Daratumumab plus lenalidomide and dexamethasone in transplant-ineligible newly diagnosed multiple myeloma: frailty subgroup analysis of MAIA

In the phase 3 MAIA study of patients with transplant-ineligible newly diagnosed multiple myeloma (NDMM), daratumumab plus lenalidomide/dexamethasone (D-Rd) improved progression-free survival (PFS) versus lenalidomide/dexamethasone (Rd). We present a subgroup analysis of MAIA by frailty status. Frailty assessment was performed retrospectively using age, Charlson comorbidity index, and baseline Eastern Cooperative Oncology Group performance status score. Patients were classified as fit, intermediate, non-frail (fit + intermediate), or frail. Of the randomized patients (D-Rd, n = 368; Rd, n = 369), 396 patients were non-frail (D-Rd, 196 [53.3%]; Rd, 200 [54.2%]) and 341 patients were frail (172 [46.7%]; 169 [45.8%]). After a 36.4-month median follow-up, non-frail patients had longer PFS than frail patients, but the PFS benefit of D-Rd versus Rd was maintained across subgroups: non-frail (median, not reached [NR] vs 41.7 months; hazard ratio [HR], 0.48; P < 0.0001) and frail (NR vs 30.4 months; HR, 0.62; P = 0.003). Improved rates of complete response or better and minimal residual disease (10–5) negativity were observed for D-Rd across subgroups. The most common grade 3/4 treatment-emergent adverse event in non-frail and frail patients was neutropenia (non-frail, 45.4% [D-Rd] and 37.2% [Rd]; frail, 57.7% and 33.1%). These findings support the clinical benefit of D-Rd in transplant-ineligible NDMM patients enrolled in MAIA, regardless of frailty status.

With longer treatment duration (36.4-month median follow-up), D-Rd continued to demonstrate a PFS benefit and deeper responses [18]. The median age was 73 years, and 43.6% of patients were aged ≥75 years [18]. In both the primary and updated analyses, D-Rd improved PFS, even in patients aged ≥75 years [14,18].
Although D-Rd improved outcomes in older patients, such patients often vary widely in fitness level [19,20]. The ability or inability to tolerate cancer treatment regimens logically impacts clinical outcomes and is dependent on overall health as determined by the functional status of numerous organ systems [20,21]. Therefore, analyses of a treatment based on frailty status should be more informative than analyses based solely on age. A scale) assessments [20]. However, use of the scoring system was not feasible with the MAIA study, as the MAIA study did not assess patients using ADL and IADL scales.
Similarly, the FIRST trial did not assess patients using ADL and IADL scales, which led to the development of a frailty scale based on age, CCI (using medical history of patients), and the physicianevaluated Eastern Cooperative Oncology Group performance status (ECOG PS) score in a retrospective subgroup analysis of the trial [19]. The frailty scale, similar to the IMWG scale, allows classification of patients into fit, intermediate, and frail subgroups; the 3-subgroup frailty classification was also used in a frailty subgroup analysis of the A.R.R.O.W. study [22]. Use of the frailty scale was further simplified to classify patients into only 2 subgroups-frail and non-frail [19]. Both 3-subgroup and simplified 2-subgroup frailty classifications were shown to be predictive of clinical outcomes in transplant-ineligible NDMM patients [19,20,23].
We present a subgroup analysis of MAIA comparing D-Rd versus Rd across frailty subgroups based on the 3-subgroup and simplified 2-subgroup frailty classifications [19].

PATIENTS AND METHODS Study design and patients
MAIA (ClinicalTrials.gov Identifier: NCT02252172) is a randomized, openlabel, phase 3 trial. The study was conducted in accordance with the principles of the Declaration of Helsinki and the International Conference on Harmonisation Good Clinical Practice guidelines. Independent ethics committees or institutional review boards at each institution approved the study protocol. All patients provided written informed consent.
The complete methodology of MAIA has been previously described [14]. Briefly, patients with documented NDMM ineligible for high-dose chemotherapy with autologous stem cell transplant due to age ≥65 years or comorbidities, an ECOG PS score ≤2, and a creatinine clearance (CrCl) ≥30 mL/min were eligible.

