Introduction

Each year in Poland, over 2000 patients are diagnosed with multiple myeloma, making it one of the most common hematologic malignancies. According to some reports, patients with multiple myeloma experience more symptoms than those with any other hematologic cancer due to the characteristic pattern of end-organ damage, including bone lesions and neuropathy.1 Therefore, despite improvements in survival rates, multiple myeloma continues to pose a significant burden on the health care system with respect to morbidity and deleterious effects on health-related quality of life (HRQoL) for patients living with an incurable cancer.2-4 Importantly, symptom intensity decreases and self-reported HRQoL increases after implementation of effective antimyeloma induction therapy.5,6 However, when disease-related symptoms are controlled, treatment-related toxicities often emerge, such as peripheral neuropathy from cumulative exposure to bortezomib.7 These treatment-emergent toxicities are particularly significant in the modern paradigm of multiple myeloma treatment, given that continuous therapy is routinely used in both transplant-eligible and transplant-ineligible patients, with effects on both short- and long-term symptom burden and HRQoL.8

For transplant-eligible patients, indefinite maintenance with single-agent lenalidomide is a well-established standard of care with respect to post-transplant therapy, with multiple lines of evidence proving its overall survival benefit over observation.9 Moreover, thanks to the convenience of this single-drug oral maintenance paired with a manageable toxicity profile, there have been no signals that this therapy negatively affects HRQoL.5,10 This combination of not only efficacy but also tolerability sets a high benchmark for other maintenance approaches to meet and potentially replace the current standard of care.

We have recently published an interim analysis of our investigator-initiated phase 3 ATLAS study that enrolled patients with multiple myeloma who had undergone autologous stem cell transplantation (ASCT). The patients were randomized to maintenance therapy with either a KRd (carfilzomib [Kyprolis; Amgen, Thousand Oaks, California, United States], lenalidomide [Revlimid; Bristol-Myers Squibb / Celgene, New York, New York, United States], dexamethasone) or an R (lenalidomide [Revlimid]) regimen.11 The results suggest a progression-free survival (PFS) benefit observed in the KRd arm (median PFS, 59.1 vs 33.8 months, respectively, for KRd and R arms; hazard ratio, 0.51). Similar findings were reported in the phase 2 FORTE trial,12 which compared lenalidomide alone with carfilzomib and lenalidomide maintenance. Acknowledging the fact that the triplet therapy, including intravenous carfilzomib, may be less convenient for the patients than single-agent oral lenalidomide, the ATLAS study design included de-escalation of the therapy in the KRd arm to lenalidomide alone for patients who achieved minimal residual disease (MRD) negativity after 6 cycles of treatment. Of note, the observed frequency of adverse events was similar between the 2 arms. There was only slightly higher frequency of grade 3 and 4 lower respiratory tract infections in the treatment arm (8% vs 1%). However, regardless of the comparable safety profile, HRQoL can be negatively affected by other factors, including the burden on social functioning, ability to return to work, and time or financial toxicity. Therefore, it is utterly important to understand the patients’ perspective by analyzing patient-reported outcomes (PROs), particularly in the context of more intensive maintenance regimens. Here, we present an analysis of PRO data in the form of self-assessment questionnaires measuring HRQoL completed longitudinally by Polish patients participating in the ATLAS trial.

Patients and methods

Patients and study design

ATLAS (NCT02659293) is an ongoing, open-label, randomized, phase 3 trial comparing post-ASCT treatment with R vs MRD-directed, risk-adapted KRd regimen. The study was approved by the appropriate institutional review board or ethics committee and the Office for Registration of Medicinal Products, Medical Devices and Biocidal Products in Poland (88216/05.10.2016). The study is conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization of Good Clinical Practice guidelines, and all patients provided written informed consent to participate.

The design of the ATLAS trial has been previously described in detail.11 Patients with newly diagnosed multiple myeloma that completed any induction therapy followed by a single ASCT were eligible for enrollment up to 100 days after the ASCT. This analysis includes only the patients enrolled at 7 Polish sites where HRQoL questionnaires were collected, excluding 21 patients treated in the United States and 25 patients treated at 2 other Polish sites.

