Daratumumab plus bortezomib, lenalidomide, and dexamethasone in transplant-deferred newly diagnosed multiple myeloma: Frailty subgroup analysis of CEPHEUS

Daratumumab, lenalidomide, and dexamethasone (DRd) is approved to treat transplant-ineligible patients with newly diagnosed multiple myeloma (NDMM) based on the phase 3 MAIA trial [ 1 , 2 , 3 ]. Daratumumab, bortezomib, lenalidomide, and dexamethasone (DVRd) is approved to treat patients with NDMM, regardless of autologous stem-cell transplant eligibility, based on the phase 3 PERSEUS [ 4 ] and CEPHEUS [ 5 ] trials. Outcomes in frail patients from CEPHEUS are unknown. The International Myeloma Working Group (IMWG) frailty index, recommended by the IMWG and the European Myeloma Network, incorporates age, comorbidities, and functional status via patient-reported assessments, i.e., the Katz Index of Independence in Activities of Daily Living (ADL) and the Lawton Instrumental Activities of Daily Living (IADL). The latter are often lacking; therefore, the Intergroupe Francophone du Myélome (IFM) simplified frailty index was developed [ 6 , 7 , 8 ] using physician-reported Eastern Cooperative Oncology Group performance status (ECOG PS) instead, and is now broadly incorporated alongside the IMWG index [ 8 ]. CEPHEUS excluded patients who were frail per IMWG criteria but not those categorized as frail per IFM criteria [ 5 ]. Outcome comparisons of differently defined frail patients are scarce [ 9 , 10 ], showing a concordance rate of approximately 75%. A post hoc analysis of MAIA using the IFM index classified approximately 50% of participants as frail and demonstrated a PFS benefit with DRd versus Rd regardless of baseline frailty [ 11 , 12 ]. To inform clinical decision making for treating these patients, we present a subgroup analysis of CEPHEUS comparing DVRd versus VRd across frailty subgroups based on protocol-defined IMWG baseline frailty status and a post hoc analysis based on the IFM index. CEPHEUS has been previously described [ 5 ]. Briefly, CEPHEUS (NCT03652064) is a randomized, phase 3 trial that enrolled patients with NDMM who were transplant ineligible or for whom transplantation was not planned as initial therapy (transplant deferred), who had an ECOG PS score of 0–2 and an IMWG frailty score of 0 or 1. Patients were randomly assigned 1:1 to DVRd or VRd; treatments were previously described [ 5 ] and are summarized in Supplemental Methods. Frailty was assessed using the IMWG and IFM frailty scales. Per protocol, baseline IMWG frailty status was assessed, incorporating patient-reported ADL and IADL assessments. Patients with an IMWG frailty score ≥2 were excluded; patients with a score of 0 or 1 were included and categorized as fit or intermediate-fit, respectively. IFM frailty status was derived post hoc, incorporating baseline physician-reported ECOG PS. Patients with an IFM frailty score <2 were categorized as nonfrail, and ≥ 2, as frail. Owing to differences in frailty score derivation, discrepancies can occur whereby patients classified as nonfrail by IMWG can be classified as frail by IFM, which has been shown in approximately 25% of patients [ 9 , 10 ]. This analysis assessed efficacy, safety, and patient-reported outcomes by treatment and frailty status. IFM nonfrail and IMWG fit were considered the fitter subgroups and IFM frail and IMWG intermediate-fit were considered frailer subgroups. In CEPHEUS, 395 patients were randomized to DVRd ( n = 197) or VRd ( n = 198); median (range) follow-up was 58.7 (0.1–64.7) months. Using the IMWG frailty scale, 256 patients were categorized as fit (DVRd, n = 124; VRd, n = 132) and 139 patients as intermediate-fit (DVRd, n = 73; VRd, n = 66). Using the IFM simplified frailty scale, 297 patients were nonfrail (DVRd, n = 140; VRd, n = 157) and 98 were frail, with the DVRd arm having more frail patients (DVRd, n = 57; VRd, n = 41). As IMWG-defined frail patients were not included, IFM-defined frail patients could not be frail according to IMWG by definition. Of the 98 frail patients, 85% were intermediate-fit according to IMWG. Considering IFM frail and IMWG intermediate-fit as frailer patients, discordance in frailty classification was observed in 24.8% of patients (Supplementary Table 1 ). A higher proportion of the frailer subgroups had ISS stage III disease and, as expected, were older, had poorer ECOG PS, and were transplant ineligible (Supplementary Tables 2 , 3 ); otherwise, baseline characteristics were generally similar between treatment arms across frailty subgroups. Median treatment duration was longer with DVRd versus VRd across frailty groups, but was longer in fitter versus frailer patients (Supplemental Results). Measurable residual disease (MRD)-negativity ≥CR rates were consistently higher with DVRd versus VRd (10 -5 and 10 -6 ) across frailty subgroups (Fig. 1 ), with similar results for patients achieving ≥12-month and ≥24-month sustained MRD-negativity (Supplementary Figs. 1 , 2 ). Compared with VRd, DVRd improved PFS (Fig. 2 ) and led to a higher ≥CR rate across all subgroups (Supplementary Fig. 3 ). Although CEPHEUS was not powered for overall survival, a trend favored DVRd in all subgroups (Supplementary Table 4 ). Aligned with the primary analysis, the treatment effect of DVRd strengthened in all subgroups when censoring for COVID-19 deaths (Supplementary Table 5 ). Fig. 1: MRD-negative ≥ CR rates in the ITT population. Full size image A MRD assessed at 10 −5 in IMWG frailty groups. B MRD assessed at 10 -5 in IFM simplified frailty groups. C MRD assessed at 10 −6 in IMWG frailty groups. D MRD assessed at 10 -6 in IFM simplified frailty groups. a P values are from Fisher’s exact test. CR complete response, DVRd daratumumab plus bortezomib, lenalidomide, and dexamethasone, IFM