Ovarian cancer is a relatively uncommon but serious disease. It ranks 10th for female cancers, 5th for mortality, and its origin is still imperfectly known. It has a silent history for a long time, is often diagnosed late, and the prognosis is poor with a high relapse rate. It is therefore necessary to assess and prevent the risk of relapse, in order to establish a diagnosis as early as possible, and thus set up the appropriate treatment. Poly-ADP-Ribose Polymerase (PARP) inhibitors such as OLAPARIB and NIRAPARIB are effective in maintenance to prevent the risk of relapse in patients with recurrent platinum-sensitive ovarian cancer, as proved by recent data from the medical literature. Nevertheless, there may be a difference between "real life" and clinical trial data. Thus, the objective of this cohort is to assess whether the efficacy and safety of PARP inhibitors is the same in Finistère patients as in the scientific literature.

Ovarian cancer is a relatively uncommon but serious disease. It ranks 10th for female cancers and 5th for mortality (4,720 new cases in France in 2017, for 3,111 deaths), and its origin is still imperfectly known. Several risk factors have been identified, the main ones being genetic (importance of heredity, BRCA mutations) and environmental (taking hormone replacement therapy based on estrogen at menopause, exposure to tobacco, asbestos…). Epithelial tumors are the most frequent (85%), with different subtypes: serous (40 to 50%); endometrioid (20%); mucinous (15%); clear cell (5%); or with transitional, undifferentiated and mixed cells. The classification used in clinical practice is the TNM classification which establishes the stage of the disease (FIGO). Initial treatment is based on surgery and chemotherapy. The standard scheme combines CARBOPLATIN and PACLITAXEL for 6 cycles. For advanced stages, IIIB, IIIC and IV, BEVACIZUMAB can be administered every 3 weeks in combination with CARBOPLATIN and PACLITAXEL, and is then continued as monotherapy for 15 months The management of recurrences depends on their time of occurrence. Before 6 months, it is a resistance to platinum, a monochemotherapy without platinum is then indicated, associated with BEVACIZUMAB, if it was not given initially and in the absence of contraindication. After 6 months, polychemotherapy with platinum must be administered, possibly associated with BEVACIZUMAB, again if it was not given initially and in the absence of a contraindication. This treatment is then continued after stopping chemotherapy, until progression or toxicity .The arrival of PARP (Poly-ADP-Ribose-Polymerase) inhibitors, such as OLAPARIB, NIRAPARIB, or even RUCAPARIB, has considerably modified the treatment regimen of these relapses known as "platinum-sensitive" These molecules act on the DNA repair system, in synergy with the loss of BRCA function by tumor cells, causing significant genetic instability leading to cell death. This is the principle of synthetic lethality. These treatments will also in the very near future modify the management of patients with ovarian cancer as soon as they are managed. In fact, the SOLO 1 study published in 2018 showed the benefit of OLAPARIB prescribed as maintenance treatment in BRCA mutated patients, in complete or partial response after chemotherapy based on platinum The PRIMA and VELIA studies, published in 2019, have shown the effectiveness of NIRAPARIB and VELIPARIB respectively, prescribed as maintenance therapy, regardless of the patients' BRCA status.The benefit of the treatment was greater in the group of patients with a defect in homologous recombination. Finally, the PAOLA study published in 2019, showed the benefit of OLAPARIB versus placebo, in combination with BEVACIZUMAB, as maintenance treatment in patients with ovarian cancer in full or partial response after chemotherapy with PLATINUM-PACLITAXEL-BEVACIZUMAB, regardless of their BRCA status PARP inhibitors are effective in maintenance to prevent the risk of relapse, in patients with recurrent ovarian cancer sensitive to platinum (relapse> 6 months after the last course of platinum), as proved by recent data from the medical literature. On the one hand, the SOLO 2 study showed an improvement in progression-free survival in patients with a BRCA constitutional or somatic mutation, in relapse of platinum-sensitive ovarian cancer, treated with OLAPARIB in maintenance (19.1 months versus 5.5 months) On the other hand, the NOVA (2016) and ARIEL 3 (2017) studies showed an improvement in progression-free survival in relapsed patients with platinum-sensitive ovarian cancer, treated with maintenance by NIRAPARIB or by RUCAPARIB respectively, regardless of their mutational status for BRCA (probable action on other homologous recombination pathways than BRCA, still unknown to date) All these molecules are now part of our therapeutic arsenal, from the 2nd line of treatment.

In this context of major progress in the management of ovarian cancer, we decided to set up a Finistère study, aiming to analyze the characteristics and survival data of patients treated with PARP inhibitor for a recurrent platinum-sensitive ovarian cancer, as well as their tolerance to these treatments. This is a retrospective study, using data from Finistère in comparison with data from the literature. Nevertheless, there may be a difference between "real life" and clinical trial data. Thus, the objective of this cohort is to assess whether the efficacy (survival) and safety (side effects) of PARP inhibitors is the same in Finistère patients as in the scientific literature.

Inclusion Criteria:

• ≥ 18 years old
• High grade serous or clear cell ovarian cancer or endometrioid, including primary peritoneal cancer and fallopian tubes
• Treatment with PARP inhibitor (OLAPARIB, NIRAPARIB) in maintenance after sensitive platinum relapse
• No objection made

Exclusion Criteria:

• - Patients under judicial protection (guardianship)
• Refusal to participate

Patients who started treatment with PARP inhibitors on maintenance, between January 1, 2014 and December 31, 2019, for recurrent platinum-sensitive ovarian cancer

All collected data that underlie results in a publication

Data will be available beginning 8 months and ending five years following the publication

Data access requests will be reviewed by the internal committee of Brest UH. Requestors will be required to sign and complete a data access agreement.