Abstract
Purpose:
Poly-(ADP-ribose) polymerase inhibitors provide an effective maintenance option for patients with BRCA- or PALB2-mutated pancreatic cancer. However, mechanisms of PARPi resistance and optimal post-PARPi therapeutic strategies are poorly characterized.
Experimental Design:
We collected paired cfDNA samples and post-PARPi clinical data on 42 patients with advanced, platinum-sensitive pancreatic cancer who were treated with maintenance rucaparib on NCT03140670, of whom 32 developed progressive disease.
Results:
Peripherally detected, acquired BRCA or PALB2 reversion variants were uncommon (5/30; 16.6%) in patients who progressed on rucaparib. Reversions were significantly associated with rapid resistance to PARPi treatment (mPFS 3.7mo vs 12.5mo, p=0.001) and poor overall survival (mOS 6.2mo vs 23.0mo, p<0.0001). All patients with reversions received re-challenge with platinum-based chemotherapy following PARPi progression and experienced faster progression on this therapy than those without reversion variants (rwTTD 2.4mo vs 5.8mo, p = 0.004). Of the patients who progressed on PARPi and received further chemotherapy, the OS from initiation of second line therapy was significantly lower in those with reversion variants than in those without (5.5mo vs 12.0 mo, p = 0.002). Finally, high levels tumor shedding were independently associated with poor outcomes in patients who received rucaparib.
Conclusions:
Acquired reversion variants were uncommon but detrimental in a population of patients with advanced BRCA or PALB2 related PDAC who received maintenance rucaparib. Reversion variants led to rapid progression on PARPi, rapid failure of subsequent platinum-based treatment and poor overall survival of patients. The identification of such variants in the blood may have both predictive and prognostic value.
Introduction
Maintenance poly-(ADP-ribose) polymerase inhibitors (PARPi) extend progression-free survival (PFS) and improve quality of life in patients with advanced platinum-sensitive pancreatic cancer harboring pathogenic germline or somatic variants in BRCA1, BRCA2 or PALB2(1,2). However, responses to PARPi in this clinical setting are highly varied, with some tumors demonstrating primary resistance, while others remain responsive to PARPi for months or even years(1,2). At present, the fundamental mechanisms of primary and secondary PARPi resistance are poorly characterized, as is the activity of further chemotherapy in a population of patients with BRCA or PALB2-related pancreas cancer who have progressed on PARPi.
We previously published the results of a single-arm phase II clinical trial testing the efficacy of maintenance rucaparib in patients with platinum-sensitive advanced pancreatic cancer and with germline or somatic pathogenic variants in BRCA1, BRCA2, or PALB2 (NCT03140670)(2). As part of this clinical trial, paired circulating cell-free (cf)DNA samples, at enrollment (post-platinum but pre-PARPi baseline) and at cancer progression on rucaparib, were collected. Clinical outcomes, including responses to post-rucaparib treatment regimens, were tabulated. Here, we report the clinical and cfDNA results of this population.
Materials and Methods
Patients
All patients were treated on NCT03140670, for which the full eligibility criteria has been previously described(2). Briefly, eligible patients had locally advanced or metastatic pancreatic cancer and a germline or somatic pathogenic variant in BRCA1, BRCA2, or PALB2 as assessed by a Clinical Laboratory Improvement Amendments- (CLIA-)certified laboratory. Prior to trial enrollment, patients had received ≥16 weeks of platinum-based combination chemotherapy for advanced disease without evidence of platinum resistance (defined as: growing tumors, new tumors, or increasing tumor marker within eight weeks of platinum exposure). Patients were eligible to enroll having received <16 weeks of platinum-based chemotherapy if there was intolerance or allergy, at the discretion of the principal investigator. Upon study enrollment, chemotherapy was discontinued and patients were treated with rucaparib 600 mg per os (PO) twice daily, days 1–28 of 28-day cycles until unacceptable toxicity or radiographic progression by Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 occurred. Upon progression, patients received physician-choice chemotherapy as per standard of care.
Clinical Data
Patients were eligible for this analysis if they received treatment with rucaparib on NCT03140670. For those who progressed on rucaparib, information on post-rucaparib treatment regimen(s) and clinical outcomes were obtained and tabulated. Correlation of baseline covariates and progression covariates was assessed with linear regression, Fisher’s exact test, ANOVA, or Pearson’s correlation coefficient, as appropriate, with significance set at p≤0.05. The primary outcome of interest was real-world time-to-treatment discontinuation (rwTTD), a surrogate in real-world databases that approximates progression-free survival seen in clinical trials(3). This outcome was defined as the time from initiation of post-rucaparib chemotherapy until treatment discontinuation (or no treatment for ≥120 days), therapy change, documented progression, or death, with patients censored at last clinical contact.
Secondary endpoints included overall survival (OS) from time of initiation of second-line chemotherapy to death, and OS2, defined as the time from study enrollment until death. Finally, we assessed real-world response rate (rwRR), which has been previously validated as approximating objective response rates in clinical trials(4). For this endpoint, responses were determined by review of radiographic reports indicating reduction of disease burden (partial response - rwPR), stable disease (stable disease- rwSD), or increase in existing tumor size or presence of new metastatic disease (progressive disease- rwPD) confirmed with clinical documentation(4). Time-to-event analysis was performed with the Kaplan-Meier method. All statistical analyses were performed in Stata version 17.0 (College Station, TX). Given the limited sample size in this post-progression analysis, only descriptive statistics were performed without formal comparisons in this group.
cfDNA Collection and Analysis
Peripheral blood samples for plasma cfDNA were collected at study enrollment (following platinum-based therapy but prior to rucaparib treatment) and upon disease progression for patients enrolled on NCT03140670. Samples were analyzed with the GuardantOMNI® 500-gene liquid biopsy platform(5). When available, somatic next-generation sequencing results were collected and tabulated. We specifically evaluated for the presence or absence of reversion mutations in BRCA1, BRCA2, and PALB2, defined as base substitutions that change a nonsense mutation to missense or synonymous, or an insertion or deletion that is predicted to restore the open reading frame of the gene in cis with a co-occurring mutation(6,7).
Detection of circulating KRAS variants was determined at baseline via cfDNA (described above) and somatic clinical next-generation sequencing of the tumor when available. Concordance for KRAS mutations between the specifically altered alleles was assessed between tumor tissue and cfDNA.
Tumor shedding was also assessed by examining the cfDNA somatic variant allele fraction as a proxy for tumor content in each sample. This was calculated as percentage of mutated allele detected in peripheral blood plasma per sample, excluding synonymous variants and adjusted for variants occurring in amplified genes, excluding variants likely to be associated with clonal hematopoiesis based on published databases. Correlations with outcomes were assessed with the maximum VAF, minimum VAF, median VAF, mean VAF, and variant count (calculated for the non-synonymous somatic variants excluding CNVs)
Associations between cfDNA results and time-to-event outcomes (OS and rwTTD) were performed using a log-rank test with index date of enrollment onto NCT03140670.
