ACR Meeting Abstracts

ACR Meeting Abstracts

  • Home
  • Meetings Archive
    • 2023 ACR/ARP PRSYM
    • ACR Convergence 2022
    • ACR Convergence 2021
    • ACR Convergence 2020
    • 2020 ACR/ARP PRSYM
    • 2019 ACR/ARP Annual Meeting
    • 2018 ACR/ARHP Annual Meeting
    • 2017-2009 Meetings
    • Download Abstracts
  • Keyword Index
  • Advanced Search
  • Your Favorites
    • Favorites
    • Login
    • View and print all favorites
    • Clear all your favorites
  • Meeting Resource Center

Abstract Number: 1363

Identification of PsA Phenotypes with Machine Learning Analytics Using Data from a Phase 3 Clinical Trial Program of Guselkumab in a Bio-naïve Population of Patients with PsA

Pascal Richette1, Marijn Vis2, Sarah Ohrndorf3, William Tillett4, Marlies Neuhold5, Michel van Speybroeck6, Elke Theander7, Wim Noel8, May Shawi9, Alexa Kollmeier10 and Alen Zabotti11, 1Lariboisiere Hospital, Paris, France, 2Erasmus MC, Badhoevedorp, Netherlands, 3Charité University Medicine Berlin, Berlin, Germany, 4Royal National Hospital for Rheumatic Diseases, Bath, United Kingdom, 5Janssen-Cilag, Zug, Switzerland, 6Janssen Pharmaceutica, Belgium, Belgium, 7Janssen Cilag, Lund, Sweden, 8Janssen Pharmaceutica, Vilvoorde, Belgium, 9Janssen Immunology Global Commercial Strategy Organization, Toronto, ON, Canada, 10Janssen Research & Development, LLC, La Jolla, CA, 11University of Udine, Udine, Italy

Meeting: ACR Convergence 2021

Keywords: Biologicals, classification criteria, clinical trial, prognostic factors, Psoriatic arthritis

  • Tweet
  • Email
  • Print
Session Information

Date: Monday, November 8, 2021

Session Title: Spondyloarthritis Including PsA – Treatment Poster II: Psoriatic Arthritis I (1329–1363)

Session Type: Poster Session C

Session Time: 8:30AM-10:30AM

Background/Purpose: Psoriatic arthritis (PsA) is mainly described based on the individual domains or clinical components of the disease.1,2 The aim of this post hoc analysis was to identify hypothesis-free clusters of phenotypes according to patients’ clinical features and characteristics at baseline (BL) using data from the guselkumab Phase 3 DISCOVER-1 and -2 clinical trials of patients with PsA.

Methods: Data from bio-naïve patients with active PsA treated with guselkumab 100 mg either every 4 or 8 weeks in DISCOVER-1 and -2 were retrospectively analyzed. Non-negative matrix factorization was applied as an unsupervised machine learning technique to identify clusters of PsA phenotypes, with BL characteristics and clinical observations as input features. Clusters are described here according to their characteristics and clinical features, including the relation to achievement of minimal disease activity (MDA) at Weeks 24 and 52 (W24; W52).

Results: Data from 661 bio-naïve patients were pooled. Eight distinct clusters of PsA phenotypes were identified. Cluster 1 was characterized by a high frequency of lower limb involvement, notably foot/ankle, and the lowest rates of severe skin involvement (Figure 1). Cluster 2 was characterized by high skin involvement, the lowest proportion of females, and the highest proportion of overweight patients (70% body mass index [BMI] 25–< 30). Cluster 3 was characterized by a high burden of disease in the hands/wrists. Cluster 4, the smallest cluster, was distinguished by the highest degree of dactylitis involvement; second highest proportion of male patients; and the highest proportion of patients with ≥3% body surface area (BSA) psoriasis involvement. Cluster 5 was differentiated based on the highest proportion of patients with enthesitis at BL. Large joint involvement was also common in this cluster. Cluster 6 was characterized by a high level of involvement of small joints in the hands and feet, but a low mean dactylitis score. Presence of nail involvement (78%) and enthesitis at BL were also common in this cluster. Cluster 7 was characterized by axial involvement at BL (100%); nearly one-half of patients with dactylitis, two thirds with enthesitis and the majority with BSA ≥3% at BL (Figure 1). Cluster 8 was characterized by low rates of joint involvement, high rates of more extensive skin involvement, and the highest proportion of obese patients (67% with BMI >30; Table 1). MDA response rates at W24 and W52 were highest in Cluster 2 and lowest in Cluster 5. Clusters 3 and 4 were characterized by low MDA response rates at W24 that increased at W52 (Figure 2).