Intergroupe Francophone du Myélome, IMWG International Myeloma Working Group, ITT intention-to-treat, MRD measurable residual disease, VRd bortezomib, lenalidomide, and dexamethasone. Fig. 2: Progression-free survival in the ITT population. Full size image A IMWG frailty scale. B IFM simplified frailty scale. a HR and 95% CI are from a Cox proportional hazards model with treatment as the sole explanatory variable. A hazard ratio < 1 indicates an advantage for DVRd. P value is based on the unstratified log-rank test. CI confidence interval, DVRd daratumumab plus bortezomib, lenalidomide, and dexamethasone, HR hazard ratio, IFM Intergroupe Francophone du Myélome, IMWG International Myeloma Working Group, int-fit intermediate-fit, ITT intention-to-treat, NR not reached, VRd bortezomib, lenalidomide, and dexamethasone. Health-related quality of life data are shown in Supplementary Table 6 and Supplementary Fig. 4 . Rates of Grade 3/4 and serious TEAEs were higher in both treatment arms in the intermediate-fit IMWG and frail IFM subgroups (Supplementary Tables 7 – 10 ). The most common serious TEAEs were infections, primarily pneumonia or COVID-19. In IMWG frailty subgroups, rates of serious infections in fit patients were 35.5% (DVRd) and 32.6% (VRd), and in intermediate-fit patients, 46.6% and 41.3%; in IFM frailty subgroups, rates in nonfrail patients were 36.4% (DVRd) and 30.8% (VRd), and in frail patients, 47.4% and 53.8%. Grade 5 TEAEs are described in Supplemental Results. TEAEs led to discontinuation of all study drugs at lower rates with DVRd versus VRd across frailty subgroups, and in fitter patients versus frailer patients within each treatment arm (Supplementary Table 11 ). Bortezomib discontinuation rates due to TEAEs were similar by treatment in the fitter subgroups (IMWG fit, 14.5% [DVRd] and 15.2% [VRd]; IFM nonfrail, 12.9% and 14.1%) and lower with DVRd in the frailer subgroups (IMWG intermediate-fit, 9.6% and 19.0%; IFM frail, 12.4% and 25.6%). Across frailty subgroups, bortezomib discontinuation rates due to peripheral sensory neuropathy were lower with DVRd versus VRd. Although frailer patients in both treatment arms had higher rates of Grade 3/4 TEAEs and generally had shorter treatment durations compared with fitter patients, frailer patients receiving DVRd experienced decreased serious infections, bortezomib discontinuations, and Grade 5 non-COVID-19 TEAEs. No detriments to health-related quality of life were observed across frailty subgroups in both treatment arms. TEAEs were generally consistent with the known safety profiles; across frailty subgroups, treatment duration was longer and TEAEs leading to treatment discontinuation were lower with DVRd versus VRd. These results are reassuring, given the risks of infection and peripheral sensory neuropathy with bortezomib that older and frailer patients face [ 8 , 13 ]. While we cannot entirely exclude an indirect role of daratumumab enabling better bortezomib management, there is no current evidence of a direct neuroprotective effect. These CEPHEUS analyses add to the previously reported post hoc MAIA frailty subgroup analysis [ 11 ]. Although both studies enrolled patients with NDMM, key differences include the lack of IMWG frailty data in MAIA, which was designed prior to publication of the IMWG frailty index and did not assess ADL and IADL scores – necessary to determine IWMG frailty status [ 11 ] – and population differences [ 1 , 5 , 11 ]. While CEPHEUS demonstrated clinical benefit with DVRd across frailty subgroups, DRd remains an appropriate option for patients who are older or unable to tolerate bortezomib whereas DVRd may confer efficacy benefits over DRd, as shown in a recent indirect treatment comparison where improvements in MRD-negative CR rate were significantly greater for DVRd versus DRd (adjusted OR: 3.04 [95% CI 2.01–4.61]) [ 14 ]. A limitation of the current analysis is that CEPHEUS excluded frail patients per IMWG criteria or those aged >80 years, warranting additional prospective studies. Approximately 25% of the population were frail based on IFM criteria, which are widely applied in clinical research, as ADL and IADL (required to determine IMWG frailty scores) are often lacking [ 13 ]. Limited information suggest the IFM index may overestimate frailty [ 9 ]. In this trial, using frail and nonfrail categorization (frail [score ≥2], nonfrail [score 0/1]), discordance between IMWG and IFM frailty classifications was observed in 24.8% of patients. While by design, CEPHEUS does not include IMWG frail patients, this may be most consistent with a systematic shift and suggests that the IFM scale may overestimate frailty relative to IMWG in this dataset. Importantly, at a more granular breakdown (score 0, 1, ≥2), the discordance was mostly confined to adjacent categories and did not materially affect treatment benefit. However, efficacy and feasibility of DVRd in patients defined as frail by IMWG remain to be established. Treatment options for transplant-ineligible patients with NDMM now include regimens such as DVRd and DRd that lead to significantly better outcomes [ 2 , 5 ], allowing treatments to be tailored to individual patients, with decisions balancing optimized efficacy with risk and frailty. This ability to individualize treatment based on age and frailty is reflected in NCCN Guidelines [ 15 ]. This analysis demonstrated the benefit of DVRd versus VRd across frailty subgroups, supporting DVRd as a standard of care. These findings, along with known DRd benefits in frailty subgroups [ 11 , 12 ], help to inform daratumumab selection in patients with NDMM who do not receive frontline transplant, with the ability to tailor the regimen based on patient fitness, bortezomib eligibility, treatment goals, and treatment access.