Efficacy of Chemotherapy Following Prior PARP-Inhibitor Treatment in Patients with Ovarian Cancer
Article information
Abstract
Purpose
Considering the current lack of consensus on post–poly(adenosine diphosphate-ribose) polymerase inhibitor (PARPi) treatment strategies, this study aimed to evaluate the efficacy of subsequent therapy and compare the outcomes of regimes in patients with recurrent ovarian cancer after PARPi treatment.
Materials and Methods
This multi-center retrospective cohort study analyzed data on patients diagnosed with ovarian cancer between January 2012 and June 2023 who had previously used PARPi after first- to fourth-line platinum-based chemotherapy. The primary endpoint was progression-free survival (PFS), which was the interval between recurrence after using PARPi and subsequent recurrence in the case of recurrence.
Results
Of 318 patients, 147/318 (46.2%) recurred after the PARPi maintenance. Patients were categorized into groups based on subsequent therapy except non-treated (11/147, 7.5%): platinum-based chemotherapy (89/147, 60.5%), non-platinum-based chemotherapy (21/147, 14.3%), other treatments (26/147, 17.7%), and the median PFS (mPFS) for each group were 7.3, 4.8, and 11.4 months, respectively. Among the platinum-based chemotherapy group, the gemcitabine+carboplatin regimen demonstrated a longer mPFS (10.1 months) than the other regimens (6.6 months, p=0.019). In non-platinum-based chemotherapy, no statistically significant differences were observed among the regimens. And, in the other therapy group, where the proportion of patients with oligometastasis was as high as 88.5%, no significant differences were observed among the therapies, including other modalities.
Conclusion
In the subsequent chemotherapy of recurrent ovarian cancer after platinum-based chemotherapy and PARPi, the gemcitabine+carboplatin regimen demonstrated a potential to delay recurrence more effectively compared to other therapies.
Introduction
Ovarian cancer remains one of the most lethal gynecologic malignancies worldwide, with high recurrence rates and limited long-term survival despite advances in treatment [1]. Although the standard treatment—cytoreductive surgery followed by platinum-based chemotherapy—offers initial disease control, most patients experience disease relapse within a few years, necessitating novel therapeutic strategies [2].
The introduction of poly(adenosine diphosphate-ribose) polymerase inhibitors (PARPi) has significantly altered the therapeutic landscape, offering a targeted approach for patients with homologous recombination deficiency, particularly those with BRCA1/2 mutations [3]. PARPi has demonstrated efficacy in prolonging progression-free survival (PFS) in both frontline and recurrent settings, especially when used as maintenance therapy after platinum-based chemotherapy [4]. This advancement has provided new hope for patients, particularly those with biomarker-defined subgroups who benefit most from PARPi therapy. However, the overall survival (OS) benefit has not been statistically proven. Additionally, subsequent analysis of the SOLO2 trial indicates a lower treatment response rate to post-relapse anticancer therapy in the olaparib-treated group [5]. Moreover, subsequent analysis of the SOLO2 trial has revealed a lower treatment response rate to post-relapse anticancer therapy in the olaparib-treated group compared to those not receiving PARPi. This suggests potential treatment resistance or altered tumor biology following PARPi exposure, raising concerns about the efficacy of subsequent therapeutic options [6].
In the changing treatment paradigms of recurrent ovarian cancer, no medical agreement exists for treatment after recurrence in patients with ovarian cancer who have been administered PARPi. Therefore, they are generally treated with platinum-based chemotherapy that has been used previously or undergo surgery and radiation therapy. Thus, this study was conducted to confirm the efficacy of chemotherapy for recurrent ovarian cancer after using PARPi.
Materials and Methods
1. Study population
This multi-center retrospective cohort study enrolled patients who met the following inclusion criteria: (1) pathologically confirmed high-grade serous ovarian, fallopian tube, and primary peritoneal carcinoma (collectively termed ovarian cancer) and (2) patients who have undergone platinum-based chemotherapy between first-line and fourth-line chemotherapy, followed by PARPi maintenance in whom progressive disease according to the Response Evaluation Criteria in Solid Tumor (RECIST) was confirmed.