Treatment

The patients were randomized in a 1:1 ratio to either R or KRd. The patients included in the KRd group received treatment in 28-day cycles (carfilzomib in cycle one, days 1 and 2, at 20 mg/m2, and on subsequent days and cycles, 36 mg/m2 administered on days 1, 2, 8, 9, 15, and 16 of cycles 1–4, and days 1, 2, 15, and 16 from cycle 5 and beyond; lenalidomide 25 mg on days 1–21; and dexamethasone 20 mg on days 1, 8, 15, and 22). The patients in the KRd arm who had no detectable MRD as defined by the International Myeloma Working Group (at least 10-5 sensitivity and at least in a complete response)13 after cycle 6 and had no high-risk cytogenetic features were de-escalated to single-agent lenalidomide (at a best-tolerated dose, up to 15 mg) starting with cycle 9. Patients included in the R maintenance group received single-agent lenalidomide at 10 mg for the first 3 cycles and then at a best-tolerated dose up to 15 mg for 28 days in 28-day cycles. After 36 cycles of treatment, the patients in both arms received or continued to receive R maintenance.

Health-related quality of life questionnaires

Functional Assessment of Cancer Therapy / Gynecologic Oncology Group – Neurotoxicity (FACT / GOG-NTX)14 questionnaires were completed by the patients, prior to any other study-related procedure at screening (approximately 3 months post-transplant) and then every 6 months (cycles 6, 12, 18, 24, 30, and 36). Collection of HRQoL self-assessments was not required by the ATLAS trial protocol, but the questionnaires were collected per local standards among patients treated at 7 of the Polish sites involved in this study. Only the patients actively receiving treatment were asked to complete the HRQoL questionnaires.

The FACT / GOG-NTX questionnaire evaluates 38 items across 5 different domains: physical well-being (PWB), 7 items; social well-being (SWB), 7 items; emotional well-being (EWB), 6 items; functional well-being (FWB), 7 items; and neurotoxicity (Ntx), 11 items. For all evaluations, higher scores indicate better quality of life. We utilized the original scoring system of the FACT / GOG-NTX self-assessment tool, which allows for adjustments in the case of missing item responses. Reliability of the analyses was assessed with Cronbach α, which equaled 0.9 for PWB, 0.85 for SWB, 0.69 for EWB, 0.86 for FWB, and 0.9 for Ntx. Detailed values for different time points in both study arms are presented in Supplementary material, Table S1. Several total summary scores can also be calculated from the FACT / GOG-NTX questionnaire, including the FACT / GOG-NTX Trial Outcome Index (TOI, combining PWB, FWB, and Ntx), FACT-G Total Score (combining PWB, SWB, EWB, and FWB), and FACT-G Total NTX (combining all 5 domains).

For evaluation of clinically meaningful changes in the instrument scores, we defined minimally important differences (MIDs) as a change of at least 10% of the instrument range, based on previous studies.15 The MID thresholds for PWB, SWB, EWB, FWB, and Ntx were 2.8, 2.8, 2.4, 2.8, and 4.4, respectively. Regarding the total summary scores, the MID thresholds for TOI, FACT-G Total Score and FACT-G Total NTX were 10.0, 10.8, and 15.2, respectively.

Statistical analysis

Continuous variables, including HRQoL scores, are presented as medians with interquartile ranges (IQRs), whereas categorical variables are presented as numbers and percentages. Changes from baseline were calculated for every follow-up time point separately. Since different numbers of questionnaires were collected at each time point, repeated measures analysis was not performed. Tests for independent samples were employed, using the t test for comparisons involving normally distributed continuous variables and the Mann–Whitney test for other continuous variables. Correction for multiple testing was not applied in the presented results. The χ2 test was used to assess differences in the distribution of categorical variables, including the proportion of patients experiencing an increase or decrease in HRQoL that reached the aforementioned MID thresholds. Statistical analysis was performed using GraphPad Prism version 10.1 (GraphPad, San Diego, California, United States) and R software version 4.3.0 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Out of 134 patients treated at the 7 sites included in this analysis, FACT / GOG-NTX questionnaires were completed at least once by 123 patients (92%). HRQoL data were collected up to the previously reported cutoff date of December 31, 2021.11 The median (IQR) follow-up time from enrollment for patients included in this analysis was 33.1 (20.1-39.1) months. Women comprised 46% of the study population, median (IQR) age was 56 (47–61) years, all participants were white, and 63% had an Eastern Cooperative Oncology Group (ECOG) performance status of 1. Baseline characteristics of the patients from the 2 study arms are presented in Supplementary material, Table S2. The patients in both arms were balanced in terms of age, ECOG status, and baseline disease characteristic. Disease progression during the study was reported in 15 out of 66 patients (23%) in the KRd arm, and 20 out of 57 patients (35%) in the R arm. Rates of HRQoL questionnaire completion were similar for patients in the KRd and R arms and equaled 85% (out of all evaluable patients) for all the analyzed time points combined (Supplementary material, Table S3). The patients who discontinued treatment for any reason before reaching the time of HRQoL evaluation were excluded from the appropriate denominators.

