500 Background: Benefit of adjuvant chemotherapy (CT) in addition to endocrine therapy (ET) remains controversial for patients (pts) aged ≥ 70 years with oestrogen receptors-positive (ER+) HER2-negative (HER2-) breast cancer (BC). In a large prospective trial, we first assessed the tumour genomic grade index (GGI) in all pts, and second, randomized pts with a high GGI between CT + ET vs. ET alone. Methods: Eligible pts were women ≥ 70 years with ER+ HER2- primary BC or isolated local relapse, irrespective of other characteristics, for whom adjuvant systemic treatment was considered. G8 score, Charlson comorbidity index (CCI) and 4-year mortality Lee score were collected at baseline. GGI was centrally performed by RT-PCR on FFPE samples. Pts with low GGI were not recommended to receive CT and were followed in an observational cohort. Pts with high (+ equivocal) GGI were randomized 1:1 to CT + ET vs. ET alone, using G8, pN and centre for stratification. Investigators chose between 3 CT regimens: 4 cycles of doxorubicin/cyclophosphamide, non-pegylated liposomal doxorubicin/cyclophosphamide or docetaxel/cyclophosphamide, given q3w with G-CSF. Standard ET consisted of 5 years of aromatase inhibitor, tamoxifen or a sequence based on tolerance. Based on CALGB 49907 results, the primary objective was to demonstrate an overall survival (OS) benefit for CT (4-year assumptions 87.5 vs 80%, HR=0.60) in the intent to treat (ITT) population. With 171 events, the trial had 90% power to demonstrate a difference with a bilateral test α=0.05. Secondary objectives included BC specific survival (BCSS), invasive disease-free survival (iDFS), event-free survival (EFS), competing events, cost-effectiveness and Q-TWiST analysis, geriatric dimensions, willingness and quality of life. Results: Between 04/2012 and 05/2016, 1,969 pts from 61 French and 12 Belgian centres were enrolled. Of them, 1,089 (55%) were randomized between CT + ET and ET alone. Median follow-up was 5.8 years at the data cut-off (17/12/2021) with 180 OS events observed. Median age was 75 (70-92), G8 score, CCI and Lee score being >14, ≤ 2, and ≤ 8 in 60%, 62% and 84% of pts, respectively. Tumours were ≥ pT2, pN+, isolated local relapses, with histological grade III, in 56%, 46%, 11% and 39% of cases, respectively. No significant OS difference was observed between arms (HR 0.85 [0.64-1.13], p=0.2538); 4-year OS was 90.5% in the CT + ET arm and 89.7% in the ET alone arm. The forest plot could not identify any subgroup deriving significant benefit from CT. ITT and per protocol analysis of secondary objectives (BCSS, iDFS, EFS) showed similar results. Conclusions: In this large phase III trial, we did not find a statistically significant OS benefit with the addition of CT to ET after surgery for ER+ HER2- BC with a high GGI. Analysis of the other outcome measures will be presented. Clinical trial information: NCT0156405.

No abstract

5508 Background: The Geriatric Vulnerability Score (GVS) combining albumin, lymphocyte count, ADL, IADL and HADS scores has been reported (Falandry C Ann Oncol 2013) to identify vulnerable elderly OC patients (pts) as those with a GVS≥3. For such pts, Carboplatin (Cb) monotherapy or weekly Cb plus paclitaxel (Pa) are often proposed as an alternative to Cb-Pa given every 3 weeks. Methods: Pts ≥70 yrs with first line FIGO stage III/IV epithelial OC were screened for GVS. Those with GVS≥3 were randomized to receive either arm A: Cb AUC5-6 + Pa 175mg/m², d1q3week or arm B: Cb AUC5-6 d1q3week or arm C:weekly Cb AUC2 + Pa 60mg/m² d1-d8-d15 q4week. Primary endpoint is treatment feasibility defined as the ability to complete 6 chemotherapy courses without disease progression, early treatment stopping due to unacceptable toxicity or death. Inclusion of 240 pts was planned. Results: Among 444 screened pts, 120 were randomized from 12/2013 to 04/2017 (armA = B = C = 40). Pts characteristics were well balanced between arms A-B-C respectively: median age (79-82-80 yrs), FIGO stage IV (32-37-27%), primary surgery (65-72-70%), absence of macroscopic residuals (CC-0) (7-5-7%), ECOG≥2 (50-50-47%). Feasibility per protocol for arms A-B-C is 65%, 47% and 60% (p = 0.15). Main reasons for treatment arrest are treatment toxicity (A:20%; B:15%; C:22.5%; p = 0.771) and disease progression (A:7.5%; B:30%; C:2%; p = 0.004). Median PFS for arm A-B-C are 12.5 mos (95%CI 10.3-15.3), 4.8 (3.8-15.3) and 8.3 (6.6-15.3), respectively (p < 0.001) and median OS for arm A-B-C is not reached (NR) (21, NR), 7.4 (5.3-NR) and 17.3 (10.8-NR), respectively (p = 0.001). At the pre-planned intermediate analysis, the IDMC recommended to prematurely close the study as survival in armB was found significantly worse and the number of potential pts required to find a significant difference between both Cb-Pa regimens (arms A&C) was out of reach. Conclusions: Compared to 3-weekly and weekly Cb-Pa regimens, Cb single agent was reported to be less active with significant worse survival outcome in vulnerable elderly pts. In this population Cb-Pa combination remains a standard. Clinical trial information: NCT02001272.