2. Data collection
Based on the medical records, the patient clinicopathological data were collected, including age at diagnosis, BRCA1/2 mutational status, Bevacizumab use in previous treatment lines, line of chemotherapy at PARPi maintenance, Objective response in line used PARPi, progression-free interval (PFI) during line used PARPi, duration of PARPi use, type of PARPi, and subsequent chemotherapy regimen after PARPi maintenance.
During the period of surveillance, patients underwent imaging (such as computed tomography and magnetic resonance imaging) approximately every 12 weeks until 2 years after chemotherapy and then every 6 months or based on clinical symptoms. Disease progression was determined according to the RECIST ver. 1.1 or GCIG cancer antigen 125 (CA-125) response criteria [7].
The duration of disease progression or death after PARPi maintenance and the duration of disease progression or death after subsequent therapy were calculated. The primary endpoint, expressed as PFS, was the interval between disease progression after PARPi maintenance and next disease progression after subsequent therapy (Fig. 1).
Study design. CR, complete remission; CTx, chemotherapy; PARPi, poly(adenosine diphosphate-ribose) polymerase inhibitors; PFS, progression-free survival; PR, partial remission; Recur, recurrence; SD, stable disease.
For this endpoint, the patients were censored if they had not experienced subsequent disease progression or death at the last date known to be alive. Based on the RECIST ver. 1.1 or GCIG CA-125 response criteria, overall response to subsequent chemotherapy after PARPi maintenance was investigated as a secondary endpoint.
3. Statistical analysis
Descriptive statistics were used for demographic data and are summarized as standard deviation or frequency (%). Differences in patient characteristics between groups were compared using chi-squared or Mann-Whitney U test. Survival analyses were conducted using Kaplan-Meier method and log-rank test. Cox proportional hazards regression analyses were conducted, and hazard ratio (HR) and 95% confidence intervals (CIs) were calculated for multivariate analyses. Statistical analyses were performed using IBM SPSS statistics software ver. 26.0 (IBM Corp.) and GraphPad Prism software (GraphPad Software). A p-value < 0.05 was considered to indicate statistical significance.
Results
The present analysis included 318 patients who received PARPi maintenance after platinum-based chemotherapy in the first to fourth lines and median PFI was 22.8 months (Fig. 1).
Among them, 171 patients showed no disease progression during this period, and 147 showed disease progression. Most patients (n=89, 60.5%) received platinum-based chemotherapy for subsequent treatments. Additionally, 21 patients (14.3%) received non-platinum-based chemotherapy, and 26 (17.7%) underwent other treatments, such as radiotherapy, surgery, or immunotherapy. Furthermore, 11 patients (7.5%) were identified as not receiving any treatment.
A significant subset received belotecan/topotecan+platinum (n=27, 18.4%), with belotecan+cisplatin being the most common combination (n=15, 10.2%). Paclitaxel/docetaxel+platinum was administered to 20 (13.6%), with pac-litaxel+car-boplatin as the most frequent regimen (n=11, 7.5%). Gemcitabine+carboplatin was used in 20 (13.6%), and pegylated liposomal doxorubicin (PLD)+carboplatin in 16 (10.9%). Additionally, six (4.1%) received other platinum-based chemotherapy regimens.
Among those on non-platinum-based regimens, PLD was the most frequently used agent (n=10, 6.8%), followed by belotecan/topotecan (n=8, 5.4%) and paclitaxel (n=3, 2.0%). In the other therapy group, radiotherapy was chosen for 16 (10.9%), debulking surgery for six (4.1%), and other treatments for four (2.7%).
Based on the RECIST ver. 1.1, overall response for each subsequent therapy was as follows: among the patients receiving platinum-based chemotherapy, two showed complete response (CR), and 25 showed partial response (PR). Among those receiving non-platinum-based chemotherapy, there were no instances of CR and four cases of PR. In the group undergoing other treatments, five CR and seven PR cases were observed (Table 1).