Circulating KRAS Variant Detection Analysis and 12-month Overall Survival ROC Analysis of VAF metrics
The association of baseline, peripherally detected KRAS and survival was further evaluated using Kaplan-Meier survival estimation and the log-rank test for baseline KRAS mutation detection within the entire cohort of patients with cfDNA (n=41 or n=39 with metastatic disease on enrollment) with an index date of enrollment. To remove potential confounding of KRAS wild-type patients (known to have longer survival) baseline peripheral blood detection of KRAS was also evaluated for those with tissue testing positive for KRAS mutation (N= 17 or n=16 with metastatic disease) and those with mutant KRAS detected in any sample at any time point (N=25 or n=24 with metastatic disease).
Baseline KRAS VAF, maximum VAF, minimum VAF, median VAF, mean VAF, and variant count (calculated for the non-synonymous somatic variants excluding CNVs) were evaluated by non-parametric receiver operator curve (ROC) for overall survival of 12-months or greater. This evaluation was performed for the full cohort (N=41 metastatic or 39 locally advanced, 9 death events within 12 months) and the cohort of patients with mutant KRAS detection in any sample at any time point (N=25 metastatic or n=24 locally advanced, 9 death events within 12 months) without correction for multiple testing. Area under the curve (AUC) and its 95% confidence interval were calculated as well as the Youden optimal cut point with sensitivity, specificity, and AUC at the that cut point.
Oversight
This study was conducted at the Abramson Cancer Center at the University of Pennsylvania and was approved by the Institutional Review Board at the University of Pennsylvania. This study was performed in accordance with the Declaration of Helsinki. All patients provided written informed consent as a condition for participation in the clinical trial. The data generated in this study are available upon request from the corresponding author.
Results
Patients
A total of 42 patients were treated with rucaparib on NCT03140670. At enrollment, 40 patients had metastatic disease and two had locally advanced, unresectable pancreas cancer. Eleven patients remained on study treatment at data cutoff (December 10, 2021), including one patient who remained on rucaparib for clinical benefit following local progression. Of the remaining 31 patients who discontinued rucaparib due to disease progression, seven patients did not receive further systemic therapy and one patient received non-cytotoxic targeted therapy. The 19 patients who remained on rucaparib or did not receive post-progression chemotherapy were not included in the post-progression outcome analysis (see Figure 1). Representativeness of all study participants is summarized in Supplementary Table S1.
Figure 1.
CONSORT diagram. 42 patients enrolled on NCT03140670, the phase II trial of rucaparib in pancreatic cancer. Of these 23, had progressed and received post-progression systemic therapy. Five patients received post-progression treatment with non-platinum-based chemotherapy, 18 received post-progression platinum-based chemotherapy. Of the remaining 9 patients who progressed on NCT03140670, 7 did not receive post progression treatment, one patient remained on rucaparib for ongoing clinical benefit, and one patient received other targeted therapy. The patients who did not receive post-progression chemotherapy were included in the cfDNA analysis
Baseline and demographic data for all 42 patients who had enrolled on NCT03140670 are described in Table 1 by the type of post-progression therapy that was received. In a linear regression model, the time receiving rucaparib on clinical trial was weakly associated with time on post-progression chemotherapy (coefficient =0.35, 95% CI 0.03–0.67, p=0.03). No association was noted between time on rucaparib and type of post-progression chemotherapy (Pearson’s Chi2 p=0.45) (Figure 2). Detection of KRAS variants at baseline was negatively associated with RECIST category of best response to rucaparib (Fisher’s exact p=0.005). However, detection of KRAS variants at progression was not associated with subsequent real-world response to post-rucaparib chemotherapy (Fisher’s p=0.41). In an ANOVA, baseline cfDNA concentration was not associated with RECIST outcome on rucaparib (model p=0.78). By Pearson’s correlation, there was no association between baseline CA19–9 levels and baseline cfDNA concentration (coefficient 0.1167, p=0.47). Among those who progressed on rucaparib and received additional chemotherapy, there was no association between progression cfDNA concentration and best real-world response to second line chemotherapy by ANOVA (model overall p=0.33), nor was there a significant association between progression cfDNA concentration and CA19–9 by Pearson’s correlation coefficient (coefficient 0.3866, p=0.09). Additional patient-level data is presented in Supplemental Table S2.
Table 1.
Demographic data for all patients enrolled on NCT03140670 by post-progression therapy.
| Non-Platinum | Platinum | Other/None | Remains on Clinical Trial | Total | |
|---|---|---|---|---|---|
| N=5 | N=18 | N=9 | N=10 | N=42 | |
| Mean Age (SD) | 58.0 (8.7) | 57.4 (11.3) | 63.2 (8.4) | 66.5 (7.7) | 60.9 (10.1) |
| Sex | |||||
| F | 4 (80.0%) | 8 (44.4%) | 6 (66.7%) | 6 (60.0%) | 24 (57.1%) |
| M | 1 (20.0%) | 10 (55.6%) | 3 (33.3%) | 4 (40.0%) | 18 (42.9%) |
| Mutation | |||||
| Germline BRCA1 | 1 (20.0%) | 6 (33.3%) | 0 (0.0%) | 0 (0.0%) | 7 (16.7%) |