5500 Background: To investigate whether adding Pembrolizumab (P) to neoadjuvant carboplatin-paclitaxel chemotherapy (CP) may increase the optimal debulking rate, assessed by Complete Resection Rate (CRR) after Interval Debulking Surgery (IDS) in patients (pts) with initially unresectable International Federation of Gynecology and Obstetrics (FIGO) stage IIIC/IV ovarian, tubal or peritoneal HGSC. Methods: Multicenter, open-label, non-comparative randomized phase II trial. Pts were randomized (2:1) to receive 4 cycles of CP ± P before IDS. After IDS, all patients received post-operative chemotherapy (2 to 4 cycles) and optional bevacizumab for 15 months in total ± P as maintenance therapy for up to 2 years. Randomization was stratified on center, FIGO stage, Bev planned after IDS and disease volume (<5cm/>5cm). Primary endpoint was the centrally reviewed CRR at IDS. 60 pts were planned in the CP+P arm (A'Hern's single-stage design P0=50%, P1=70%). Safety (particularly due to P addition), surgical morbidity, ORR, PFS and OS were secondary endpoints. Results: 91 pts were randomized from 02/18 to 04/19 with a median Peritoneal Cancer Index at 24 (range 7-39). 80 pts (88%) received Bev in combination with CP followed by bev ± P in maintenance. In the CP+P group (n=61), 58 (95%) pts had IDS and 78% achieved complete resection. The CRR in this group was 74%, statistically superior to the pre-defined hypothesis. In the CP group, CRR was 70% (29/30 pts underwent IDS). Complete resection after strictly 4 cycles of CP±P was obtained for 41 pts (71%) and 17 (58%) pts in CP+P and CP group, respectively (sensitivity analysis). For CP+P group, numerically higher ORRs were observed before IDS compared to CP group (76% vs 61%). Grade ≥3 adverse events (AE) occurred in 75% of the CP+P group and 67% in the CP group: mainly blood and lymphatic, gastrointestinal and vascular disorders. Postoperative AE (mainly infectious, vascular and gastrointestinal) occurred in 20% and 13% of the pts in CP+P and CP arm, respectively. No difference in the number of fatal events between the two arms: 2 in the experimental arm vs 1 in the control arm. Progression free survival rate at 18 months was 61% (95CI% [47-73]) and 57% (95CI% [37-72]) in CP+P and CP arm, respectively. Conclusions: P may be safely added to preoperative treatment in pts deemed non-optimally resectable. The primary objective was met with an improved CRR on CP+P arm. The CRR in the control group was higher than expected. Survival data and translational research including PDL1 status are ongoing to better define P as treatment option in this setting. Clinical trial information: 2016-004-163-39. Clinical trial information: NCT03275506.

Background: We have shown that neoadjuvant carboplatin and paclitaxel (NACP) increased tumor-infiltrating lymphocytes and PDL1 expression in OC pts. Combined PDL1/CTLA4 blockade is active in relapsed OC. INEOV is the 1st trial of neoadjuvant D +/-T with NACP in pts with unresectable OC. Main endpoints are feasibility and safety. 2 endpoints include macroscopically complete resection (CCO) and major pathological response rates after 3 cycles (C3).Methods: Pts with stage IIIC/IV OC were randomized to C1 of NACP alone (immune microenvironment priming) followed by NACP + D (1125mg) at C2&C3 without (arm A) or with T (75mg) at C2 (arm B). Interval debulking surgery (IDS) was planned after C3, or delayed after C6. Major pathological response after C3 was either a chemotherapy response score (CRS) of 3 (Bohm et al) or no viable tumor cells (pCR) at IDS. Data on delayed IDS are not yet available. abstracts Annals of Oncology Volume 32 -Issue S5 -2021 S731