Overall response during subsequent chemotherapy
The demographics of the patients who underwent these regimens were examined and multivariate analysis was conducted. While no significant differences among the three groups for age, BRCA status, line of chemotherapy in PARPi, bevacizumab use in previous chemotherapy, objective response to previous chemotherapy, and kind of PARPi were observed, PFI during line used PARPi and duration of PARPi use showed significant differences (Table 2).
Demographics in each subsequent therapy
Fig. 2 depicts the Kaplan-Meier survival curve comparing the PFS among the patients treated with platinum-based chemotherapy, non-platinum-based chemotherapy, and other treatments. The median PFS (mPFS) rates for platinum-based chemotherapy, non-platinum–based chemotherapy, and other treatment groups were 7.3, 4.8, and 11.4 months, respectively, and a statistically significant difference in PFS was observed among the three groups (p=0.001) (Fig. 2).
Progression-free survival of subsequent treatment. mPFS, median progression-free survival.
The four most administered chemotherapy regimens were belotecan/topotecan+platinum, paclitaxel/docetaxel+platinum, gemcitabine+carboplatin, and PLD+carboplatin (Table 1). There was no statistically significant difference in the demographics between the groups (S1 Table), and the mPFS for each regimen was 8.4, 6.4, 10.1, and 6.4 months, respectively (S2 Fig.). The PFS analysis indicated a statistically significant difference across the treatment groups (p=0.025). Specifically, compared to gemcitabine+carboplatin, paclitaxel/docetaxel+platinum-based medication exhibited an HR of 2.039 (95% CI, 1.035 to 4.669; p=0.024) and PLD+carboplatin showed an HR of 2.375 (95% CI, 1.109 to 5.083; p=0.008), suggesting that gemcitabine+carboplatin regimen has a more substantial impact on survival compared to the other regimens.
The PFS between the gemcitabine+carboplatin regimen and other platinum-based regimens shows a significant difference with a p-value of 0.019 (HR, 1.893; 95% CI, 1.174 to 3.053) (Fig. 3). Additionally, on comparing the gemcitabine+carboplatin regimen with the other three most frequent regimens, a significant difference was observed, as evidenced by a p-value of 0.025 (HR, 2.039; 95% CI, 1.035 to 4.669) (S3 Fig.).
Progression-free survival of gemcitabine+carboplatin and other platinum-based chemotherapy. CI, confidence interval; HR, hazard ratio; mPFS, median progression-free survival.
In a multivariate analysis conducted on the subsequent therapy, the objective response in line used, compared between CR and PR, demonstrated an HR of 1.463 (95% CI, 1.003 to 2.134) with a p-value of 0.048, indicating a significant impact. Additionally, the PFI during line used PARPi use also showed an HR of 0.965 (95% CI, 0.936 to 0.994) with a p-value of 0.018, suggesting a significant influence (S4 Table).
No statistically significant differences were observed in PFS among patients treated with paclitaxel (n=3, mPFS: 6.3 months), PLD (n=10, mPFS: 5.4 months), and belotecan/topotecan (n=8, mPFS: 3.6 months) in the non-platinum-based chemotherapy regimen (S5 Fig.).
Among other treatments, including operation (n=7), radiotherapy (n=15), and immunotherapy (n=4), no statistically significant differences in PFS were found among the three groups (p=0.067) (S6 Fig.). The operation group did not reach mPFS, while the mPFS for the radiotherapy and immunotherapy groups were 11.4 and 4.6 months, respectively. However, an investigation into the number of recurrent sites was possible only in the other treatments group, which confirmed that 23/26 (88.5%) of patients had fewer than three recurrent sites.
