Abstract Background/Aims Guselkumab, a human monoclonal antibody targeting the interleukin-23 p19 subunit, demonstrated joint and skin efficacy in patients with PsA in the Phase III DISCOVER-1/-2 trials. MDA, a multi-domain composite outcome, is a clinically relevant measure of therapeutic response in PsA. However, response dynamics and the effect of individual domains on achieving MDA are not well understood. This analysis aimed to characterise MDA response clusters according to domain changes over time and identify potential baseline predictors in patients with PsA receiving guselkumab. Methods Data from bio-naïve patients with active PsA receiving guselkumab 100 mg every 4 or 8 weeks were pooled across DISCOVER-1/-2. Unsupervised machine learning using the time-series K-means clustering algorithm was performed to identify clusters according to MDA domain responses over 52 weeks. Missing data were not imputed. Results This analysis included 571 of 669 patients receiving guselkumab and distinguished four response clusters (C1-4; Table 1). Mean age and body mass index were similar across clusters; C3 had a lower proportion of female patients vs. other clusters. Relative to C3, a high burden of baseline disease was observed in C4 across clinical measures and PROs, and across PROs only in C2. Through Week 52, mean values for all MDA domains showed continuous improvement across clusters (Table 1). MDA response rates were highest in C3, in which all individual domain thresholds were rapidly approached, and lowest in C4. In C1 and C2, improvement in clinical measures paralleled that of C3; both met the PASI threshold and showed substantial reduction in SJC. PROs appeared to take longer to improve but did so earlier in C1 than C2. Improvements were slowest in C4, though still substantial given the high baseline disease burden. Conclusion Machine learning differentiated four clusters of patients with PsA receiving guselkumab based on response patterns in individual MDA domains. Response may differ due to baseline disease burden, especially in patients with high pain, PtGA and functional disability scores. These results offer an innovative, complementary approach to identifying treatment response patterns across diverse clusters of bio-naïve patients with PsA, which may facilitate clinical decision-making. Disclosure A. Zabotti: Member of speakers’ bureau; AbbVie, Amgen, Janssen, Lilly, Novartis and UCB. Grants/research support; Novartis. S. Ohrndorf: Other; Speaker fees or travel expense reimbursements from AbbVie, BMS, Janssen, Novartis and Pfizer. W. Tillett: Other; Research funding, consulting, speaker fees and/or honoraria from AbbVie, Amgen, Celgene, GlaxoSmithKline, Janssen, Lilly, MSD, Novartis, Pfizer and UCB. M. Neuhold: Other; Former employee of Janssen and is now an employee of Takeda Pharmaceuticals International AG. M. van Speybroeck: Shareholder/stock ownership; Employee of Janssen and owns stocks in Johnson & Johnson. E. Theander: Other; Former employee of Janssen. C. Contré: Other; Former employee of Janssen and is now an employee of Chiesi. M. Sharaf: Shareholder/stock ownership; Employee of Janssen and owns stocks in Johnson & Johnson. M. Shawi: Shareholder/stock ownership; Employee of Janssen and owns stocks in Johnson & Johnson. M. Perate: Shareholder/stock ownership; Employee of Janssen and owns stocks in Johnson & Johnson. A. Kollmeier: Shareholder/stock ownership; Employee of Janssen and owns stocks in Johnson & Johnson. P. Richette: Other; Received fees from AbbVie, Amgen, Celgene, Janssen, Lilly, MSD, Novartis, Pfizer and UCB.
Read full abstract