Treatment
Patients (N = 737) were randomized 1:1 to D-Rd or Rd; randomization was stratified by International Staging System (ISS) disease stage (I vs II vs III), geographic region (North America vs other), and age (<75 vs ≥75 years). During each 28-day cycle, all patients received lenalidomide 25 mg (10 mg recommended if CrCl 30-50 mL/min) orally on Days 1-21 and dexamethasone 40 mg (20 mg if aged >75 years or body mass index <18.5 kg/m 2 ) orally on Days 1,8,15,and 22. Patients in the D-Rd cohort received daratumumab 16 mg/kg intravenously once weekly during Cycles 1-2, every 2 weeks during Cycles 3-6, and then every 4 weeks thereafter. Treatment in both cohorts continued until disease progression or unacceptable toxicity.

Frailty evaluation
Frailty assessment was performed retrospectively on all patients using age, CCI (based on retrospective review of each patient's medical history), and baseline ECOG PS score (Supplementary Table 1) [19]. Frailty scores were used to classify patients into fit (0), intermediate (1), or frail (≥2) subgroups. Frailty status was further simplified into 2 categories: total-non-frail (0-1; a combination of the fit and intermediate subgroups) and frail (≥2). Patients within the total-non-frail and frail subgroups were further divided by ISS stage (I/II vs III). Patients with missing data were excluded from frailty evaluation.

Assessments and statistical analyses
The primary endpoint was PFS. Post-hoc analyses were performed by patient frailty status. Efficacy endpoints were assessed based on the intentto-treat population (all randomized patients). Safety was assessed in the safety population (patients who received ≥1 dose of study treatment). See Supplementary Information for details on statistical analyses. In the frail subgroup, 4 patients randomized to D-Rd and 3 patients randomized to Rd did not receive treatment. Demographics and baseline characteristics were generally balanced between the treatment cohorts within each frailty subgroup ( Table 1).

Patient disposition and treatment
The dispositions of patients according to frailty status are summarized in Table 2. For both the D-Rd and Rd cohorts, the proportion of patients who discontinued treatment was highest in the frail subgroup; the proportion was lower in the D-Rd cohort versus the Rd cohort across frailty subgroups. Among patients randomized to D-Rd, a higher proportion of patients discontinued treatment during the first 12 months of treatment in the frail subgroup versus other frailty subgroups (Table 2). Overall, the 2 most common reasons for treatment discontinuation with D-Rd and Rd in all frailty subgroups were progressive disease and adverse event (AE). Median (range) time to treatment discontinuation with D-Rd and Rd was 22.    Table 3). The median cumulative dose of lenalidomide was 525 mg in each of the first 6 cycles with D-Rd and Rd for all frailty subgroups except for the D-Rd cohort in the frail subgroup (Cycle 1, 338 mg; Cycle 2, 315 mg; Cycle 3, 300 mg; Cycle 4, 308 mg; Cycles 5 and 6, 210 mg) and the Rd cohort of the frail subgroup (Cycle 2, 513 mg).  Fig. 1A and B). The 36-month PFS rate was higher in the D-Rd cohort in all subgroups, with decreasing rates from fit to frail (fit, D-Rd, 78.3% vs Rd, 53.6%; intermediate, 70.4% vs 51.7%; total-non-frail, 73.2% vs 52.1%; frail, 61.5% vs 39.5%). In the total-non-frail and frail subgroups subdivided by ISS stage (I/II vs III), the PFS benefit of D-Rd versus Rd was also maintained in most subgroups, with the exception of frail + ISS III ( Fig. 2A and B). Regardless of lenalidomide starting dose, the PFS benefit of D-Rd versus Rd was observed in the intent-to-treat population ( Higher overall response rates (ORRs) were achieved with D-Rd versus Rd across frailty subgroups, with the total-non-frail subgroup achieving higher ORRs compared with the frail subgroup in each treatment cohort (fit, 100.0% vs 83.3%; P = 0.0004; intermediate, 96.9% vs 85.2%; P = 0.0012; total-non-frail, 98.0% vs 84.5%; P < 0.0001; frail, 87.2% vs 78.1%; P = 0.0265). Higher ≥CR rates and MRD-negativity (10 -5 sensitivity threshold) rates were achieved with D-Rd versus Rd across frailty subgroups; total-non-frail patients had higher MRD-negativity rates than frail patients in the D-Rd cohort (33.2% vs 23.8%, respectively), and in the Rd cohort, total-non-frail and frail patients had similarly low MRD-negativity rates (8.5% vs 10.1%;

