In addition to efficacy and safety, novel regimens for the treatment of multiple myeloma (MM) should be assessed in terms of disease-related symptoms, the adverse events patient's experience, and their effect on health-related quality of life (HRQOL). Patient-reported outcome (PRO) measures capture these experiences from the patient's perspective. However, an interpretative challenge of PROs is understanding the magnitude of change that can be considered clinically meaningful.
The European Organisation for Research and Treatment of Cancer (EORTC) QLQ-C30, supplemented with the multiple myeloma-specific module (QLQ-MY20), are commonly used PRO measures in MM trials. However, the minimally important difference (MID) in QLQ-MY20 scores, between groups or for an individual who may be categorized as a responder, has not been established. Previously, estimates based on the statistical variability of scores, or a universal 10-point threshold, have been applied to approximate an MID or responder definition (RD).
This study aims to provide empirical estimates of the MID and RD for the QLQ-MY20 in an adult newly diagnosed (ND) and relapsed/refractory (RR) MM population using clinical trial data and qualitative patient interviews.
Data were pooled from three completed Phase III clinical trials assessing the efficacy and safety of carfilzomib-containing regimens in patients with RRMM (ASPIRE [NCT01080391], ENDEAVOR [NCT01568866]) or NDMM (CLARION [NCT01818752]). A recommended MID or RD should come from triangulating estimates from several different techniques. MIDs were estimated quantitatively by evaluating the mean PRO score changes in patients grouped as 'improved', 'stable' or 'deteriorated' according to external anchors. The anchors were selected based on variables common to the source datasets expected to reflect changes in PRO scores, and assessed for suitability through correlations. Those included were based on a PRO measuring chemotherapy-induced neuropathy (FACT/GOG-Ntx), ECOG performance status, and matched CTCAE toxicity events. RDs for deteriorations and improvements in health were estimated using the same anchor measures through receiver operating characteristic (ROC) curves, supported by 'distribution-based' estimates incorporating the statistical variability of scores. Qualitative telephone interviews were conducted to obtain patient insights from patients with MM regarding what constitutes meaningful change, to further support RD estimates and triangulate quantitative findings. This directly utilizes patient input to define an important treatment benefit and translate HRQOL change scores into tangible and clinically relevant impacts of the treatment on a patient. Interview participants were independent to the clinical trial samples and recruited via patient associations and clinical sites in the US and UK. Ethical approval was obtained for the study.
The pooled sample from the clinical trials comprised 2,147 patients for analysis; 14 additional patients (to date) were interviewed. MID and RD estimates varied as a result of the different methods, anchors and time points used to derive them. Based on the clinical trial sample, the range of MID and RD estimates for the QLQ-MY20 subscales were calculated and are presented in Table 1. The qualitative interview findings to date suggest that patients consistently understood the subscale concepts, relevance and response options for each domain of the QLQ-MY20. Patients discussed the concept of meaningful change at the domain level and provided reliable meaningful change estimates. The quantitative and qualitative findings will be triangulated to derive weighted estimates recommended for use in future studies. A single global estimate may not be appropriate for all subscales, in which case a range may be recommended.
Updated data based on full triangulation of the qualitative and quantitative evidence will be presented at the annual meeting. Findings from the current study will inform the development of new recommended interpretation guidelines for QLQ-MY20. These values will facilitate power analysis when designing studies and enable consistent interpretation of treatment effect on patients' HRQOL by regulators, payers and clinicians.
Yucel:Amgen: Employment, Equity Ownership. Trigg:Adelphi Values: Employment; Amgen: Consultancy. Sully:Adelphi Values Ltd: Employment; Amgen: Consultancy. Harper:Adelphi Values: Employment; Amgen: Consultancy. Bonner:Amgen: Consultancy; Adelphi Values: Employment. Shah:Nkarta: Consultancy; Amgen: Consultancy; Sutro Biopharma: Research Funding; University of California San Francisco: Employment; Indapta Therapeutics: Consultancy; Bristol Myers Squibb: Consultancy; Bluebird: Research Funding; Janssen: Research Funding; Kite: Consultancy; Nekktar: Consultancy; Teneobio: Consultancy; Celgene: Research Funding; Indapta Therapeutics: Equity Ownership; Sanofi: Consultancy; Takeda: Consultancy. Panjabi:Amgen: Employment, Equity Ownership. Cocks:Amgen: Consultancy; Bristol-Myers Squibb: Consultancy; Celgene: Consultancy; Endomag Ltd.: Consultancy; Adelphi Values: Employment.
Asterisk with author names denotes non-ASH members.