Discussion
1. Summary of main results
This study investigates the efficacy of subsequent therapies for ovarian cancer patients who have received PARPi as maintenance therapy after platinum-based chemotherapy. A retrospective cohort analysis of 318 patients showed that platinum-based chemotherapy yielded an mPFS of 7.3 months, with an overall response rate of 30.3%. Among platinum-based regimens, gemcitabine plus carboplatin resulted in the longest PFS (10.1 months), demonstrating a statistically significant advantage over other regimens (p=0.019). The study indicates that platinum-based regimens, particularly gemcitabine combined with carboplatin, could offer better outcomes in recurrent ovarian cancer patients post-PARPi treatment.
2. Results in the context of published literature
With the increasing use of PARPi, addressing the issue of subsequent treatment is essential. Unlike the era when conventional treatments were the sole option, optimizing patient care requires a careful combination of various medications for an effective and tailored approach.
This study presents the retrospective research results on PARPi use in patients experiencing recurrence. The aim was to identify the conventional chemotherapy regimen that can demonstrate a more favorable impact after post-PARPi treatment. The observed superiority of the gemcitabine+carboplatin regimen in mPFS compared to other regimens raises uncertainties regarding a clear theoretical explanation. Considering the existing studies, gemcitabine is metabolized intracellularly by nucleoside kinases to two active forms, gemcitabine diphosphate (dFdCDP) and gemcitabine triphosphate (dFdCTP). dFdCDP is an inhibitor of ribonucleotide reductase, which is responsible for producing deoxynucleotides (dNTPs) [8]. Thus, dFdCDP depletes the dNTP pools and further potentiates dFdCTP- mediated blockage of DNA strand elongation. dFdCTP is a deoxycytidine analog that competes with deoxycytidine triphosphate (dCTP) and disrupts DNA synthesis [9,10]. According to Thompson et al. [11], gemcitabine triggers ATR/Chk1-dependent replication stress response related to cytotoxicity. Recently, in a cell-level experiment, the results of effectively inducing apoptosis have been derived through a test using gemcitabine in combination with a Chk1 blockade such as pristimerin [12]. This effect of gemcitabine may be the reason for the results of this study.
Park et al. [13] reported in a multi-institutional retrospective cohort study that the use of PARPi in patients with ovarian cancer with BRCA mutations who experienced recurrence after the initial treatment negatively impacted the PFS, which is the period from subsequent therapy to recurrence. The worst effect was observed in patients who used PARPi for ≥ 12 months. The study, irrespective of BRCA mutation status, investigated the differences in subsequent regimens after using PARPi in various treatment lines to determine which regimen yields a better subsequent PFS, demarcated as ‘PFS2-PFS1’ in the study [13]. Contrary to the correlation between PARPi and the impact on subsequent treatments reported in the study, this research found differences. On multivariate analysis, the BRCA mutation status and the type of PARPi did not directly influence PFS in patients with recurrence. However, overall response to previous chemotherapy and PFI during line used PARPi impacted PFS after PARPi use. Consequently, factors affecting the impact of subsequent treatments upon recurrence in the current trend of ovarian cancer treatment with the prevalent use of PARPi require further examination through large-scale studies.
The mechanisms of PARPi resistance may also influence the effectiveness of chemotherapy after recurrence. In the PAOLA-1 study, exploratory analysis revealed that in cases where recurrence occurred while actively using PARPi, the role of subsequent chemotherapy was found to be more limited during follow-up after discontinuation compared to cases where recurrence occurred post-PARPi usage [14]. The SOLO2 study indicated a reduction in the effectiveness of platinum-based subsequent chemotherapy after recurrence in patients undergoing prolonged PARPi maintenance therapy [5]. While no conclusive evidence exists on the reasons for resistance to platinum-based subsequent chemotherapy after PARPi use in ovarian cancer, studies in breast cancer have demonstrated that patients with BRCA-mutated cancer experience partial restoration of homologous recombination due to the loss of the non-homologous end-joining protein p53-binding protein1 (53BP1) [15]. This restoration leads to PARPi resistance and resistance against DNA-damaging agents such as cisplatin and doxorubicin [16].