| Germline BRCA2 | 4 (80.0%) | 8 (44.4%) | 7 (77.8%) | 8 (80.0%) | 27 (64.3%) |
| Germline PALB2 | 0 (0.0%) | 3 (16.7%) | 2 (22.2%) | 1 (10.0%) | 6 (14.3%) |
| Somatic BRCA2 | 0 (0.0%) | 1 (5.6%) | 0 (0.0%) | 1 (10.0%) | 2 (4.8%) |
| Histology | |||||
| Acinar | 0 ( 0.0%) | 1 ( 5.6%) | 0 ( 0.0%) | 0 ( 0.0%) | 1 ( 2.4%) |
| Adeno | 5 (100.0%) | 16 (88.9%) | 9 (100.0%) | 9 (90.0%) | 39 (92.9%) |
| Mucinous | 0 ( 0.0%) | 0 ( 0.0%) | 0 ( 0.0%) | 1 (10.0%) | 1 ( 2.4%) |
| Squamous | 0 ( 0.0%) | 1 ( 5.6%) | 0 ( 0.0%) | 0 ( 0.0%) | 1 ( 2.4%) |
| Median Weeks of Platinum Prior to Rucaparib (IQR) | 26.1 (13.0–26.1) | 17.3 (17.3–21.7) | 17.3 (8.7–17.3) | 33.7 (17.3–39.1) | 17.3 (17.3–26.1) |
| cfDNA Concentration at Enrollment (ng/mL) | 4.2 (3.4–8.0) | 7.7 (4.7–12.7) | 6.8 (6.5–10.7) | 9.3 (5.0–18.8) | 7.7 (4.7–11.7) |
| cfDNA Concentration on Rucaparib Progression (ng/mL) | 5.7 (2.7–6.0) | 4.2 (2.9–15.2) | 5.6 (3.1–33.1) | 4.8 (2.8–15.2) | |
| Median CA19–9 on Enrollment (u/mL) (IQR) | 32.0 (22.0–36.0) | 111.0 (13.0–242.0) | 75.0 (25.0–132.0) | 91.5 (32.0–140.0) | 78.0 (25.0–140.0) |
| Median CA19–9 on Progression (u/mL) (IQR) | 80.0 (50.0–99.0) | 317.0 (35.0–5542.0) | 51.0 (29.0–308.0) | 66.0 (19.0–432.0) | |
| Circulating KRAS Detected on Enrollment | |||||
| Not Detected | 4 (80.0%) | 11 (61.1%) | 8 (88.9%) | 8 (80.0%) | 31 (73.8%) |
| Detected | 1 (20.0%) | 6 (33.3%) | 1 (11.1%) | 2 (20.0%) | 10 (23.8%) |
| Not Tested | 0 ( 0.0%) | 1 ( 5.6%) | 0 ( 0.0%) | 0 ( 0.0%) | 1 ( 2.4%) |
| Circulating KRAS Detected on Rucaparib Progression | |||||
| Not Detected | 3 (60.0%) | 2 (11.1%) | 5 (55.6%) | 10 (23.8%) | |
| Detected | 2 (40.0%) | 15 (83.3%) | 2 (22.2%) | 19 (45.2%) | |
| Not Tested | 0 ( 0.0%) | 1 ( 5.6%) | 2 (22.2%) | 3 ( 7.1%) | |
| No Progression | 0 ( 0.0%) | 0 ( 0.0%) | 0 ( 0.0%) | 10 (100.0%) | 10 (23.8%) |
| Best RECIST Response to Rucaparib | |||||
| Stable Disease | 2 (40.0%) | 9 (50.0%) | 4 (44.4%) | 5 (50.0%) | 20 (47.6%) |
| Partial Response | 2 (40.0%) | 3 (16.7%) | 4 (44.4%) | 3 (30.0%) | 12 (28.6%) |
| Complete Response | 0 (0.0%) | 0 (0.0%) | 1 (11.1%) | 2 (20.0%) | 3 (7.1%) |
| Progression | 1 (20.0%) | 6 (33.3%) | 0 (0.0%) | 0 (0.0%) | 7 (16.7%) |
| Second Line | |||||
| None | 0 (0.0%) | 0 (0.0%) | 6 (66.7%) | 6 (14.3%) | |
| FOLFIRI | 1 (20.0%) | 0 (0.0%) | 0 (0.0%) | 1 (2.4%) | |
| FOLFIRINOX | 0 (0.0%) | 2 (11.1%) | 0 (0.0%) | 2 (4.8%) | |
| Gemcitabine and Cisplatin | 0 (0.0%) | 9 (50.0%) | 0 (0.0%) | 9 (21.4%) | |
| FOLFOX | 0 (0.0%) | 5 (27.8%) | 0 (0.0%) | 5 (11.9%) | |
| Bevacizumab and Niraparib | 0 (0.0%) | 0 (0.0%) | 1 (11.1%) | 1 (2.4%) | |
| Gemcitabine and Nab-Paclitaxel | 3 (60.0%) | 0 (0.0%) | 0 (0.0%) | 3 (7.1%) | |
| FLOX | 0 (0.0%) | 1 (5.6%) | 0 (0.0%) | 1 (2.4%) | |
| CapeOx | 0 (0.0%) | 1 (5.6%) | 0 (0.0%) | 1 (2.4%) | |
| Remains on First-Line Rucaparib | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 10 (100.0%) | 10 (23.8%) |
| 5FU+Lip-Iri | 1 (20.0%) | 0 (0.0%) | 0 (0.0%) | 1 (2.4%) | |
| Radiation | 0 (0.0%) | 0 (0.0%) | 1 (11.1%) | 1 (2.4%) | |
| Progressed, Remains on Rucaparib | 0 (0.0%) | 0 (0.0%) | 1 (11.1%) | 1 (2.4%) | |
| Number of Additional Lines of Systemic Therapy After Second Line | 2.5 (1.0–3.0) | 1.0 (0.0–2.0) | 0.0 (0.0–0.0) | 1.0 (0.0–2.0) | |
| Best Real-World Second Line Response | |||||
| N/AA | 0 (0.0%) | 1 (5.6%)A | 1 (4.3%) | ||
| rwPD | 1 (20.0%) | 11 (61.1%) | 12 (52.2%) | ||
| rwPR | 2 (40.0%) | 4 (22.2%) | 6 (26.1%) | ||
| rwSD | 2 (40.0%) | 2 (11.1%) | 4 (17.4%) | ||
| Reversion Mutation Detected at Baseline or Progression | 0 (0.0%) | 5 (27.8%) | 2 (22.2%) | 0 (0.0%) | 7 (16.7%) |
Died prior to first tumor assessment on chemotherapy.
SD=standard deviation.
Figure 2.
Patient leel factors and outcomes on rucaparib and post-progression chemotherapy (if applicable). In order to enroll on NCT03140670, patients had to have at least stable disease while receiving ≥16 weeks of platinum-based chemotherapy or a medical exception (allergy or intolerance). We have previously demonstrated no relationship to length of pre-trial platinum exposure and outcomes on rucaparib. Here we also detail the specific qualifying mutation, somatic NGS and which panel was used, detection of circulating KRAS variants at baseline and progression, and presence of reversion mutations at baseline and progression. Swimmer’s plot details time on rucaparib and post-rucaparib chemotherapy, censored at last clinical contact prior to database lock. Patients are grouped by best RECIST response to rucaparib. (g=germline, s=somatic)
Clinical Outcomes
Post-rucaparib chemotherapy was selected at the discretion of the treating physician and was independent of the clinical trial. Of the 23 patients who received post-rucaparib systemic therapy, 18 (78.3%) received platinum-based regimens (n=9 oxaliplatin; n=9 cisplatin) and five (21.7%) received non-platinum-based regimens (n=3 gemcitabine; nab-paclitaxel and n=2 irinotecan-based regimens) (Table 1 and Figure 1). We assessed rwTTD This outcome was defined as the time from time of initiation of post-rucaparib chemotherapy until treatment discontinuation (or no treatment for ≥120 days), therapy change, documented progression, or death (Figure 3A)(3). Further, we assessed OS from time of initiation of second-line chemotherapy to death, and OS2, defined as the time from study enrollment on NCT03140670 until death. Patients who had not met the endpoints of interest were censored at date of last clinical contact.
