Real-world efficacy and safety data of the use of PARP inhibitors in advanced ovarian cancer: the experience of the Hellenic Cooperative Oncology Group

1. Introduction

Ovarian cancer remains the leading cause of death among gynaecological malignancies [1]. Despite improvements in the five-year survival rate through complete debulking surgery and platinum-based chemotherapy, significant unmet needs persist. The introduction of poly (ADP-ribose) polymerase inhibitors (PARPi) has notably expanded treatment options for patients with advanced high-grade serous ovarian cancer (HGSOC) [2, 3, 4]. Olaparib has demonstrated efficacy in multiple maintenance settings for patients with advanced Breast Cancer gene 1, Breast Cancer gene 2 (BRCA1/2)-mutated (germline and/or somatic) HGSOC, who achieve complete or partial response following first-line platinum-based chemotherapy [2]; in combination with bevacizumab for patients with homologous recombination deficiency (HRD)-positive advanced HGSOC, following response to first-line platinum-based chemotherapy [5]; and for patients with platinum-sensitive relapsed HGSOC who respond to platinum-based chemotherapy [4].

Similarly, niraparib is approved for maintenance treatment as monotherapy for patients with advanced HGSOC who achieve complete or partial response following first-line platinum-based chemotherapy [3] and for patients with platinum-sensitive relapsed HGSOC who respond to platinum-based chemotherapy. These advancements reflect a growing emphasis on targeted maintenance therapies, thus improving progression-free survival (PFS) and providing personalised treatment strategies in ovarian cancer management.

While randomised trials provide valuable data on the safety and efficacy of novel treatment regimens, real-world studies comprise a broader patient population, which includes elderly patients and those with comorbidities [6]. These data are important for everyday clinical practice, offering insights that may support informed treatment decisions and guideline development.

The purpose of this study was to evaluate toxicity and clinical outcome data in patients with advanced HGSOC who received PARPi as maintenance after first- or second-line treatment in a real-world setting.

2. Materials and methods

3. Results

3.2 Clinical outcomes

The median follow-up period was 57.8 months (min 3.2, max 276.6) for patients who did not receive any maintenance treatment, 35.9 months (min 4.0, max 57.5) for patients who received olaparib, and 64.3 months (min 14.7, max 123.7) for patients who received bevacizumab as maintenance treatment.

In first-line treatment, maintenance therapy significantly improved PFS compared to no maintenance (Supplementary Table 2). Patients who received maintenance treatment had a median PFS of 28.2 months versus a median PFS of 17.1 months for those without maintenance treatment (hazard ratio (HR) = 0.53, p < 0.001; adjusted hazard ration (aHR) = 0.44, p < 0.001) (Supplementary Fig. 2). Moreover, olaparib as maintenance treatment demonstrated greatest benefit—with a non-estimable (NE) median PFS—compared to no maintenance and was associated with an aHR of 0.16 (p < 0.001) (Fig. 1a). Bevacizumab also demonstrated a statistically significant improvement in PFS, with a median of 26.3 months (aHR = 0.61, p = 0.024) compared to no maintenance. Additionally, patients who received maintenance treatment with olaparib after first-line chemotherapy had longer PFS compared to patients who received bevacizumab (aHR = 0.30, p < 0.001) (Fig. 1b).

In terms of OS, no significant differences were observed between patients who received maintenance compared to the remaining patients (Supplementary Fig. 3). In addition, there was no difference among subgroups who received different maintenance treatments (olaparib and bevacizumab) (Fig. 1c).

Fig. 1.Progression-free survival and overall survival according to maintenance treatment after first-line therapy. (a) Progression-free survival (PFS) in patients who received maintenance treatment with olaparib or bevacizumab after first-line therapy compared to patients who did not receive any maintenance treatment. (b) Progression-free survival (PFS) in patients who received maintenance treatment with olaparib after first-line therapy compared to patients who received maintenance treatment with bevacizumab. (c) Overall survival (OS) in patients who received maintenance treatment with olaparib after first-line therapy compared to patients who received maintenance treatment with bevacizumab.