At screening, median (IQR) domain scores for EWB (22 [18–25] for KRd vs 21 [19–25] for R), PWB (20 [17–25] vs 20 [16–24.8]), SWB (22 [19–23] vs 21 [18–22], FWB (20 [18–25] vs 19.5 [17.3–22]), and Ntx (34 [29–38] vs 35.5 [25.3–39]) as well as the combined scores were similar for both study arms (Supplementary material, Figure S1). The longitudinal changes in the scores for the analyzed domains were also not different between the 2 arms. For the most domains, the scores were numerically the highest at screening. The exceptions included PWB and FWB, which reached the highest scores in the KRd arm after cycle 6 (median [IQR], 22 [18–26]) and cycle 30 (21 [18.8–25.3]), respectively, while FWB and SWB peaked after cycle 12 (20 [17.3–24]) and 36 (21 [18–24]) in the R arm, respectively. For the total summary scores, the only exception was noted for TOI, which reached the highest value after cycle 6 in the KRd arm. The lowest scores were observed later in the course of therapy: for PWB after cycle 30 in both the KRd (20.5 [14.8–25.2]) and R (19 [16–24]) arms, for SWB after cycle 24 (20 [16–22]) in the KRd arm and after cycle 30 (20 [17.5–24]) in the R arm, for EWB after cycle 36 (20 [15.5–25]) in the KRd arm and after cycle 24 (20 [17–23]) in the R arm, for FWB after cycle 12 (20 [17–24]) in the KRd arm and after cycle 30 (18 [14.3–24.8]) in the R arm, and for Ntx after cycle 36 (32 [23.5–41]) in the KRd arm and after cycle 24 (26.5 [23.8–36.3]) in the R arm. In the KRd arm all combined scores reached the lowest values after cycle 18, and in the R arm after cycle 24. Notably, the differences between the lowest and highest scores at different time points were relatively small, as compared with the instrument ranges. The values are summarized in Supplementary material, Table S4.

Similar patterns were observed when comparing changes in scores from baseline at different time points (Figure 1). For Ntx, EWB, and SWB, the scores in both arms were numerically lower than at the baseline, although these differences were nonsignificant. For FWB and PWB, the values remained stable throughout the analyzed time points. Importantly, there were no significant differences between the study arms at any of the analyzed time point in any of the analyzed domains or combined scores (Supplementary material, Table S5).

Figure 1. Changes from baseline (assessed at screening) in the evaluated domains of the Functional Assessment of Cancer Therapy / Gynecologic Oncology Group – Neurotoxicity questionnaire at different time points. Data are presented as medians with interquartile ranges.

The improvements and deteriorations above the MID threshold were also not significantly different between the study arms (Supplementary material, Figure S2). The rates of improvement in different domains for the KRd arm and the R arm were as follows: 31% vs 27% in PWB, 7% vs 15% in SWB, 31% vs 29% in EWB, 34% vs 42% in FWB, 25% vs 23% in Ntx, 15% vs 21% in TOI, 19% vs 25% in FACT-G Total, and 15% vs 12% in FACT-G Total NTX. Respectively, the rates of deterioration equaled 36% vs 38% in PWB, 51% vs 40% in SWB, 47% vs 54% in EWB, 44% vs 40% in FWB, 41% vs 48% in Ntx, 32% vs 38% in TOI, 39% vs 35% in FACT-G Total, and 41% vs 37% in FACT-G Total NTX. Notably, except for FWB in the R arm, in both groups for all analyzed domains, the percentage of patients experiencing deterioration was numerically higher than the rate of patients reporting significant improvement in their HRQoL. It was particularly evident for SWB and EWB, although none of the differences was statistically significant.