Figure 3.
Outcomes by receipt of post-progression therapy. Twenty-three of the forty-two patients who enrolled on NCT03140670 progressed on rucaparib and received post-rucaparib chemotherapy, 18 (78.3%) received platinum-based regimens (nine oxaliplatin; nine cisplatin) and five (21.7%) received non-platinum-based regimens. Given the limited sample size, no statistical comparisons were made. A. Graphical representation of endpoints assessed in this study. Patients received at least 16 weeks of platinum therapy before enrolling on the clinical trial to receive maintenance rucaparib. We measured rwTTD (time from initiation of second-line chemotherapy to initiation of third line treatment or death), OS (survival from initiation of second-line chemotherapy to death or last follow-up), and OS2 (survival from enrollment on clinical trial to death or last follow-up). B. The median rwTTD across the entire cohort that received post-PARPi chemotherapy was 3.9 months (95% CI 2.4–5.9 months). The median rwTTD of patients who received post-PARPi platinum therapy was 3.7 months (95% CI 2.1–5.9 months) and rwTTD of patients who received post-PARPi non-platinum chemotherapy was 5.8 months (95% CI 1.0-NR). C. OS from receipt of second line therapy. For all 23 patients, the median OS after starting post-PARPi chemotherapy was 10.8 months (95% CI 6.2–12.8 months). For patients who received platinum, the median OS was 9.9 months (95% CI 4.4–11.5 months) compared to 18.7 months (95% CI 12.6-NR) for those who received non-platinum-based chemotherapy D. OS2 by receipt of post-progression platinum- or non-platinum-based chemotherapy. The median OS2 from trial enrollment was 18.3 months (95% CI 11.9–22.4 months) for the entire cohort. The median OS2 was 14.8 months (95% CI 9.3–20.7) for those who received post-progression platinum-based chemotherapy and 28.9 months (14.4-not reached) for patients who received non-platinum therapy.
For those who received platinum, the median rwTTD from initiation of post-progression chemotherapy was 3.7 months (95% CI 2.1–5.9 mo). For those who received non-platinum, the median rwTTD was 5.8 months (95% CI 1.0-NR) (Figure 3B). No patients received post-PARPi immunotherapy in the second line; one patient received CAR-T cell treatment in the third line on a clinical trial.
For all 23 patients who received post-progression chemotherapy, the median overall survival (OS) after starting chemotherapy was 10.8 months (95% CI 6.2–12.8 months) and the median OS2 (time from trial enrollment until death) was 18.3 months (95% CI 11.9–21.5 months). For patients who received platinum re-challenge after PARPi progression, the median OS from initiating chemotherapy was 9.9 months (95% CI 4.4–11.5 months) and OS2 was 14.8 months (95% CI 9.3–20.7). For patients who received non-platinum therapy, the median OS was 18.7 months (95% CI 12.6- not reached (NR)) and median OS2 was 28.9 months (14.4-NR) (Figure 3C and D).
We further assessed real-world response rates (rwRR), an outcome that approximates RECIST outcomes in clinical trials(4). Two of five patients (40%) who received non-platinum-based therapy experienced a real-world partial response (rwPR) and four of 18 patients (22.2%) treated with platinum-based chemotherapy experienced a rwPR. Two patients who received platinum-based therapy (11.1%) experienced real-world stable disease (rwSD) as the best response. The remainder of patients experienced real-world progressive disease (rwPD) as best response (three of five non-platinum, 12 of 18 platinum) (Table 2). One patient who received platinum died from progressive disease prior to first radiographic tumor assessment and is therefore tabulated as having rwPD.
Table 2.
Key outcomes by type of post-progression chemotherapy received. OS2 defined from study enrollment until death. OS2 and rwTTD data presented are medians with 95% confidence intervals. rwRR is a summary statistic with 95% confidence intervals. NR= not reached
| Comparison | OS2 (months) | rwTTD (months) | rwRR (%, 95% CI) | |
|---|---|---|---|---|
| Platinum vs Non-platinum | Platinum (n=18) | 14.8 (9.3–20.7) | 3.7 (2.1–5.9) | 22.2 (6.0–56.9) |
| Non-platinum (n=5) | 28.9 (14.4-NR) | 5.8 (1.0-NR) | 40.0 (4.8–100) | |
| Platinum Type | Cisplatin (n=9) | 13.8 (5.8–21.5) | 3.9 (1.2–7.4) | 11.1 (0.3–48.2) |
| Oxaliplatin (n=9) | 19.0 (5.0–23.0) | 3.7 (0.6–9.9) | 33.3 (7.5–70.0) |
cfDNA Analysis
Of the 42 patients enrolled on NCT03140670, 41 had cfDNA samples obtained at baseline (platinum-sensitive, pre-rucaparib), 36 of whom (88%) had detectable tumor cfDNA. Of the 31 patients who developed disease progression on rucaparib, 30 had cfDNA samples available for analysis at progression, all of whom (100%) had detectable cfDNA (Figure 2). Results from these assays were performed in batches after completion of the clinical trial and thus were not available to the post-progression treating physicians. Therefore, these results did not influence treatment decisions.
Of the 30 patients who progressed on rucaparib and had a cfDNA sample at progression, seven patients had a reversion mutation detected at any time point and five (16.6%) had acquired a newly detected reversion mutation that restored the open reading frame of the BRCA or PALB2 gene and thus restored function of the protein(7). Both patients with reversion mutations at time of trial enrollment experienced rapid progressive disease on PARPi. One of these patients had a reversion mutation detected only at baseline and one patient had an (identical) reversion mutation detected both at baseline and at disease progression. In the context of all patients who had progressive disease, reversions at any time point were associated with a significantly shorter OS (median OS 6.2 months (95% CI 1.3–11.3) vs 23.0 months (95% CI 19.0–28.9), log-rank p<0.0001) and PFS on rucaparib compared to those without reversions (median PFS 3.7 months (95% CI 0.9–4.1) vs 12.5 months (95% CI 5.3–16.0), log-rank p=0.001) (Figure 4A and 4B). Newly acquired reversion mutations were similarly associated with a significantly shorter OS (median 9.3 months (95% CI 4.6-NR) vs 23.0 months (95% CI 19.0–28.9), p<0.001) and PFS (median 3.8 months (95% CI 1.8-NR) vs 12.5 months (95% CI 5.3–16.0), p=0.009) from initiation of rucaparib compared to those who did not have reversions (Figure 4C and 4D). Specific patient-level data regarding patients with reversion mutations can be found in Supplemental Table 3.