In the second-line setting, maintenance therapy was also associated with improved PFS. Patients who received maintenance had a median PFS of 23.0 months compared to 12.1 months in patients without maintenance (HR = 0.37, p < 0.001; aHR = 0.44, p < 0.001) (Fig. 2a). Olaparib exhibited the most pronounced improvement, with a median PFS of 25.4 months (aHR = 0.36, p < 0.001). Bevacizumab provided a smaller, non-significant benefit, with a median PFS of 19.5 months (aHR = 0.76, p = 0.399) (Fig. 2b). With regard to patients who received maintenance treatment, there was a trend for longer PFS in patients who received olaparib compared to patients who received bevacizumab (aHR = 0.52, p = 0.062) (Fig. 2c). The results were similar for older patients (>65 years) in terms of benefit of olaparib as maintenance treatment after first-line chemotherapy. However, no benefit was noted for this group after second-line treatment compared to bevacizumab (Supplementary Table 3).

Fig. 2.Progression-free survival according to maintenance treatment after second-line therapy. (a) Progression-free survival (PFS) in patients who received maintenance treatment after second-line therapy compared to patients who did not receive any maintenance treatment. (b) Progression-free survival (PFS) in patients who received maintenance treatment with olaparib after second-line therapy compared to patients who received maintenance treatment with bevacizumab or no maintenance treatment. (c) Progression-free survival (PFS) in patients who received maintenance treatment with olaparib after second-line therapy compared to patients who received maintenance treatment with bevacizumab.

Further, the reported aHRs were adjusted for key factors, including age, FIGO stage, germline BRCA1/2 status, and family history of cancer (Supplementary Table 3).

Patients with germline BRCA1/2 mutations showed no significant differences in PFS or OS compared to those without germline BRCA1/2 mutations. The median PFS was 22.7 months (95% CI (Confidence Interval): 20.4, 33.3) in mutation carriers versus 22.6 months (95% CI: 16.4–28.9) in wild-type patients (p = 0.160) (Supplementary Fig. 4). Similarly, the median OS was 83.3 months (95% CI: 72.8–95.9) in mutation carriers versus 77.2 months (95% CI: 56.7, 137.4) in wild-type patients (p = 0.990) (Supplementary Fig. 5).

Patients with germline BRCA1/2 mutations who received PARP inhibitors (olaparib, niraparib, rucaparib, or olaparib and bevacizumab as maintenance after first-line therapy) had a median PFS1 of 31.5 months (95% CI: 25.0–50.4) compared to 19.1 months (95% CI: 16.4–24.1) in the remaining patients (p = 0.0018) (Fig. 3a). With regard OS, patients with germline BRCA1/2 mutations who received PARP inhibitors had a median OS of 83.7 months (95% CI: 64.4–NE), compared to 78.2 months (95% CI: 65.9–96.0) in the remainder of the patients (p = 0.490) (Fig. 3b).

Fig. 3.Comparison of progression-free and overall survival between patients with germline BRCA1/2 mutations receiving maintenance treatment with PARP inhibitors and the rest of the patients. (a) Progression-free survival (PFS) in patients with germline BRCA1/2 mutations who received maintenance treatment with PARP inhibitors compared to the rest of patients. (b) Overall survival (OS) in patients with germline BRCA1/2 mutations who received maintenance treatment with PARP inhibitors compared to the rest of patients. BRCA1/2: Breast Cancer gene 1/Breast Cancer gene 2.

The response data are summarised in Supplementary Table 4. Specifically, the selection of maintenance treatment is depicted according to clinical response after first- and second-line platinum-based chemotherapy treatment.

4. Discussion

The introduction of PARPi in recent years has significantly transformed the outcomes and management of advanced ovarian cancer. Their use as maintenance therapy, both in the first-line setting and in patients with platinum-sensitive relapsed HGSOC and who have achieved complete or partial response to platinum-based chemotherapy, has demonstrated substantial benefits. Real-world data from everyday clinical practice confirm the efficacy and toxicity results observed in landmark clinical trials. However, the optimal sequencing of maintenance treatments remains unclear.

In this study, we evaluated real-world clinical outcomes and toxicity data in patients with advanced HGSOC who received maintenance treatment with PARPi, bevacizumab, or no maintenance therapy in the first- or second-line treatment setting. It is important to note that during the patient registration period, niraparib was not yet approved for first-line maintenance nor was the combination of bevacizumab and olaparib for HRD-positive patients. Moreover, during the period in which the study was conducted, the administration of PARP inhibitors as second-line maintenance therapy after platinum-sensitive relapse did not require biomarker testing. Patients could receive olaparib or other PARPi as first-line maintenance without prior biomarker assessment as part of clinical trials. The use of olaparib as first-line maintenance therapy for patients with BRCA mutations was approved in Greece at the end of 2019, thus explaining the limited number of patients who received olaparib as first-line maintenance therapy in our study. Consequently, the number of patients who received niraparib or the combination of bevacizumab and olaparib was negligible and, thus, excluded from our analysis. Moreover, due to the time gap in the approval of these therapies, more patients received olaparib compared to niraparib.

