Date: Monday, November 9, 2020
Session Type: Abstract Session
Session Time: 4:00PM-4:50PM
Background/Purpose: Methotrexate (MTX) is the most common anchor drug for rheumatoid arthritis (RA), but the risk of missing the opportunity for early effective treatment with alternative medications is substantial given the delayed onset of MTX action and 30-40% inadequate response rate. There is a compelling need to accurately predict MTX response before treatment initiation, in order to effectively identify patients at RA onset who are likely to respond to MTX. We aimed to test the ability of machine learning (ML) approaches with clinical and genomic biomarkers to predict MTX response in patients with RA.
Methods: Age, sex, clinical, serological and genome wide association study (GWAS) data on patients of European ancestry with early RA available through the PhArmacogenetics of Methotrexate in RA (PAMERA) consortium were used. A total of 647 patients were included: 336 recruited in the United Kingdom [UK]; 307 recruited across Europe (70% female; 72% rheumatoid factor [RF] positive; mean age 54 years; mean baseline Disease Activity Score with 28-joint count [DAS28] 5.65). The genomic data comprised 160 genome-wide significant single nucleotide polymorphisms (SNPs) with p< 1x10-5 that were associated with risk of RA and MTX metabolism. DAS28 scores were available at baseline and 3-month follow-up. Response to MTX monotherapy (>/=15 mg/week) was defined as good or moderate by the EULAR response criteria at 3-month follow-up. Supervised ML methods were trained with 5 repeats and 10-fold Cross validation using data from the UK patients. Class imbalance in training was accounted for by using simulated minority oversampling technique. Prediction performance was validated in the European patients (not used in training).
Results: Age, sex, RF positivity and baseline DASA28 data predicted response to MTX with area under the receiver operating curve (AUC) 0.54 in the UK subjects and 55% accuracy in European patients (p=0.98). However, supervised ML methods that combined demographics, RF status, baseline DAS28 and genomic SNPs predicted EULAR response at 3 months with AUC 0.84 (p=0.05) in UK patients, and achieved prediction accuracies of 76.2% (p=0.05) in the European patient’s (sensitivity 72%, specificity 77%). The addition of genomic data improved the predictive accuracies of MTX response by 19% and achieved cross-site replication. Baseline DAS28 and SNPs in or near the CASC15 (rs12446816), B3GNT2 (rs13385025), PARK2 (rs113798271), and ATIC (rs2372536) genes were among the top predictors of MTX response.
Conclusion: Pharmacogenomic biomarkers including gene variants for cancer susceptibility genes (CASC15) and important MTX pathway enzymes (ATIC) combined with baseline DAS28 score predicted MTX response in patients with early RA more reliably than demographics and baseline DAS28 alone, with replication in an independent cohort. Using pharmacogenomic biomarkers for the identification of MTX responders in early RA may help to guide effective RA treatment choices, including timely escalation of RA therapies. Further studies of personalized prediction of response to MTX and other antirheumatic treatments are needed to optimize control of RA disease and improve outcomes in patients with RA.
To cite this abstract in AMA style:Myasoedova E, Athreya A, Crowson C, Weinshilboum R, Wang L, Matteson E. Individualized Prediction of Response to Methotrexate Treatment in Patients with Rheumatoid Arthritis: A Pharmacogenomics-driven Machine Learning Approach [abstract]. Arthritis Rheumatol. 2020; 72 (suppl 10). https://acrabstracts.org/abstract/individualized-prediction-of-response-to-methotrexate-treatment-in-patients-with-rheumatoid-arthritis-a-pharmacogenomics-driven-machine-learning-approach/. Accessed November 25, 2020.
« Back to ACR Convergence 2020
ACR Meeting Abstracts - https://acrabstracts.org/abstract/individualized-prediction-of-response-to-methotrexate-treatment-in-patients-with-rheumatoid-arthritis-a-pharmacogenomics-driven-machine-learning-approach/