Figure 4.
Reversion mutations were associated with particularly poor outcomes. Through the course of NCT03140670, seven patients were found to have a reversion at any timepoint, including two who were found to have reversions at baseline A. Of all patients who enrolled on NCT03140670 and experienced progression (n=32), reversions at any time point were associated with shorter median OS (6.2 months (95% CI 1.3–11.3) vs 23.0 months (95% CI 19.0–28.9), log-rank p<0.0001) and B. shorter median PFS on rucaparib (3.7 months (95% CI 0.9–4.1) vs 12.5 months (95% CI 5.3–16.0), log-rank p=0.001). C. Newly acquired reversion mutations were also associated with a shorter OS (mOS 9.3 months (95% CI 4.6-NR) vs 23.0 months (95% CI 19.0–28.9), p<0.0001) and D. PFS (mPFS 3.8 months (95% CI 1.8-NR) vs 12.5 months (95% CI 5.3–16.0), p=0.009) from initiation of rucaparib on NCT03140670 compared to those who did not have reversions. E. For patients treated with platinum in the post-PARPi setting, the median OS from initiation of post-rucaparib chemotherapy was 5.5 months (95% CI 4.4-NR) for the patients with a reversion mutation detected (n=4) and was 10.3 months (95% CI 3.2–14.4, log-rank p = 0.13) for those without a detected reversion mutation (n=14). F. The median rwTTD for the patients with a reversion mutation detected at progression treated on post-progression platinum-based chemotherapy was 2.1 months (95% CI 0.6-NR). The 14 patients who did not have reversion mutations detected at progression and received second line chemotherapy with a platinum-based regimen had a median rwTTD of 3.9 months (95% CI 1.7–8.4, log-rank p=0.03).
Patients with reversion mutations at any timepoint had significantly worse post-rucaparib outcomes compared to those who did not. In patients who received further therapy after progression on rucaparib, the median rwTTD and OS were 2.4 months (95% CI 0.6-NR) and 5.5 months (95% CI 3.2-NR), respectively, compared to 5.8 months (95% CI 2.6–9.9) and 12.0 months (95% CI 9.9–15.0) from time of initiation of post-progression therapy for those without reversion mutations (log-rank rwTTD p= 0.004, OS p= 0.002). Further, patients with reversion mutations, compared to those without, had worse outcomes when treated with post-rucaparib platinum-based regimens. Four patients with reversion variants detected at progression and fourteen patients without reversion variants were treated with platinum-based chemotherapy after progression on rucaparib. The median OS from initiation of post-rucaparib chemotherapy was 5.5 months (95% CI 4.4-NR) for the patients with a reversion mutation (n=4) and 10.3 months (95% CI 3.2–14.4, log-rank p = 0.13) for those without a detected reversion mutation (Figure 4E). The median rwTTD for those with reversion mutations was 2.1 months (95% CI 0.6-NR) compared to 3.9 months (95% CI 1.7–8.4, log-rank p=0.03) for the 14 patients who did not have reversion mutations detected at progression and who received second line platinum therapy, (Figure 4F). Finally, three patients experienced loss-of-heterozygosity (LOH) events, however only one of these patients had LOH in BRCA (this patient had LOH in BRCA1 and BRCA2). No patients experienced LOH in PALB2.
Among 21 of the 42 enrolled patients with available somatic tissue NGS, 17 (81%) had a mutant KRAS variant detected within the tumor tissue. Ten of the 36 patients (27.7%) with detectable baseline cfDNA had mutant KRAS variants detected while 26 (72.2%) did not. At disease progression, mutant KRAS was detected in 17 of 31 (53.1%) of the cfDNA samples. Patients with a peripherally detected KRAS variant at baseline (n=10) had a significantly shorter mPFS on rucaparib compared to those who did not (n=31): mPFS 1.7 months vs mPFS of 16.1 months (log-rank p=0.026). Similarly, those with a peripherally detectable KRAS variant at baseline had a significantly shorter median OS compared to those who did not have a detectable KRAS variant at baseline (9.1 months vs 29.0 months, p=0.014). On area under the receiver operatic curve (AUC/ROC) analysis of variant allele frequency (VAF) metrics of the binary outcome of 12-month OS in the entire cohort (n=41), KRAS VAF (AUC= 0.771, 95% CI 0.597–0.944), median VAF (AUC= 0.745, 95% CI 0.570–0.919), mean VAF (AUC=0.703, 95% CI 0.520–0.886), and variant count (AUC= 0.727, 95% CI 0.559–0.896) were associated with the outcome, whereas maximum VAF (AUC=0.679, 95% CI 0.490–0.868) and minimum VAF (AUC=0.663, 95% CI 0.486–0.850) were not. When enriched only for those with known mutated KRAS (either in cfDNA or somatic sequencing, n=25), KRAS VAF (AUC=0.708, 95% CI 0.512–0.905), maximum VAF (AUC=0.742, 95% CI 0.537–0.949), minimum VAF (AUC=0.746, 95% CI 0.555–0.938), median VAF (AUC=0.833, 95% CI 0.672–0.994), mean VAF (AUC=0.781, 95% CI 0.590–0.972) and variant count (AUC=0.760, 95% CI 0.567–0.954) were all associated with 12 month overall survival. This implies that tumor shedding (representing outgrowth of tumor while receiving active therapy), rather than mutation status alone, is likely a driver of worse outcomes for these patients. (Supplemental Figure S1A–D and Supplemental Table S4). Of note, both patients with locally advanced disease at enrollment were classified as having high-shedding tumors on VAF analysis.
Discussion
In this study, we describe the post-rucaparib treatment outcomes and cfDNA findings of a cohort of patients with advanced pancreatic cancer and germline or somatic BRCA1, BRCA2, or PALB2 variants who received rucaparib maintenance treatment on a previously published clinical trial(2). In this follow-up analysis of paired cfDNA samples and post-PARPi clinical data, we have identified that reversion mutations are uncommon in a population of patients who have progressed on PARPi, and that they associate with poor outcomes, both on rucaparib and on post-PARPi platinum-based chemotherapy.
Indeed, we observe that no patient (0/7) with a detectable reversion variant at either enrollment or progression responded to rucaparib for more than six months, while 10/35 (28.5%) of those without reversion mutations responded for over six months (odds ratio for failure at six months = 31.6, 95% CI 1.6–612.3, p=0.02). This highlights that reversion variants appear to be an early resistance event compared to other PARPi resistance mechanisms. There are reports in the literature, however, of patients with pancreatic cancer developing reversion mutations after more than a year on PARPi, highlighting that although likely uncommon as a late resistance mechanism, reversion variants in this setting can occur (8).