A key aspect that arose from our analysis that we questioned is the relapse-free interval or PFS of patients with germline BRCA1/2 (gBRCA1/2) mutations who did not receive any first-line maintenance therapy. In our dataset, 33 gBRCA1/2-mutant patients who received no first-line maintenance treatment had a PFS of 17 months (95% CI: 14.3–19.2), which is consistent with findings from other real-world data studies [7].

Current consensus suggests that bevacizumab is most effective in cases with a significant disease burden [8]. Our observations align with previous findings that indicate that bevacizumab demonstrates reduced efficacy as a first-line maintenance treatment for ovarian cancer compared to its performance in subsequent maintenance lines [9]. According to Sznurkowski et al. [9], bevacizumab may be better reserved for second-line maintenance therapy, while PARPi should be offered to all advanced ovarian cancer patients who respond to first-line platinum-based chemotherapy. However, additional molecular predictors are needed to better identify patients who would benefit most from bevacizumab. In contrast, PARPi regimens have consistently shown substantial benefits in PFS and OS in newly diagnosed advanced ovarian cancer [2, 10, 11, 12]. Nevertheless, their effectiveness in the second-line maintenance setting—particularly for patients previously treated with PARPi as the first line—remains controversial [9]. Regardless of individual risk factors, maintenance treatment has been associated with a longer time to next therapy (TTNT) in real-world Advanced ovarian cancer (AOC) patients [13].

Moreover, the PFS observed in the first-line maintenance group treated with bevacizumab appears to be higher compared to that reported in clinical trials [14, 15]. In our study, 54% of patients achieved a complete response following first-line therapy, while only 55% (66 of 120) of patients who received maintenance treatment in the first-line setting were treated with bevacizumab. The high rate of complete response, combined with the relatively small number of patients who received bevacizumab, may have contributed to the notably prolonged PFS observed in our study.

Other real-world studies have explored switch maintenance with niraparib or olaparib monotherapy, thus suggesting that it may be a viable option for patients with advanced ovarian cancer [13, 16, 17]. Based on current evidence, bevacizumab appears more suitable for second-line maintenance treatment, while PARPi should remain the standard of care for all advanced ovarian cancer patients who achieve a positive response to first-line platinum-based chemotherapy. Further, our results confirm the results of other real-world data studies on PARPi use [18, 19, 20]. For example, in one study, investigators demonstrated that PARPi provided stable disease in a high proportion of recurrent ovarian cancer patients who had pathogenic HR mutations, with toxicities comparable to major trials [18]. Patients with non-BRCA HR and somatic BRCA mutations could benefit from PARPi [18]. Other investigators reported that olaparib and niraparib were effective and well-tolerated for those sensitive to platinum-based chemotherapy in a real-world setting [19]. Finally, in another multicentre real-world study, the use of PARPi demonstrated efficacy and toxicity rates similar to those in clinical trial data. According to our data, maintenance therapy regardless of regimen significantly increases PFS in both first- and second-line treatments.

The benefit in PFS is clearly higher with olaparib compared with bevacizumab as maintenance therapy in first-line treatment, particularly for BRCA-mutated patients. However, no statistically significant difference in OS was demonstrated in our study. These results can be attributed to the use of PARPi mainly in second-line treatments and not in first-line ones. In addition, the limited sample size may have impacted this result. It is important to note that a proportion of patients were not tested for genetic mutations and, consequently, received maintenance therapy regardless of mutation status and likely not in the first-line setting. Our data support the importance of administering PARPi to mutation carriers in the first-line setting.

