Abstract GS1-01: KEYNOTE-522 study of neoadjuvant pembrolizumab + chemotherapy vs placebo + chemotherapy, followed by adjuvant pembrolizumab vs placebo for early-stage TNBC: Event-free survival sensitivity and subgroup analyses

Abstract Background: KEYNOTE-522 (NCT03036488) is a phase 3 study of neoadjuvant pembro + chemo vs placebo + chemo, followed by adjuvant pembro vs placebo in patients with early-stage TNBC. The primary analysis showed a statistically significant and clinically meaningful improvement in event-free survival (EFS) with pembro + chemo followed by pembro. To assess the robustness and consistency of the primary EFS result, prespecified sensitivity and subgroup analyses for EFS were performed. Methods: Patients with previously untreated, non-metastatic, centrally confirmed TNBC (stage T1c N1-2 or T2-4 N0-2 per AJCC) were randomized 2:1 to pembro 200 mg Q3W or placebo, both given with 4 cycles of paclitaxel + carboplatin, then with 4 cycles of doxorubicin or epirubicin + cyclophosphamide (neoadjuvant phase). After definitive surgery, patients received pembro or placebo for 9 cycles or until recurrence or unacceptable toxicity (adjuvant phase). Patients were stratified by nodal status (positive or negative), tumor size (T1/T2 or T3/T4), and carboplatin schedule (Q3W or QW). Dual primary endpoints are pCR rate and EFS. Five prespecified sensitivity analyses for EFS were performed, including 2 that assessed the impact of different censoring rules and 3 that assessed the impact of different event definitions. Treatment effects on EFS were examined in prespecified patient subgroups defined by nodal involvement (positive or negative), disease stage (II or III), menopausal status (pre-menopausal or post-menopausal), HER2 status (2+ by IHC but FISH- or 0-1+ by IHC), and LDH (>ULN or ≤ULN). Results: Among 1174 patients randomized, 784 were randomly assigned to the pembro + chemo group and 390 were randomly assigned to the placebo + chemo group. Median follow-up was 39.1 months at the time of the March 23, 2021 data cutoff. The benefit of neoadjuvant pembro + chemo followed by adjuvant pembro vs neoadjuvant chemo alone was generally consistent with the primary EFS results for all five sensitivity analyses and in each subgroup evaluated (Table). Conclusion: EFS sensitivity analyses show a robust treatment benefit of neoadjuvant pembro + chemo followed by adjuvant pembro for previously untreated non-metastatic TNBC. This benefit was generally consistent across a broad selection of patient subgroups. Table. EFS Sensitivity and Subgroup Analyses in KEYNOTE-522EFS Analyses (ITT Population)Pembro + Chemo n/N (%)*Placebo + Chemo n/N (%)*HR (95% CI)†Primary Analysis‡123/784 (15.7)93/390 (23.8)0.63 (0.48 - 0.82)Sensitivity Analyses1. Alternate censoring rules§112/784 (14.3)84/390 (21.5)0.64 (0.48 - 0.84)2. “New anticancer therapy for metastatic disease” considered an EFS event123/784 (15.7)93/390 (23.8)0.63 (0.48 - 0.82)3. “Positive margin at last surgery” removed from EFS definition122/784 (15.6)90/390 (23.1)0.65 (0.50 - 0.85)4. “Positive margin at last surgery” and “second primary malignancy” removed from EFS definition116/784 (14.8)88/390 (22.6)0.63 (0.48 - 0.84)5. “Second breast malignancy” included in EFS definition126/784 (16.1)95/390 (24.4)0.63 (0.48 - 0.82)Subgroup AnalysesNodal involvement‖Positive80/408 (19.6)57/196 (29.1)0.65 (0.46 - 0.91)Negative43/376 (11.4)36/194 (18.6)0.58 (0.37 - 0.91)Overall disease stageII69/590 (11.7)54/291 (18.6)0.60 (0.42 - 0.86)III54/194 (27.8)39/98 (39.8)0.68 (0.45 - 1.03)Menopausal statusPre-menopausal60/438 (13.7)47/221 (21.3)0.62 (0.42 - 0.91)Post-menopausal63/345 (18.3)46/169 (27.2)0.64 (0.44 - 0.93)HER2 status2+ by IHC (but FISH-)32/188 (17.0)24/104 (23.1)0.73 (0.43 - 1.24)0-1+ by IHC91/595 (15.3)69/286 (24.1)0.60 (0.44 - 0.82)LDH>ULN29/149 (19.5)23/80 (28.8)0.65 (0.37 - 1.12)≤ULN93/631 (14.7)69/309 (22.3)0.63 (0.46 - 0.86)*Number of events/total number of patients (%). †Hazard ratios (HR) and 95% CIs in the primary analysis and sensitivity analyses were based on a stratified Cox regression model; analyses in subgroups were based on an unstratified Cox model. ‡EFS was defined as the time from randomization to the time of first documentation of disease progression that precludes definitive surgery, local or distant recurrence, a second primary cancer or death from any cause, whichever occurs first; patients who did not experience an event at the time of data cutoff were censored at the date they were last known to be alive and event-free. §Events after 2 consecutive missed disease assessments or initiation of post-surgery new anticancer therapy were censored at last disease assessment prior to the earlier date of ≥2 consecutive missed disease assessments and initiation of post-surgery new anticancer therapy; if no events before new anticancer therapy, events were censored at last disease assessment before initiation of post-surgery new anticancer treatment. ‖Determined by the study investigator by physical exam, sonography/MRI and/or biopsy. Data cutoff: March 23, 2021. Citation Format: Peter Schmid, Javier Cortes, Rebecca Dent, Lajos Pusztai, Heather McArthur, Sherko Kümmel, Jonas Bergh, Carsten Denkert, Yeon Hee Park, Rina Hui, Nadia Harbeck, Masato Takahashi, Michael Untch, Peter A. Fasching, Fatima Cardoso, Jay Andersen, Debra Patt, Michael Danso, Marta Ferreira, Marie-Ange Mouret-Reynier, Seock-Ah Im, Jin-Hee Ahn, Maria Gion, Sally Baron-Hay, Jean-Francois Boileau, Yalin Zhu, Wilbur Pan, Konstantinos Tryfonidis, Vassiliki Karantza, Joyce O’Shaughnessy. KEYNOTE-522 study of neoadjuvant pembrolizumab + chemotherapy vs placebo + chemotherapy, followed by adjuvant pembrolizumab vs placebo for early-stage TNBC: Event-free survival sensitivity and subgroup analyses [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr GS1-01.