Approximating International Myeloma Working Group Uniform Response Criteria to Derive Response for Multiple Myeloma (MM) Patients Using Data from Electronic Health Records (EHR)

Background

International Myeloma Working Group (IMWG) consensus criteria for response in MM require reductions in both serum protein electrophoresis (SPEP) and 24-hour urine protein electrophoresis (UPEP) for patients without oligosecretory disease. Requiring both SPEP and UPEP results to assess response significantly limits evaluability in the real-world, given that 24-hour UPEP tests are not used often in routine clinical practice (Foster RB, et al. Blood. 2018;132(Suppl. 1):3536). We sought to determine if response events and response rates could be derived using data readily available in the EHR for patients with MM who received lenalidomide, bortezomib, and dexamethasone (RVD) therapy.

Methods

This retrospective cohort study used de-identified information from the nationwide Flatiron Health EHR-derived database; Institutional Review Board approval with a waiver of informed consent was obtained. The cohort included patients from oncology clinics across the U.S. with a confirmed MM diagnosis via chart review from 01/01/2011 to 06/30/2019. The cohort was limited to patients who received at least one line of therapy. Patients whose MM treatment start (captured through chart review) was > 30 days before start of activity in the database were excluded, as this may indicate missing therapy data.

Patients were followed for response by either SPEP, UPEP, or serum free light chain (SFLC) lab tests. Patients were eligible if they had at least one measurable test (specimen type), consistent with IMWG criteria (SPEP ≥ 1 g/dL, 24-hour UPEP ≥ 200 mg, SFLC of any numeric value). Specimen type was determined using a hierarchy (SPEP > 24-hour UPEP > SFLC) when multiple test types were available. An initial baseline lab was assigned to each line of therapy: the highest lab value within 90 days of the start of the line of therapy, or the earliest lab within the line of therapy if there were only labs > 90 days after therapy start. Subsequent labs within the line were evaluated for a partial response (PR) or very good partial response or better (≥ VGPR) according to IMWG criteria based on percent change from baseline.

Derived response rate (dRR) was defined as the proportion of patients who had at least one PR or better assessment during RVD therapy, among patients with a baseline lab value and a subsequent lab value during RVD ("assessable for response").

Results

Out of 7,506 patients, 4,626 (61.6%) had at least one measurable SPEP test (n=4,257) or at least one measurable 24-hour UPEP test (n=788). Only 419 (9.1%) had at least one measurable test for both parameters.

Patients who received first-line (1L) RVD and were assessable for response by SPEP or UPEP (n=1,379) had a dRR of 0.89. Among 1,196 patients with a response event tracked by SPEP during 1L RVD, only 102 (8.5%) had at least one of these events with a measurable 24-hour UPEP test conducted within 14 days. Those who received RVD in second-line and were assessable by SPEP or UPEP (n=222) had a dRR of 0.75.

These response rates are generally consistent with results from the phase III SWOG S0777 trial investigating upfront RVD (overall response rate [ORR] 0.82) (Durie BGM, et al. Lancet. 2017;89(10068):519-527), and a phase II study of RVD in the relapsed/refractory setting (ORR 0.64) (Richardson PG, et al. Blood. 2014;123(10):1461-1469).

Conclusions

Strictly applying IMWG response criteria in routine practice settings may be challenging, as patients rarely have both SPEP and UPEP measured as part of routine clinical care. However, the results of this study suggest that response in real-world patients with MM can be effectively derived from the results of a single test type. Results from this single-specimen derivation are reasonably consistent with select published results across different lines of therapy. It should be noted that the real-world response rates reported are not confirmed by subsequent assessment and any formal comparison would have to account for differences in the patient characteristics between the real-world cohorts and those in clinical trials.

Further studies may be warranted to investigate alternatives for defining CR in the real-world. Adherence to IMWG criteria in the real-world may be further limited by the requirement of a bone marrow biopsy for confirmation of complete response (CR), which may be performed less frequently than SPEP or UPEP lab tests. As such, discerning a VGPR from a CR may not be feasible with this real-world response approach.