Despite the particularly poor outcomes, reversion mutations were relatively uncommon in the entire patient population, with only 16.6% of patients demonstrating a peripherally detectable acquired reversion mutation at time of progression on rucaparib. Patients who developed detectable reversion variants had rapid progression on rucaparib, while this was not universally the case for most patients without a reversion variant. This finding suggests that acquired resistance – and specifically delayed acquired resistance - to rucaparib in patients with BRCA- or PALB2- related pancreatic cancer may be largely dependent on other mechanisms or may not be readily detectable in the currently available peripheral blood assays. Multiple potential mechanisms of resistance have been previously described in the literature in addition to reversion mutations. Epithelial-mesenchymal transition, loss of SLFN11, overexpression of p-glycoprotein, loss of PARG (poly(ADP-ribose) glycohydrolase), mutation of PARP1, and upregulation of BRCA-independent end resection enzymes (e.g. loss of TP53BP1 or amplification of MRE11A) have all been noted in both preclinical models and reported in patients with PARPi resistance, yet the full clinical implications and incidence have yet to be realized(9–15). Functional testing of homologous recombination such as demonstration of RAD51 foci in tissue, genomic sequencing of tumor material, and the development and validation of other resistance mechanisms such as promoter methylation will be key in characterizing patients who may rapidly develop resistance to PARPi (13,16). The recently developed HRD-RNA platform may further identify additional patients with homologous recombination deficiency phenotype outside of BRCA loss using whole-exome RNA sequencing(17).
Additionally, we found that all patients with reversion mutations who received post-rucaparib platinum-based chemotherapy experienced rapid disease progression on this therapy. One might hypothesize that once functional BRCA protein production is restored via reversion mutation, platinum-sensitivity may be lost, a finding that has been demonstrated in other cancer types (18–23). The existing literature in other BRCA-related cancers has shown that reversion mutations detected prior to or at progression on PARPi are associated with shorter OS and PFS on PARPi in ovarian cancer(7). Patients with reversion variants in our study also had a significantly lower overall survival compared to their non-reversion counterparts, noting that all patients with reversion variants received platinum in the post-PARPi setting. Whether the patients with reversions would have had better outcomes with non-platinum-based regimens remains unknown and should be studied in a prospective manner.
Regardless of reversion mutation status, we also observed that those who received platinum re-challenge following PARPi progression did less well than those who received non-platinum-based treatment. Although the small sample size limits our ability to draw statistical conclusions, numerically, patients seemed to derive more benefit from non-platinum-based chemotherapies after progression on PARPi from an overall survival and time-to-treatment-discontinuation endpoints. This finding highlights the fact that our understanding of post-PARPi therapeutic efficacy in this population remains limited, and that prospective studies will be required to address this issue. Indeed, despite PARPi availability for the treatment of breast, ovarian, prostate and pancreatic cancer, data to guide post-PARPi progression therapies is limited mostly to retrospective series and post-hoc analyses of clinical trials, similar to the present analysis. These other analyses provide context for the present analysis, especially regarding the use of post-progression platinum chemotherapies. In ovarian cancer for example, a platinum-free interval of at least six months is associated with increasing likelihood of platinum sensitivity when rechallenged(24). Similarly, a retrospective study of ovarian cancer patients treated with PARPi followed by platinum chemotherapy found that patients who responded to post-PARPi platinum chemotherapy were likely to experience further benefit from PARPi rechallenge on progression(25). A post-hoc analysis of SOLO2 (maintenance olaparib vs placebo in platinum-sensitive BRCA-mutated ovarian cancer) specifically analyzing post-progression outcomes identified a reduced benefit with post-progression chemotherapy in the olaparib-treated group compared to the placebo-treated group, suggesting that prior exposure to (and progression on) PARPi decreases subsequent efficacy of platinum-based chemotherapies compared to placebo exposure(26). Although outcomes with platinum versus non-platinum-based chemotherapies were not directly compared in that study, platinum-based chemotherapies produced a longer time-to-second progression following progression on PARPi in those patients ovarian cancer (26). Finally, a real-world analysis of BRCA1- and BRCA2-mutated recurrent platinum-sensitive ovarian cancer treated with maintenance olaparib similarly showed diminishing activity of post-progression chemotherapy after PARPi(27).
Finally, we noted an association of peripherally detectable KRAS mutations with poor outcomes. However, in the ROC analysis, which used the binary outcome of 12-month overall survival, we also identified a significant association between tumor shedding (by VAF detection) and outcomes. Given the context that most patients in our cohort had known KRAS variants, we hypothesize that our observed association with peripherally detectable KRAS and poor outcomes was most likely a simple reflection of the fact that those patients with higher tumor burden have inferior outcomes on rucaparib. A concordant observation was made clinically (by using presence or absence of measurable disease on imaging at study start) in our previously published manuscript(2).
Importantly, this study has several limitations. Although patients contributed data to the clinical trial in a prospective manner and remained followed over time for clinical outcomes, the post-progression clinical data were collected in a retrospective fashion and post-progression therapy was determined by individual treating clinician off-protocol and independent of the clinical study. Although we assessed the objective endpoint of overall survival, we also analyzed real-world outcomes that have been previously validated as approximating RECIST outcomes in clinical trials(3,4). Still, the possibility exists that our outcomes and interpretation could differ if they were assessed in a more formal, protocolized manner. However, given that treatment decisions for post-progression therapy were made at the discretion of the treating physician, patients were not necessarily seen or assessed at the University of Pennsylvania and outcome assessment relied on the treating physician assessment as well as radiographic reports. Given the limited sample size, we did not perform statistical comparisons by post-progression treatment and results should be cautiously interpreted. Additionally, 10 patients remained on NCT03140670 at data cutoff and thus did not have progression cfDNA available for review. Lastly, in the present analysis we only describe variants detected by cfDNA. Paired tissue genomic analyses or RNA-based assays may reveal additional mechanisms of resistance. Notably, LOH was a rare event in this analysis but may not be adequately captured by current cfDNA assays. Similarly, circulating KRAS variants were detected in just over half (53%) of available samples although 81% of NGS samples had mutated KRAS. Altogether, this may be reflective of cfDNA assay limitations rather than absence of circulating mutated KRAS variants. Archival, baseline, and progression tissue genomic analyses which may provide additional insights are currently ongoing and will be described in a future report.