Discussion

This study offers important insights into the impact of 2 distinct maintenance strategies on HRQoL in patients with multiple myeloma. Although the analysis was not preplanned to formally compare PROs, the adherence among the eligible population was high throughout the analyzed time points, and the presented results do not reveal any alarming signals regarding the negative impact of the more intensive triplet maintenance on patient HRQoL.

It is important to emphasize that the baseline evaluations in this analysis were conducted at the time of the screening visit, which, as outlined in the ATLAS trial protocol, took place 3 months post-ASCT. Previous studies have shown that in transplant-eligible patients, HRQoL scores tend to improve within the first months of induction therapy, as compared with the time of diagnosis.5,16 Subsequently, ASCT is associated with a brief worsening of HRQoL due to the toxic effects of high-dose alkylating therapy and hospitalization. However, this decline is transient, and by the 3-month mark, most patients report full recovery and even better QoL than immediately before the transplant.17 Using this time point as a baseline in the present analysis may explain why we did not observe a trend for improvement with time in the analyzed domains, regardless of the study arm. This is in line with previous reports on HRQoL in 169 patients receiving lenalidomide maintenance, based on the registry data from the United States, as well as a recently published analysis of more than 750 patients treated in the BMTCTN0702 (STaMINA) trial.10,18 Moreover, the baseline values for SWB, EWB, FWB, and FACT-G Total Score reported in our analysis are numerically higher than normative data of the general United States population.19

One of the concerns regarding the KRd maintenance strategy used in the ATLAS trial is that it requires up to 3 years of more intensive treatment. To that point, it is important to note that the longitudinal results of PROs, as presented, remained stable throughout the observed period with almost 3 years of follow-up data available for most participants. The study design allowed for treatment de-escalation after 8 cycles of therapy in the subset of MRD-negative, standard-risk patients, and indeed, 29 out of 66 patients from the KRd arm included in this analysis de-escalated the therapy to lenalidomide alone. Due to a low number of patients in this subgroup, we did not present their results separately.

While the progress achieved in myeloma therapy is remarkable, it is crucial to bear in mind that patients, even when in deep response, still endure the burden of both the disease and its treatment. This has been demonstrated previously and is reaffirmed in the current analysis.20 The rates of deterioration reaching the MID threshold were relatively more common than improvements of the same magnitude. This trend remained consistent across both study arms and at all analyzed time points. Of note, this HRQoL deterioration was particularly evident concerning social and emotional well-being, underscoring an important factor that is often overlooked by health care professionals. These results strongly encourage future efforts to support myeloma patients on maintenance therapy with appropriate psychological care.

A notable limitation of this analysis is that the evaluation of PROs was not mandated by the ATLAS trial protocol and only a subset of patients was included in this analysis. Therefore, according to the Setting International Standards in Analyzing Patient-Reported Outcomes and Quality of Life Endpoints Data Consortium Guidelines,21 this analysis is essentially descriptive. Secondly, the questionnaires were completed only by patients actively treated in the study, thereby excluding results of patients with disease progression or those who withdrew from treatment for any other reasons; HRQoL scores may have been lower in this group of patients. Thirdly, the FACT / GOG-NTX questionnaires used in this study were not specifically designed for patients with multiple myeloma and may not capture some aspects of HRQoL in this particular population, such as dyspnea resulting from carfilzomib treatment.22 Additionally, these HRQoL questionnaires do not assess time toxicity, which refers to the duration of myeloma-related health care interactions, which is another crucial consideration for myeloma patients receiving extended maintenance requiring repeated intravenous infusions.

Conclusions

This analysis indicates that there is no difference in the tolerability of carfilzomib, lenalidomide, and dexamethasone maintenance, as compared with lenalidomide alone, among actively treated multiple myeloma patients. It also emphasizes the need for paying careful attention to patient emotional and social wellness during a period that is often considered uneventful for patients. Future clinical trials, especially in the maintenance setting, should strive to prospectively evaluate PROs to ensure the provision of the most accurate data. In this rapidly evolving field, considering patient perspectives alongside clinical efficacy is essential for informing and guiding optimal care.