Exploring clusters based on ultrasound-detected inflammation in patients with psoriatic arthritis: a post-hoc analysis from the ULTIMATE trial

TL;DR

This study used ultrasound to identify different groups of psoriatic arthritis patients based on inflammation levels and found that treatment responses varied across these groups.

Contribution

The study introduces a novel method of clustering PsA patients using ultrasound-detected synovitis to better understand treatment responses.

Findings

Three distinct patient clusters were identified based on ultrasound-detected synovitis severity.

Cluster 2 showed higher power Doppler signal scores compared to clusters 1 and 3.

Secukinumab showed a trend of better response compared to placebo across all clusters.

Abstract

Psoriatic arthritis (PsA) is characterized by heterogeneous musculoskeletal manifestations. The aim of this post-hoc cluster analysis from the ULTIMATE study was to explore whether objective assessment of inflammation, using power-Doppler ultrasound (PDUS), identified homogeneous groups of patients based on the level of severity of ultrasound-detected synovitis and then to determine the longitudinal trajectory of response with secukinumab vs. placebo for each cluster up to week 12.

The ULTIMATE study enrolled patients with PsA and active clinical and PDUS synovitis. Patients were randomized to either secukinumab 150 or 300 mg, according to psoriasis severity, or placebo for the first 12 weeks. Multiple factor analysis was performed on baseline B-mode and power Doppler (PD) signal core components of the composite Global EULAR-OMERACT synovitis score (GLOESS) at joint level.

Three clusters were identified from all patients enrolled in the ULTIMATE study (N = 166), with 49 (30%), 54 (33%), and 63 (38%) patients assigned to clusters 1, 2, and 3, respectively. Higher GLOESS PDUS scores (± standard deviation) were observed in clusters 2 (35 ± 18) and 3 (30 ± 11) than cluster 1 (10 ± 6). While the B-mode core components of GLOESS were higher in clusters 2 (34 ± 18) and 3 (29 ± 11) than cluster 1 (10 ± 6), the PD signal core component was higher in cluster 2 (13 ± 10) than clusters 1 (5 ± 3) and 3 (5 ± 4). Overall, the pattern of distribution of PDUS-detected synovitis showed heterogeneity, with a predominance on small joints in cluster 2, more diffuse across small and large joints in cluster 3 and low frequency in small joints and knees in cluster 1. Change in GLOESS from baseline to week 12 showed a numerical trend toward a treatment difference favoring secukinumab vs. placebo in all three clusters.

Three clusters of patients were identified by ultrasound highlighting the heterogeneity of a typical PsA trial population. The addition of ultrasound to complement clinical assessment of swollen joints may help in better selecting active patients for inclusion in PsA clinical trials, thereby limiting the variability of the therapeutic response usually observed.

ClinicalTrials.gov (identifier: NCT02662985; date: 26/01/2016).

The online version contains supplementary material available at 10.1186/s12891-025-09434-w.

Figures