In conclusion, we describe outcomes of patients with innately platinum-sensitive, BRCA- or PALB2-related pancreas cancer that progressed on maintenance rucaparib in a clinical trial. We found that a minority of these patients had acquired reversion mutations detected in cfDNA, however reversion mutations were correlated with rapid progression on rucaparib as well as on post-rucaparib platinum-based therapy. Further, we identified that high levels of tumor shedding prior to PARPi therapy was associated with worse outcomes, which is consistent with our previously published observation that patients with higher volume disease on imaging at time of PARPi initiation had significantly more rapid progression. Last, we describe outcomes of patients treated with post-progression chemotherapy with either platinum-based or non-platinum-based chemotherapies, noting that platinum re-challenge was not necessarily the right clinical choice for most patients, particularly those with reversion variants. Larger prospective studies are needed to further clarify the predictive and prognostic implications of reversion mutations, and to broaden the understanding of other, perhaps more common, PARPi resistance mechanisms in patients with pancreatic cancer, so that we may learn to better select and sequence PARPi and subsequent therapies.
Supplementary Material
Statement of Translational Relevance:
There is little, if any, guidance surrounding treatment selection following progression on PARPi in BRCA and PALB2-related pancreas cancer. Furthermore, the prominent mechanisms and predictors of PARPi resistance remain poorly defined in this disease. We analyzed paired cfDNA samples and post-PARPi outcomes in a population of patients who were treated and progressed on maintenance rucaparib on a previously published clinical trial. We identify that peripherally detected reversion variants are associated with rapid progression on rucaparib, rapid progression on post-PARPi platinum-based therapy and poor overall survival. Additionally, we show that patients with higher levels of tumor shedding in the peripheral blood had inferior outcomes on maintenance therapy. To our knowledge, this is the first study describing post-PARPi treatment outcomes and the rates of reversion variants in this population.
Acknowledgements
Funding Acknowledgement: NIH T32CA009679 and NIH L30CA274783 (TJB); Clovis Oncology (KAR), Anonymous Foundation (KAR), Basser Young Leadership Council (KAR), The Konner Fund (KAR); The Philip and Pearl Basser Fund (KAR)
Footnotes
Conflict of Interest Disclosures:
AY, JY, and LAK are employees/shareholders of Guardant Health. Remaining authors declare no other potential conflicts of interest
Previous Presentation: Portions of this manuscript were presented at ASCO 2022, June 4, 2022 Chicago, Il and ASCO GI Cancers Symposium 2023, January 19-23, 2023, San Francisco, CA.
References
- 1.Golan T, Hammel P, Reni M, Van Cutsem E, Macarulla T, Hall MJ, et al. Maintenance Olaparib for Germline BRCA -Mutated Metastatic Pancreatic Cancer. N Engl J Med [Internet]. 2019. [cited 2022 Oct 12];381:317–27. Available from: 10.1056/NEJMoa1903387 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Reiss KA, Mick R, O’Hara MH, Teitelbaum U, Karasic TB, Schneider C, et al. Phase II Study of Maintenance Rucaparib in Patients With Platinum-Sensitive Advanced Pancreatic Cancer and a Pathogenic Germline or Somatic Variant in BRCA1, BRCA2, or PALB2. JCO [Internet]. 2021. [cited 2022 Mar 22];39:2497–505. Available from: 10.1200/JCO.21.00003 [DOI] [PubMed] [Google Scholar]
- 3.Stewart M, Norden AD, Dreyer N, Henk HJ, Abernethy AP, Chrischilles E, et al. An Exploratory Analysis of Real-World End Points for Assessing Outcomes Among Immunotherapy-Treated Patients With Advanced Non–Small-Cell Lung Cancer. JCO Clinical Cancer Informatics [Internet]. 2019. [cited 2022 Feb 14];1–15. Available from: 10.1200/CCI.18.00155 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ma X, Bellomo L, Magee K, Bennette CS, Tymejczyk O, Samant M, et al. Characterization of a Real-World Response Variable and Comparison with RECIST-Based Response Rates from Clinical Trials in Advanced NSCLC. Adv Ther [Internet]. 2021. [cited 2022 Oct 28];38:1843–59. Available from: 10.1007/s12325-021-01659-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Helman E, Artieri C, Vowles JV, Yen J, Nance T, Sikora M, et al. Abstract 5603: Analytical validation of a comprehensive 500-gene ctDNA panel designed for immuno-oncology and DNA damage research. Cancer Research [Internet]. 2018. [cited 2023 Jan 31];78:5603–5603. Available from: 10.1158/1538-7445.AM2018-5603 [DOI] [Google Scholar]
- 6.Sakai W, Swisher EM, Karlan BY, Agarwal MK, Higgins J, Friedman C, et al. Secondary mutations as a mechanism of cisplatin resistance in BRCA2-mutated cancers. Nature [Internet]. 2008. [cited 2022 Dec 19];451:1116–20. Available from: http://www.nature.com/articles/nature06633 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lin KK, Harrell MI, Oza AM, Oaknin A, Ray-Coquard I, Tinker AV, et al. BRCA Reversion Mutations in Circulating Tumor DNA Predict Primary and Acquired Resistance to the PARP Inhibitor Rucaparib in High-Grade Ovarian Carcinoma. Cancer Discovery [Internet]. 2019. [cited 2022 Nov 21];9:210–9. Available from: https://aacrjournals.org/cancerdiscovery/article/9/2/210/10756/BRCA-Reversion-Mutations-in-Circulating-Tumor-DNA [DOI] [PubMed] [Google Scholar]
- 8.Pishvaian MJ, Biankin AV, Bailey P, Chang DK, Laheru D, Wolfgang CL, et al. BRCA2 secondary mutation-mediated resistance to platinum and PARP inhibitor-based therapy in pancreatic cancer. Br J Cancer [Internet]. 2017. [cited 2023 May 21];116:1021–6. Available from: https://www.nature.com/articles/bjc201740 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Herzog M, Puddu F, Coates J, Geisler N, Forment JV, Jackson SP. Detection of functional protein domains by unbiased genome-wide forward genetic screening. Sci Rep [Internet]. 2018. [cited 2022 Dec 19];8:6161. Available from: http://www.nature.com/articles/s41598-018-24400-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pettitt SJ, Krastev DB, Brandsma I, Dréan A, Song F, Aleksandrov R, et al. Genome-wide and high-density CRISPR-Cas9 screens identify point mutations in PARP1 causing PARP inhibitor resistance. Nat Commun [Internet]. 2018. [cited 2022 Dec 19];9:1849. Available from: http://www.nature.com/articles/s41467-018-03917-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lheureux S, Oaknin A, Garg S, Bruce JP, Madariaga A, Dhani NC, et al. EVOLVE: A Multicenter Open-Label Single-Arm Clinical and Translational Phase II Trial of Cediranib Plus Olaparib for Ovarian Cancer after PARP Inhibition Progression. Clinical Cancer Research [Internet]. 2020. [cited 2022 Dec 19];26:4206–15. Available from: https://aacrjournals.org/clincancerres/article/26/16/4206/82777/EVOLVE-A-Multicenter-Open-Label-Single-Arm [DOI] [PubMed] [Google Scholar]