Recent studies have compared the OS outcomes associated with niraparib and olaparib. The SOLO-2 trial did not show a statistically significant improvement in patient OS with olaparib, also long-term follow-up results from the NOVA study did not show the same for niraparib [17,18]. This study focused not on OS but on the PFS following subsequent therapy after PARPi use. While a direct comparison of the studies is challenging due to differing endpoints, the type of PARPi did not appear to influence the difference in PFS across the sort of treatment. This study did not aim to determine OS outcomes, and long-term follow-up was not conducted.
Among the groups receiving platinum-based chemotherapy, non-platinum-based chemotherapy, and other treatments, the group demonstrating the longest mPFS was the other treatments group, with 11.4 months. However, this study was limited in accurately assessing the extent of metastatic lesions upon recurrence following PARPi use; this group consisted mostly of patients who underwent radiotherapy and surgery, which primarily included patients with oligometastasis defined by 1 to 3 metastatic lesions where removal may be feasible [19]. These could be different compared to those who undergo chemotherapy for systemic metastasis. Analysis of these differences in the future is necessary. Recent clinical trials have investigated the applications of secondary cytoreductive surgery following PARPi use, adding significance to the interpretation of these results [20].
3. Strengths and weaknesses
This study is the multi-center and large cohort design increase the generalizability of the results. And detailed subgroup analysis of different chemotherapy regimens allows for identifying the most effective treatment strategies post-PARPi. The study also provides significant clinical insights into the real-world efficacy of different therapeutic approaches following PARPi maintenance therapy.
But the study is retrospective, which may introduce biases such as patient selection and treatment allocation biases. The variability in treatment protocols across different institutions could affect the consistency of the results. And limited long-term follow-up data for OS make it challenging to draw conclusions about long-term patient outcomes.
4. Implications for practice and future research
This study highlights the potential benefit of gemcitabine plus carboplatin as a preferred chemotherapy regimen following PARPi in patients with recurrent ovarian cancer. In clinical practice, this regimen should be considered as a viable option in this patient population. Future research should focus on prospective trials to confirm these findings and explore the mechanisms behind differential responses to subsequent therapies post-PARPi. Additionally, studies should aim to establish standardized treatment guidelines and explore combination therapies, including targeted agents and immunotherapy.
In the subsequent chemotherapy treatment of recurrent ovarian cancer using platinum-based chemotherapy and PARPi, the gemcitabine+carboplatin regimen may show slightly better results than other chemotherapies, on which additional research is necessary.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
The institutional review boards of the participating three centers reviewed and approved this study (YUHS, 4-2022-1027; SNUH, H-2107-103-1234; SMC, 2022–01-122). This study was conducted in accordance with the Declaration of Helsinki. The requirement for informed consent was waived.
Author Contributions
Conceived and designed the analysis: Kim JC, Lee JY.
Collected the data: Kim JC, Park J, Lee YJ, Nam EJ, Kim SW, Kim SH, Kim YT, Kim SI, Kim JW, Kim BG, Lee JY.
Contributed data or analysis tools: Kim JC, Lee JY.
Performed the analysis: Kim JC, Lee JY.
Wrote the paper: Kim JC, Lee JY.
Data curation, formal analysis, investigation, software, visualization, validation: Kim JC.
Funding acquisition, methodology, project administration, resources, supervision, writing-review & editing: Lee JY.
Conflicts of Interest
The authors declare no conflict of interest. The study received financial support from Chong Kun Dang Pharmaceutical Corp., which had no role in the design, conduct, or reporting of the research.
Funding
This work was supported by Chong Kun Dang Pharmaceutical Corp. and the Soonchunhyang University Research Fund. We thank Chong Kun Dang Pharmaceutical Corp. and the Soonchunhyang University Research Fund for their financial support.
Acknowledgments
Data collection was performed at Yonsei University College of Medicine, while data analysis and manuscript preparation were carried out at Soonchunhyang University Bucheon Hospital.