Carfilzomib, lenalidomide, and dexamethasone (KRd) versus bortezomib, lenalidomide and dexamethasone (VRd) in patients with newly diagnosed multiple myeloma (NDMM) – interim results from the randomized Phase III COBRA trial.

Abstract Background: VRd-based regimens are currently the standard for the frontline treatment in NDMM. KRd and KRd-based regimens show high efficacy but in direct comparison in the ENDURANCE randomized phase III trial there was no difference in PFS between KRd and VRd in NDMM. Here we report the pre-specified analysis of the randomized phase III COBRA trial (NCT03729804), evaluating the efficacy and safety of KRd vs VRd in patients with NDMM regardless of eligibility for upfront autologous hematopoietic stem cell transplantation or cytogenetic risks. Methods: Patients with NDMM were randomized 1:1 to receive either KRd or VRd. In the KRd arm, K was given i.v. at 56 mg/m² on days 1, 8, and 15 (20 mg/m² on days 1 cycle 1), R 25 mg orally on days 1–21, and D 40 mg weekly during 1-12 cycles of induction phase, and K 56 mg/m² on days 1 and 15, R 25 mg on days 1–21, and D 20 mg weekly during cycles 13-24 in the maintenance phase for a total of 24 cycles. In the VRd arm, patients received V 1.3 mg/m² s.c. on days 1, 4, 8, and 11; R 25 mg on days 1–14; and D 40 mg on days 1, 2, 4, 5, 8, 9, 11, and 12, for eight 21-day cycles in the induction phase, which followed by sixteen 28-day cycles of Rd maintenance (R 25 mg on days 1–21, D 20 mg weekly), also for a total of 24 months. After maintenance, both arms received R 15 mg daily until progression or intolerance. The co-primary endpoints were minimal residual disease (MRD) negative complete response (CR) after 12 months and progression-free survival (PFS). Secondary endpoints included overall response rate (ORR), overall survival (OS), and safety. MRD was assessed centrally using next-generation sequencing (clonoSEQ, Adaptive Biotechnologies). This primary analysis was conducted after all randomized patients had the opportunity to reach the 12-month MRD assessment. To allow for two co-primary endpoints, a p-value of 0.025 is needed for statistical significance. At this interim analysis, 86 of a projected 106 PFS events have occurred (81%). Based on O'Brien-Fleming type error-spending, the required p-value for this analysis is 0.011. The data cutoff for this analysis was July 25, 2025. All efficacy analyses are in intent-to-treat (ITT) population. Results: Between July 29, 2019, and July 25, 2024, a total of 250 patients were randomized 1:1 to KRd (n = 126) or VRd (n = 124). Baseline patient characteristics were balanced between the two groups regarding median age (66.0 vs 67.0), ISS stage III (26% vs 22%), ECOG performance status 1 or 2 (75% vs 75%) and the presence of high-risk cytogenetic abnormalities (23% vs 23%), defined as presence of del(17p), t(4;14), or t(14;16). After a median follow-up of 35 months (range: 0–66), both co-primary endpoints were met. At 12 months, MRD was measured in 163 patients, of whom 147 had available MRD results (patients without MRD results due to progressive disease, death or other reasons were counted as MRD-positive on ITT). A higher proportion of patients achieved MRD-negative CR at the 10⁻⁵ threshold in the KRd arm compared to the VRd arm [31% vs 18%; OR=2.08, 95%CI: 1.15-3.77); p = 0.016]. This benefit was also observed for MRD-negative CR at the 10-6 threshold (19% vs 7%; OR=3.01, 95%CI: 1.34-6.77, p = 0.008]. PFS was significantly longer with KRd vs VRd treatment [median not reached vs 49 months; HR = 0.57, 95% CI: 0.37–0.88; logrank p = 0.0095]. At the time of data cutoff, 38 patients had died: 20 in the KRd arm and 18 in the VRd. ORR were similar between the arms: 94% (KRd) vs 91% (VRd). However, a significantly greater proportion of patients achieved CR or better in the KRd arm (71% vs 53%; OR = 2.11, 95% CI: 1.26–3.56; p = 0.005). Grade ≥3 adverse events (AEs) were observed in 72% of patients in the KRd arm and 62% in the VRd arm. The most frequent Grade ≥3 hematologic AE was neutropenia, occurring in 21% of patients in the KRd arm and 11% in the VRd arm. The most common non-hematologic Grade ≥3 AE was lower respiratory tract infection (10% in both arms). Peripheral neuropathy (any grade) occurred in 17% and 56% of patients, and cardiac events (any grade) in 18% and 10% of patients in the KRd vs VRd arm, respectively. Two patients in the KRd arm died from treatment-related AEs (pneumonia, COVID-19). Conclusions:The randomized phase III COBRA trial showed superior efficacy of KRd compared VRd in NDMM, with safety profile consistent with prior reports. These findings support further evaluation of KRd-based induction regimens in NDMM.