Conclusion: Unsupervised machine learning was able to identify 8 clusters of PsA phenotypes with significant differences in demographic and clinical features, including patterns of involvement of joint, skin/nail, dactylitis, and enthesitis manifestations as well as MDA responses. These clusters seem to differ in their initial vs. later responses to guselkumab. Further analysis will clarify which MDA domains play a role in this response kinetics.

1. Coates C et al. Arthritis Rheumatol 2016; 68: 1060–1071.

2. Gossec L et al. Ann Rheum Dis 2012; 71: 4–12.


Disclosures: P. Richette, AbbVie, 1, 6, Amgen, 1, 6, Celgene, 1, 6, Janssen, 1, 6, Eli Lilly, 1, 6, MSD, 1, 6, Novartis, 1, 6, Pfizer, 1, 6, UCB, 1, 6; M. Vis, AbbVie, 2, 5, 6, Amgen, 2, 5, 6, Celgene, 2, 5, 6, Janssen, 2, 5, 6, Eli Lilly, 2, 5, 6, Novartis, 2, 5, 6, Pfizer, 2, 5, 6, UCB, 2, 5, 6; S. Ohrndorf, AbbVie, 2, 6, Bristol Myers Squibb, 2, 6, Janssen, 2, 6, Novartis, 2, 6, Pfizer, 2, 6; W. Tillett, AbbVie, 1, 2, 6, Amgen, 1, 2, 6, Celgene, 1, 2, 6, Eli Lilly, 1, 2, 6, Janssen, 1, 2, 6, Novartis, 1, 2, 6, MSD, 1, 2, 6, Pfizer, 1, 2, 6, UCB, 1, 2, 6, Merck Sharp & Dohme, 2; M. Neuhold, Janssen Scientific Affairs, LLC (a subsidiary of Johnson & Johnson), 3, 11; M. van Speybroeck, Janssen Scientific Affairs, LLC (a subsidiary of Johnson & Johnson), 3, 11; E. Theander, Janssen Scientific Affairs, LLC (a subsidiary of Johnson & Johnson), 3, 11; W. Noel, Janssen Global Services, LLC, 3, 12, Owns stock in Johnson & Johnson; M. Shawi, Janssen Global Services, LLC (a subsidiary of Johnson & Johnson), 3, 11; A. Kollmeier, Janssen Research & Development, LLC (a subsidiary of Johnson & Johnson), 3, 11; A. Zabotti, Novartis, 5, AbbVie, 6, Amgen, 6, Eli Lilly, 6, UCB, 6, Novartis, 6.

To cite this abstract in AMA style:

Richette P, Vis M, Ohrndorf S, Tillett W, Neuhold M, van Speybroeck M, Theander E, Noel W, Shawi M, Kollmeier A, Zabotti A. Identification of PsA Phenotypes with Machine Learning Analytics Using Data from a Phase 3 Clinical Trial Program of Guselkumab in a Bio-naïve Population of Patients with PsA [abstract]. Arthritis Rheumatol. 2021; 73 (suppl 9). https://acrabstracts.org/abstract/identification-of-psa-phenotypes-with-machine-learning-analytics-using-data-from-a-phase-3-clinical-trial-program-of-guselkumab-in-a-bio-naive-population-of-patients-with-psa/. Accessed June 2, 2023.
  • Tweet
  • Email
  • Print

« Back to ACR Convergence 2021

ACR Meeting Abstracts - https://acrabstracts.org/abstract/identification-of-psa-phenotypes-with-machine-learning-analytics-using-data-from-a-phase-3-clinical-trial-program-of-guselkumab-in-a-bio-naive-population-of-patients-with-psa/

Advanced Search

Your Favorites

You can save and print a list of your favorite abstracts during your browser session by clicking the “Favorite” button at the bottom of any abstract. View your favorites »

© COPYRIGHT 2023 AMERICAN COLLEGE OF RHEUMATOLOGY

Wiley

  • Home
  • Meetings Archive
  • Advanced Search
  • Meeting Resource Center
  • Online Journal
  • Privacy Policy
  • Permissions Policies
  • Cookie Preferences