DISCUSSION
After >3 years of follow-up, D-Rd demonstrated improved efficacy versus Rd in transplant-ineligible NDMM patients, regardless of frailty status. Compared with the total-non-frail subgroup, patients in the frail subgroup had poorer outcomes in both treatment cohorts. Nevertheless, D-Rd reduced the risk of disease progression or death by 52% in total-non-frail patients and by 38% in frail patients. The PFS results demonstrated that D-Rd leads to outcomes in frail patients that are at least as good as those observed with Rd in fit patients. Median PFS was NR in the D-Rd cohort for any frailty subgroup, whereas the Rd cohort did reach this milestone in both the total-non-frail (41.7 months) and frail (30.4 months) subgroups. Importantly, a greater PFS benefit of D-Rd over Rd was seen in total-non-frail and frail patients with lower ISS disease stage (I/II) and total-non-frail patients in the ISS stage III category. A greater PFS benefit of D-Rd over Rd was also seen in total-non-frail and frail patients who received a lenalidomide starting dose of <25 mg, with the benefit less pronounced in frail patients. Regardless of frailty status, deep responses were achieved with D-Rd versus Rd, with improved rates of ≥CR and MRD negativity. Consistent with the findings in the FIRST trial [19], the use of the ECOG PS score-containing frailty scale predicted clinical outcomes in transplant-ineligible NDMM  T. Facon et al. patients, with frail patients demonstrating worse prognosis in terms of PFS and response rates versus total-non-frail patients. The safety profile of D-Rd in frailty subgroups was generally consistent with the overall population of MAIA [14]; although higher rates of grade 3/4 neutropenia and pneumonia were observed with D-Rd in the frail subgroup than in the total-non-frail subgroup, these events were clinically manageable. The frail subgroup had an increased incidence of hematologic and non-hematologic grade 3/4 TEAEs, serious TEAEs, and deaths in both treatment cohorts versus the total-non-frail subgroup, but this was not unexpected based on the additional comorbidities frequently associated with frailty. Among patients in the frail subgroup, a higher incidence of grade 3/4 neutropenia was observed with D-Rd versus Rd. Across frailty subgroups, the incidences of treatment discontinuation overall and due to AEs were higher with Rd versus D-Rd, while the incidences of lenalidomide dose modifications overall and due to AEs were higher with D-Rd versus Rd. These findings may indicate that clinicians were more likely to modify the dose of lenalidomide due to AEs, such as neutropenia, with D-Rd versus Rd, as patients in the D-Rd cohort were also receiving daratumumab. Although there was no clear association observed between a CCI ≥1 and higher rates of grade 3/4 TEAEs or serious TEAEs in frail patients, a greater proportion of deaths occurred within 60 and 90 days of receipt of the first dose of study treatment in the frail subgroup versus other frailty subgroups, and almost all of these deaths were due to AEs; overall, the TEAE with an outcome of death observed most frequently in the frail subgroup was pneumonia (D-Rd, 1.2%; Rd, 1.8%). Consistent with the increased incidence of grade 3/4 TEAEs in the frail subgroup versus other frailty subgroups in both treatment cohorts, the frail subgroup had a shorter duration of treatment and a higher frequency of treatment discontinuations. The median RDI of lenalidomide was lower with D-Rd versus Rd in all frailty subgroups; this difference was most pronounced in the frail subgroup. The median RDI of daratumumab was nearly identical across frailty subgroups. A reduced starting dose of lenalidomide (<25 mg) was given more frequently to daratumumab-treated patients in all frailty subgroups, with the highest frequency reported in the frail subgroup. Growth factors were used most commonly in the frail subgroup and were more commonly used with D-Rd versus Rd across frailty subgroups.
