Towards stratified treatment of JIA: machine learning identifies subtypes in response to methotrexate from four UK cohorts
Stephanie J.W. Shoop-Worrall,Saskia Lawson-Tovey,Lucy R. Wedderburn,Kimme L. Hyrich,Nophar Geifman,Aline Kimonyo,Alyssia McNeece,Andrew Dick,Andrew Morris,Annie Yarwood,Athimalaipet Ramanan,Bethany R. Jebson,Chris Wallace,Daniela Dastros-Pitei,Damian Tarasek,Elizabeth Ralph,Emil Carlsson,Emily Robinson,Emma Sumner,Fatema Merali,Fatjon Dekaj,Helen Neale,Hussein Al-Mossawi,Jacqui Roberts,Jenna F. Gritzfeld,Joanna Fairlie,John Bowes,John Ioannou,Kimme L. Hyrich,Lucy R. Wedderburn,Melissa Kartawinata,Melissa Tordoff,Michael Barnes,Michael W. Beresford,Michael Stadler,Nophar Geifman,Paul Martin,Rami Kallala,Sandra Ng,Samantha Smith,Sarah Clarke,Saskia Lawson-Tovey,Soumya Raychaudhuri,Stephanie J.W. Shoop-Worrall,Stephen Eyre,Sumanta Mukherjee,Teresa Duerr,Thierry Sornasse,Vasiliki Alexiou,Victoria J. Burton,Wei-Yu Lin,Wendy Thomson,Zoe Wanstall
DOI: https://doi.org/10.1016/j.ebiom.2023.104946
IF: 11.205
2024-01-10
EBioMedicine
Abstract:Summary Background Methotrexate (MTX) is the gold-standard first-line disease-modifying anti-rheumatic drug for juvenile idiopathic arthritis (JIA), despite only being either effective or tolerated in half of children and young people (CYP). To facilitate stratified treatment of early JIA, novel methods in machine learning were used to i) identify clusters with distinct disease patterns following MTX initiation; ii) predict cluster membership; and iii) compare clusters to existing treatment response measures. Methods Discovery and verification cohorts included CYP who first initiated MTX before January 2018 in one of four UK multicentre prospective cohorts of JIA within the CLUSTER consortium. JADAS components (active joint count, physician (PGA) and parental (PGE) global assessments, ESR) were recorded at MTX start and over the following year. Clusters of MTX 'response' were uncovered using multivariate group-based trajectory modelling separately in discovery and verification cohorts. Clusters were compared descriptively to ACR Pedi 30/90 scores, and multivariate logistic regression models predicted cluster-group assignment. Findings The discovery cohorts included 657 CYP and verification cohorts 1241 CYP. Six clusters were identified: Fast improvers (11%), Slow Improvers (16%), Improve-Relapse (7%), Persistent Disease (44%), Persistent PGA (8%) and Persistent PGE (13%), the latter two characterised by improvement in all features except one. Factors associated with clusters included ethnicity, ILAR category, age, PGE, and ESR scores at MTX start, with predictive model area under the curve values of 0.65–0.71. Singular ACR Pedi 30/90 scores at 6 and 12 months could not capture speeds of improvement, relapsing courses or diverging disease patterns. Interpretation Six distinct patterns following initiation of MTX have been identified using methods in artificial intelligence. These clusters demonstrate the limitations in traditional yes/no treatment response assessment (e.g., ACRPedi30) and can form the basis of a stratified medicine programme in early JIA. Funding Medical Research Council , Versus Arthritis , Great Ormond Street Hospital Children's Charity , Olivia's Vision , and the National Institute for Health Research .
medicine, research & experimental