- 12.Ordonez LD, Hay T, McEwen R, Polanska UM, Hughes A, Delpuech O, et al. Rapid activation of epithelial-mesenchymal transition drives PARP inhibitor resistance in Brca2 -mutant mammary tumours. Oncotarget [Internet]. 2019. [cited 2022 Dec 19];10:2586–606. Available from: 10.18632/oncotarget.26830 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Waks AG, Cohen O, Kochupurakkal B, Kim D, Dunn CE, Buendia Buendia J, et al. Reversion and non-reversion mechanisms of resistance to PARP inhibitor or platinum chemotherapy in BRCA1/2-mutant metastatic breast cancer. Annals of Oncology [Internet]. 2020. [cited 2022 Dec 20];31:590–8. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0923753420360555 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hu K, Wu W, Li Y, Lin L, Chen D, Yan H, et al. Poly(ADP-ribosyl)ation of BRD 7 by PARP 1 confers resistance to DNA -damaging chemotherapeutic agents. EMBO Rep [Internet]. 2019. [cited 2022 Dec 20];20. Available from: 10.15252/embr.201846166 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gogola E, Duarte AA, De Ruiter JR, Wiegant WW, Schmid JA, De Bruijn R, et al. Selective Loss of PARG Restores PARylation and Counteracts PARP Inhibitor-Mediated Synthetic Lethality. Cancer Cell [Internet]. 2018. [cited 2023 May 30];33:1078–1093.e12. Available from: https://linkinghub.elsevier.com/retrieve/pii/S1535610818302228 [DOI] [PubMed] [Google Scholar]
- 16.Prados-Carvajal R, Irving E, Lukashchuk N, Forment JV. Preventing and Overcoming Resistance to PARP Inhibitors: A Focus on the Clinical Landscape. Cancers [Internet]. 2021. [cited 2022 Dec 16];14:44. Available from: https://www.mdpi.com/2072-6694/14/1/44 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Leibowitz BD, Dougherty BV, Bell JSK, Kapilivsky J, Michuda J, Sedgewick AJ, et al. Validation of genomic and transcriptomic models of homologous recombination deficiency in a real-world pan-cancer cohort. BMC Cancer [Internet]. 2022. [cited 2023 Jun 1];22:587. Available from: 10.1186/s12885-022-09669-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Cheng HH, Salipante SJ, Nelson PS, Montgomery B, Pritchard CC. Polyclonal BRCA2 Reversion Mutations Detected in Circulating Tumor DNA After Platinum Chemotherapy in a Patient With Metastatic Prostate Cancer. JCO Precision Oncology [Internet]. 2018. [cited 2023 Mar 3];1–5. Available from: 10.1200/PO.17.00169 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ganesan S Tumor Suppressor Tolerance: Reversion Mutations in BRCA1 and BRCA2 and Resistance to PARP Inhibitors and Platinum. JCO Precision Oncology [Internet]. 2018. [cited 2023 Mar 3];1–4. Available from: 10.1200/PO.18.00001 [DOI] [PubMed] [Google Scholar]
- 20.Domchek SM. Reversion Mutations with Clinical Use of PARP Inhibitors: Many Genes, Many Versions. Cancer Discovery [Internet]. 2017. [cited 2023 Mar 3];7:937–9. Available from: https://aacrjournals.org/cancerdiscovery/article/7/9/937/6026/Reversion-Mutations-with-Clinical-Use-of-PARP [DOI] [PubMed] [Google Scholar]
- 21.Goodall J, Mateo J, Yuan W, Mossop H, Porta N, Miranda S, et al. Circulating Cell-Free DNA to Guide Prostate Cancer Treatment with PARP Inhibition. Cancer Discovery [Internet]. 2017. [cited 2023 Mar 3];7:1006–17. Available from: https://aacrjournals.org/cancerdiscovery/article/7/9/1006/6097/Circulating-Cell-Free-DNA-to-Guide-Prostate-Cancer [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Murciano-Goroff YR, Schram AM, Rosen EY, Won H, Gong Y, Noronha AM, et al. Reversion mutations in germline BRCA1/2-mutant tumors reveal a BRCA-mediated phenotype in non-canonical histologies. Nat Commun [Internet]. 2022. [cited 2023 Jan 12];13:7182. Available from: https://www.nature.com/articles/s41467-022-34109-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Banda K, Swisher EM, Wu D, Pritchard CC, Gadi VK. Somatic Reversion of Germline BRCA2 Mutation Confers Resistance to Poly(ADP-ribose) Polymerase Inhibitor Therapy. JCO Precision Oncology [Internet]. 2018. [cited 2023 May 30];1–6. Available from: 10.1200/PO.17.00044 [DOI] [PubMed] [Google Scholar]
- 24.Gadducci A, Cosio S, Landoni F, Lissoni AA, Zola P, Laudani ME, et al. Response to Chemotherapy and Clinical Outcome of Patients With Recurrent Epithelial Ovarian Cancer After PARP Inhibitor Maintenance Treatment: A Multicenter Retrospective Italian Study. Anticancer Res [Internet]. 2022. [cited 2022 Dec 19];42:2017–22. Available from: 10.21873/anticanres.15681 [DOI] [PubMed] [Google Scholar]
- 25.Moubarak M, Harter P, Ataseven B, Traut A, Welz J, Baert T, et al. Re-treatment with PARPi in patients with recurrent epithelial ovarian cancer: A single institutional experience. Gynecologic Oncology Reports [Internet]. 2022. [cited 2023 May 20];40:100939. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2352578922000194 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Frenel J-S, Kim J-W, Berton-Rigaud D, Asher R, Vidal L, Pautier P, et al. Efficacy of subsequent chemotherapy for patients with BRCA1/2 mutated platinum-sensitive recurrent epithelial ovarian cancer (EOC) progressing on olaparib vs placebo: The SOLO2/ENGOT Ov-21 trial. Annals of Oncology [Internet]. 2020. [cited 2022 Dec 16];31:S615. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0923753420409482 [DOI] [PubMed] [Google Scholar]
- 27.Cecere SC, Giannone G, Salutari V, Arenare L, Lorusso D, Ronzino G, et al. Olaparib as maintenance therapy in patients with BRCA 1–2 mutated recurrent platinum sensitive ovarian cancer: Real world data and post progression outcome. Gynecologic Oncology [Internet]. 2020. [cited 2022 Dec 19];156:38–44. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0090825819316075 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.