In terms of toxicity, manageable adverse effects of maintenance treatment with either regimen were observed, with anaemia grade III being the most serious one and mostly observed in patients treated with olaparib. Importantly, we did not identify significant increases in toxicity rates among elderly patients. Adverse events were documented starting from the administration of maintenance therapy. Consequently, low rates of neutropenia and neurotoxicity were documented. To the best of our knowledge, this is the first study based on real-world data to report on differences in toxicity and discontinuation rates of treatment between the elderly and other patients. Additionally, in our study, among patients who received PARPi as maintenance treatment, one patient was diagnosed with acute myeloid leukaemia and two with myelodysplastic syndrome, which agrees with the results of clinical trials [21, 22]. Specifically, after being treated with olaparib for three years, one patient was diagnosed with myelodysplastic syndrome with ring sideroblasts (MDS-RS), Revised International Prognostic Scoring System (IPSS-R) score 7.5 (very high risk), WHO Classification-based Prognostic Scoring System (WPSS) 3, and IPSS 1.5. The patient received azacytidine 75 mg/m²; two years later, the patient was referred for evaluation regarding the possibility of an allogeneic hematopoietic stem cell transplantation due to cytogenetic progression. The second patient was diagnosed with acute myeloid leukaemia after being treated with rucaparib for 27 months and died a month later. The third patient was diagnosed with myelodysplastic syndrome (MDS-EB 2) after 33 months of treatment with olaparib and died 9 months later.

Finally, germline and somatic testing revealed mutations in a significant proportion of the patients of our cohort. The reported rate of germline mutations (53.7%) was higher than what is generally reported in most published series, which may have been associated with selection bias. As previously described, selected germline pathogenic variants may be missed by tumour testing [23, 24, 25]. This could be attributed to methodological limitations and differences in the assays used for testing, definitions of pathogenicity, and the number of genes or variants tested [23]. In addition, missing germline alterations may have significant implications both for the family and for healthy relatives. Therefore, based on international guidelines, germline testing needs to be performed in all patients with epithelial ovarian cancer, irrespective of the results of genomic testing [10, 26]. In our cohort, 73.5% of patients had undergone both germline and somatic testing, possibly due to diagnosis in earlier years, but mostly due to the lack of reimbursement for somatic testing and, until recently, germline testing in our country.

The limitations of this study include its retrospective nature, lack of patients undergoing HRD testing to assess treatment efficacy, and the use of PARPi as maintenance treatment in a large proportion of patients, thus confounding clinical outcomes and the relatively limited number of patients. It is estimated that almost 800 women in Greece are diagnosed with advanced ovarian cancer each year. While a higher number of patients could have been included in this study based on the number of newly diagnosed cases per year, it is important we acknowledge that not all HeCOG-affiliated oncology centres were specialised in ovarian cancer, thereby leading to a lower number of patients eligible for this study. Despite these inherent study limitations, our real-world findings complement clinical trial data and emphasise the need for larger studies with systematic HRD evaluation to refine patient selection and improve long-term outcomes in advanced ovarian cancer. Finally, the frequent detection of BRCA1/2 mutations highlights the importance of universal molecular profiling for precise treatment selection.

While our study supports the benefit of PARPi, particularly olaparib, in improving PFS in both first- and second-line treatment settings, emerging data raise important concerns regarding their impact on overall survival (OS). Recent analyses and regulatory actions by the US Food and Drug Administration (FDA) have led to the withdrawal or narrowing of several PARPi indications for the treatment and maintenance of advanced epithelial ovarian cancer (EOC) due to the potential OS detriment observed in multiple randomised controlled trials (e.g., SOLO3, ARIEL3, NOVA, and ARIEL4) [27].

In particular, benefits from PARPi appear to be more clearly sustained in patients with germline or somatic BRCA mutations, whereas patients without BRCA mutations or HRD-negative status may not derive long-term survival benefit and could potentially experience harm. Consequently, maintenance indications have been restricted to BRCA-mutant populations, and regulatory authorities now require stricter biomarker-based selection for treatment eligibility.

5. Conclusions

This multicentre real-world study by HeCOG confirms the clinical benefit of PARPi maintenance, particularly olaparib, in patients with HGSOC. Olaparib was found to significantly improved PFS in both first- and second-line settings compared to no maintenance or maintenance treatment with bevacizumab, with an acceptable and manageable safety profile. Hematologic toxicities—such as anaemia grade III—were more common with olaparib, but toxicity was similar across age groups, thus suggesting that age should not limit the use of PARPi.

Further, while there was a clear improvement in PFS, there was no observation of any statistically significant OS benefit, thereby highlighting the need for longer follow-up and further investigation into optimal treatment sequencing. Further, the inclusion of patients without BRCA or HRD alterations also highlights the importance of early and comprehensive molecular testing to guide treatment decisions and avoid exposing patients to costly therapies with limited survival benefits.

Overall, equitable access to genomic testing and biomarker-driven care should be prioritised in clinical practice and supported by appropriate policy and reimbursement frameworks. Future research should focus on refining patient selection and maximising the real-world impact of PARP inhibitors within a precision oncology framework.

