P0928 Patient endotypes in ulcerative colitis revealed by longitudinal symptom trajectories under therapy

Abstract

Background: Ulcerative colitis (UC) poses unique challenges in patient care. Although therapies have led to better outcomes, only 20-30% of patients enter remission, and even fewer sustain it over time. A critical need remains in addressing three key areas: i) predicting which patients will enter remission, ii) tailoring treatments to individual patient needs, and iii) identifying novel therapeutic targets that can sustain long-term remission. Symptom trajectories under therapy offer an accessible approach to addressing these challenges and this work investigates their alignment with molecular responses to assess their clinical value. Methods: Using a reverse translation approach, this study analyses longitudinal data from the etrolizumab phase 3 trials to gain insights into the UC patient population. Symptom response trajectories of 437 UC patients were tracked over 54 to 66 weeks using the partial Mayo Clinical Score (pMCS). Through unsupervised clustering of pMCS, we identified distinct subpopulations based on their patterns of disease activity. Clinical outcomes within these clusters were then examined, allowing us to uncover significant predictors of rapid and sustained remission. Molecular profiling using gene expression data was then performed, revealing the biological mechanisms that differentiate these patient groups and providing insights into their treatment responses. Results: Our analysis revealed five distinct patterns of disease response and progression, from patients achieving rapid and deep remission to those with persistently high disease activity. These five symptom-based subgroups match closely with traditional clinical endpoints. We found that predictors of sustained remission, such as elevated albumin levels, low endoscopic scores, and reduced markers of inflammation (faecal calprotectin and C-reactive protein) aligned with these subgroups. In addition, molecular profiling highlighted specific differences in gene expression and immune cell activity, notably showing that patients with deep remission had an enrichment of anti-inflammatory signalling pathways, suggesting a biological basis for their favourable outcomes. Conclusion: Longitudinal monitoring of symptom trajectories offers a non-invasive, easily measured and reliable proxy for early assessment of the likelihood of achieving deep, sustained remission in UC patients. These symptom patterns show strong concordance with endoscopic, laboratory, and molecular remission profiles. This approach aligns with the principles of personalised medicine by enabling i) dynamic symptom monitoring ii) application in Phase II trials to guide drug development by assessing if a patient is on the right path toward therapeutic success and iii) identification of novel therapeutic targets.