RWD-derived response in multiple myeloma
Tao Xu, James Roose, Mellissa Williamson , Ahmed Sawas,Wan-Jen Hong, Huan Jin, Kathleen Maignan, Alberto Rocci, Kasra Yousefi, Shaji Kumar , Stefka Tyanova
Abstract
Real-world data (RWD) are important for understanding the treatment course and response patterns of patients with multiple myeloma. This exploratory pilot study establishes a way to reliably assess response from incomplete laboratory measurements captured in RWD. A rule-based algorithm, adapted from International Myeloma Working Group response criteria, was used to derive response using RWD. This derived response (dR) algorithm was assessed using data from the phase III BELLINI trial, comparing the number of responders and non-responders assigned by independent review committee (IRC) versus the dR algorithm. To simulate a real-world scenario with missing data, a sensitivity analysis was conducted whereby available laboratory measurements in the dataset were artificially reduced.
Introduction
Multiple myeloma (MM) is a bone marrow malignancy accounting for almost 10% of all haematologic cancers [1]. Nearly all patients with MM experience relapse after initial or salvage therapy [2]. Despite numerous advances in treatment options for MM, including the use of second-generation proteasome inhibitors (PIs) and immunomodulatory drugs (IMiDs), as well as antibody therapies [3], there is an unmet need for improved treatment and management options for patients with relapsed/refractory MM (RRMM) [4]. Assessment of patient response to therapy is an important element when determining appropriate treatments for MM [5, 6]. In 2006, the International Myeloma Working Group (IMWG) developed a set of response criteria that are commonly used by physicians in the assessment of patients with MM [7], which were further updated in 2016 [8]. These response criteria are based on the comparison of serial MM-specific laboratory measures, including levels of monoclonal (M) protein in the serum and urine, and serum free light chains (FLCs), as well as radiologic images and bone marrow investigations when appropriate.
Material and methods
Description of the dR algorithm
Flexible IMWG criteria, with exclusion of bone marrow biopsy data and imaging results, and reduction in either serum or urine M protein levels (but not both), were used to define the following dR categories: partial response (PR), very good PR (VGPR), complete response (CR) and stringent CR (sCR) (Table 1). PR was assigned if any of the following criteria were met: (i) a reduction of >50% in at least two consecutive measurements of serum M protein, given that the requirement for measurable disease was met for serum M protein; (ii) a reduction of >90% in at least two consecutive urine M protein measurements, given that the requirement for measurable disease was met for urine M protein; or (iii) a reduction of >50% in FLC difference in two consecutive measurements, if M protein was unmeasurable (or unavailable) in serum or urine. For patients meeting the requirement for measurable disease for both serum and urine M protein, VGPR was assigned if (i) serum and urine M protein were detectable by immunofixation but not on electrophoresis; or (ii) there was a >90% reduction in serum M protein plus urine M protein level of <100 mg/24 hours. CR was assigned in case of negative immunofixation on serum and urine (with no requirement for bone marrow assessment). sCR was assigned when the flexible criterion for CR was met, in addition to a normal FLC ratio.
Results
In total there were 16 discrepant cases for which different responses were assigned by the IRC and the dR algorithm. In the only case assigned as a non-responder by the dR algorithm but as a PR by the IRC, there were not enough consecutive laboratory test results of the same type to confirm response, although other criteria were met. It is worth noting that the investigator in the BELLINI trial assessed this patient as having minimal response. Of the 15 cases in which the dR algorithm disagreed with the IRC assessment, eight were assigned as responders by investigator’s assessment, showing that there are cases in which the dR algorithm agrees more closely with the investigator’s assessment, and suggesting that these cases could be difficult to assess.
Discussion
Automated algorithms are increasingly used in clinical decision management due to their reliability and reproducibility, timely assessment of a large sample of patients, and adherence to recommended clinical practice [19]. Diverse applications of automated algorithms can be found for disease diagnosis [20], patient risk stratification [21], and prognostic scores [22, 23], and previous studies have demonstrated their potential for assessment of treatment response and disease progression [24, 25].
Acknowledgments
Third-party medical writing assistance, under the direction of all authors, was provided by Helen Cathro, PhD, of Ashfield MedComms, an Inizio company.
Citation: Xu T, Roose J, Williamson M, Sawas A, Hong W-J, Jin H, et al. (2023) RWD-derived response in multiple myeloma. PLoS ONE 18(5): e0285125.
https://doi.org/10.1371/journal.pone.0285125
Editor: Tao Huang, Chinese Academy of Sciences, CHINA
Received: October 21, 2022; Accepted: April 15, 2023; Published: May 11, 2023
Copyright: © 2023 Xu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: For eligible studies qualified researchers may request access to individual patient level clinical data through a data request platform. At the time of writing this request platform is Vivli (https://vivli.org/ourmember/roche/). For up-to-date details on Roche's Global Policy on the Sharing of Clinical Information and how to request access to related clinical study documents, see the website (https://go.roche.com/data_sharing). Anonymised records for individual patients across more than one data source external to Roche cannot, and should not, be linked due to a potential increase in risk of patient re-identification. The real-world data that support the findings of this study have been originated by Flatiron Health, Inc. These de-identified data may be made available upon request, and are subject to a license agreement with Flatiron Health; interested researchers should contact DataAccess@flatiron.com to determine licensing terms. The data are subject to a license agreement with Flatiron Health to protect patient privacy and ensure compliance with measures necessary to reduce the risk of re- identification. For example, the data necessary to replicate the study include numerous specific dates, including visit dates (i.e., laboratory or examination dates), treatment start and stop dates, and month of death, as well as laboratory test results. Other measures to maintain de-identification without contractual agreements in place are not feasible due to the study question, methods used, and data elements required.
Funding: The BELLINI trial is co-sponsored by AbbVie and Genentech, Inc. Third-party medical writing assistance, under the direction of all authors, was provided by Helen Cathro, PhD, of Ashfield MedComms, an Inizio company, and was funded by F. Hoffmann-La Roche Ltd. Genentech, Inc. and F. Hoffmann-La Roche Ltd. provided support in the form of salaries for authors [TX, MW, WJH, HJ, KM, AR, KY and ST], and these authors were involved in study design, data collection and analysis, decision to publish, and preparation of the manuscript.
Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: TX and ST report employment at Roche and Genentech. JR and AS report employment at Flatiron Health, Inc., which is an independent subsidiary of the Roche Group, and report stock ownership in Roche.KM, AR and KY report employment at Roche. MW, WJH and HJ report employment at Genentech. This does not alter our adherence to PLOS ONE policies on sharing data and materials.
Abbreviations: CI, confidence interval; CMH, Cochran-Mantel-Haenszel; CR, complete response; dR, derived response; ECOG, Eastern Cooperative Oncology Group; EHR, electronic health record; FDA, Food and Drug Administration; FLC, free light chain; h, hours; HR, hazard ratio; IFE, immunofixation; IMiD, immunomodulatory drugs; IMWG, International Myeloma Working Group; IRC, independent review committee; M, monoclonal; Max, maximum; Min, minimum; MM, multiple myeloma; MR, minimal response; OR, overall response; ORR, overall response rate; OS, overall survival; PI, proteasome inhibitors; PR, partial response; RRMM, relapsed refractory multiple myeloma; RWD, real-world data; sCR, stringent complete response; VGPR, very good partial response; Y, year
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0285125#ack