Availability of Data and Materials

The data presented in this study are available in the article while further details can be obtained on request from the corresponding author.

Author contributions

AN, EF, GF—conceptualization, resources, writing–original draft preparation. ML, PP, GP, SK, AmP, AngK, GA, AnnK, IB, KoP, AnP, NT, CR, PK—resources, writing–review and editing. KD—formal analysis, writing–original draft preparation. KyP—investigation, writing–original draft preparation. FF—investigation, writing–review and editing. AC—conceptualization, resources, writing–original draft preparation.

Ethics approval and consent to participate

This study was approved by the scientific committee of S. Andrew Hospital, Patras, Greece (Prot. Number S.C. 484/09-12-20). Informed consent was obtained from all alive participants, while waiver of consent was obtained for deceased patients.

Acknowledgment

The authors are indebted to all patients and their families for their trust and participation in the Hellenic Cooperative Oncology Group (HeCOG) studies. The authors also thank the data managers of the Group for data collection and Maria Moschoni and Anastasia Kitka for coordinating the study.

Funding

This study was supported by a Hellenic Society of Medical Oncology (HeSMO) research grant and by a Hellenic Cooperative Oncology Group (HeCOG) internal research grant.

Conflict of interest

Michalis Liontos is serving as one of the Editorial Board members of this journal. We declare that Michalis Liontos had no involvement in the peer review of this article and has no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to TM.

Adamantia Nikolaidi: Advisory Board: Pfizer, Novartis, AstraZeneca, Glaxo, Ipsen, Roche, Gilead. Speaker fees (Invited Speaker): Pfizer, Novartis, AstraZeneca, Glaxo, Gilead.

Michalis Liontos: Employment—Pfizer. Stock and Other Ownership Interests—Pfizer. Honoraria—Astellas Pharma; AstraZeneca; Bristol-Myers Squibb/Celgene; Ipsen; Janssen; MSD Oncology; Roche; Sanofi. Consulting or Advisory Role—Amgen; AstraZeneca; GlaxoSmithKline; Janssen. Travel, Accommodations, Expenses—AstraZeneca; Bayer; Pfizer; Roche.

Elena Fountzilas: Advisory Role: Amgen. Invited speaker fees: Roche, Lilly, AstraZeneca, GSK. Travel grant: AstraZeneca, Pfizer, Genesis, and K.A.M. Oncology/Hematology. Stock ownership: Genprex Inc., Deciphera Pharmaceuticals Inc.

Sofia Karageorgopoulou: Consultation/Advisory: Astra Zeneca, Roche. Research: Novartis, Roche. Speaker: Astra Zeneca, Novartis, Roche.

Amanda Psyrri: Consultation Fees: Amgen, Merck Serono, Roche, BMS, AstraZeneca, MSD. Honoraria: Amgen, Merck Serono, Roche, BMS, AstraZeneca, MSD. Research funding: BMS, Kura Oncology, DEMO, Roche.

Angelos Koutras: Advisory Board: Pierre Fabre, AstraZeneca, Gilead, Pfizer, Genesis, MSD, BMS. Invited Speaker: Sanofi, Gilead, AstraZeneca, Sandoz. Travel grant: Rafarm, Lilly, Ipsen, Gilead, Pfizer, Genesis, AstraZeneca.

Anna Koumarianou: Advisory Role: Genesis Pharma. Honoraria: Pfizer. Speaker’s bureau: Roche. Research Funding: Merck. Travel: MSD. Educational grants: Novartis, Pfizer, Merck, Roche, BMS, MSD, Genesis, and Ipsen.

Anastasios Papadopoulos: Advisory role: Gilead. Speeker fees (Invited Speaker): Novartis, Remedica.

Paris Kosmidis: Advisory board: Bristol, Gilead, Lilly; Travel Congress sponsor: MSD, Pfizer.

George Fountzilas: Advisory Board of Pfizer, Novartis. Honoraria from AstraZeneca, Novartis. Stock ownership: Genprex, Daiichi Sankyo, RFL Holdings, FORMYCON.

Athina Christopoulou: Advisory and sponsored conferences: Astra, Roche, MSD, BMS, Sanofi, Amgen, Pfizer, Novartis.

The rest of the authors declare no conflict of interest.

Supplementary material

Supplementary material associated with this article can be found, in the online version, at https://oss.ejgo.net/files/article/2000434795576737792/attachment/Supplementary%20material.docx.