 2020. Accessed March 5, 2021. https://bit.ly/3qgIh7v}

_{8. Jensen K, Konnick EQ, Schweizer MT, et al. Association of clonal hematopoiesis in DNA repair genes with prostate cancer plasma cell-free DNA testing interference. JAMA Oncol. 2021;7(1):107-110. doi:10.1001/jamaoncol.2020.5161}

_{9. Reichert ZR, Jones MA, Alumkal JJ. A CHIP in the armor of cell-free DNA–based predictive biomarkers for prostate cancer. JAMA Oncol. 2021;7(1):111-112. doi:10.1001/jamaoncol.2020.5140}