In a separate analysis of MAIA age subgroups, D-Rd reduced the risk of disease progression or death by 37% in patients aged ≥75 years and by 50% in patients aged <75 years, similar to results reported in the frail and total-non-frail subgroups [24]. Thus, the results of the current MAIA frailty subgroup analysis combined with the MAIA age subgroup analysis highlight the key role D-Rd can play as first-line treatment in transplant-ineligible NDMM patients. In the real-world treatment of NDMM, each additional line of therapy is associated with worse outcomes [25]. In transplant-ineligible NDMM patients, attrition was found to be as high as 50% per line of therapy, with the high attrition level associated with older age and poor comorbidity status [26]. These data suggest that the most effective treatment regimen should be provided upfront, as frail patients may not have the opportunity to be treated with additional lines of therapy later.
A frailty subgroup analysis using the same frailty scale as our study was conducted on the phase 3 ALCYONE study of daratumumab plus bortezomib/melphalan/prednisone (D-VMP) versus bortezomib/melphalan/prednisone (VMP) [27]. After a 40.1-month median follow-up, the overall survival (OS) and PFS benefit of D-VMP versus VMP was observed in all frailty subgroups. With OS data not yet mature at the time of this analysis, the effect of frailty on OS remains to be seen in MAIA. Patients in ALCYONE in the D-VMP cohort received single-agent daratumumab starting in Cycle 10; thus, a better safety profile in frail daratumumab-treated patients was observed in ALCYONE compared with in MAIA. The results of these frailty subgroup analyses of MAIA and ALCYONE support the use of daratumumabbased regimens in transplant-ineligible NDMM patients.
This study provides validation of the simplified frailty score implemented in the FIRST trial [19]. The retrospective assessment of frailty score was a limitation of this study. Retrospective CCI calculations were based on reported medical history, which may contain missing data and result in underestimating or overestimating the number of patients in each frailty subgroup. Additionally, the ECOG PS score parameter used for frailty score calculations in our study is more subjective, with susceptibility to intra-and inter-observer bias, compared with the ADL and IADL scales used in the IMWG scoring system [28,29]. Furthermore, while the frailty scale used in our study is based on parameters that are routinely assessed in clinical practice and is therefore practical for clinical use, the use of comprehensive frailty assessments that more accurately reflect biological or functional frailty will remain important for the further optimization of treatment strategies for frail patients [29]. Finally, patients with an ECOG PS score ≥3 and patients with comorbidities that may interfere with the study procedures were excluded from MAIA; the inclusion and exclusion criteria for the study limits the generalizability of these results to more frail patients seen in clinical practice.
In conclusion, improved efficacy with D-Rd versus Rd was observed across frailty subgroups, consistent with the overall study population. Our findings, although based on a retrospective assessment of frailty, support the clinical benefit of D-Rd in patients with transplant-ineligible NDMM enrolled in MAIA, regardless of frailty status.

DATA AVAILABILITY
The data sharing policy of Janssen Pharmaceutical Companies of Johnson & Johnson is available at https://www.janssen.com/clinical-trials/transparency. As noted on this site, requests for access to the study data can be submitted through Yale Open Data Access (YODA) Project site at http://yoda.yale.edu.