Click here to read the peer-reviewed journal article on the Journal of Crohn’s and Colitis website.
Understanding inflammation in ulcerative colitis (UC) has long relied on complementary assessments: endoscopy for macroscopic evaluation and histology for microscopic assessment. While endoscopy remains the gold standard for assessing mucosal healing, microscopic inflammation can persist even in areas that appear normal, creating challenges in accurately characterizing disease activity—particularly for patients with mild to moderate UC.
In a recent publication in the Journal of Crohn’s and Colitis, Dr. Walter Reinisch, Dr. David T. Rubin, and colleagues leverage artificial intelligence (AI)-enabled tools to better quantify and link macroscopic and microscopic inflammation. Using retrospective data from the Phase 3 LUCENT-1 trial of mirikizumab (NCT03518086), the study demonstrates how AI can reveal subtle patterns of inflammation that traditional scoring methods may miss.
The Problem: Subtle Inflammation in Mild-to-Moderate UC
Conventional metrics, such as the Mayo Endoscopy Score (MES) and Geboes Score (GS), are highly correlated for patients with inactive or severe disease. However, in mild-to-moderate UC, correlation is low, leaving gaps in understanding disease activity and treatment response. Standard human assessment may underrepresent residual microscopic inflammation, which can influence long-term outcomes and clinical decision-making.
The AI-Driven Solution
By applying AI to both endoscopy and histology, the study uncovered granular insights across mild-to-moderate disease:
- AI Endoscopic Severity and Extent (AI-ESe) Score: Measures the proportion of colon surface affected by moderate-to-severe inflammation. For MES 1, median involvement was 11% vs 0% in patients with histologic improvement; for MES 2, 38% vs 20%.
- IBDExplore v2 Digital Pathology: Quantifies histologic features such as neutrophils and granulation tissue across tissue regions, aligning closely with endoscopic findings.
- Revealing Discordance: AI assessment highlighted where microscopic and macroscopic inflammation diverge, particularly in patients with MES 1–2, providing a more nuanced understanding of disease activity.
These findings suggest that AI can dissect the drivers of inflammation and mucosal barrier repair, offering the potential to better predict response to treatment and guide therapeutic strategies.
Why This Approach Matters
Embedding AI into clinical research allows for:
- Granular, objective measurement: AI captures subtle disease patterns beyond human visual scoring.
- Improved trial insights: Identifying microscopic inflammation can enhance evaluation of treatment efficacy in clinical trials.
- Patient-centered care: More precise disease characterization supports individualized treatment planning.
At Iterative Health, we are leveraging these AI-driven innovations to advance research in inflammatory bowel disease (IBD),
bridging clinical and microscopic data to accelerate insights and improve patient outcomes.
A Step Toward Smarter UC Research
This study underscores a broader trend: combining AI-enabled endoscopy and digital pathology offers a scalable, reproducible method to enhance UC trial outcomes. By providing a clearer picture of disease activity, these tools help investigators, sponsors, and patients navigate the high-variance space of mild-to-moderate UC with greater confidence.
Click here to read the peer-reviewed journal article on the Journal of Crohn’s and Colitis website.
About Iterative Health
Iterative Health is a healthcare technology and services company transforming clinical research and patient outcomes. By combining deep expertise in clinical trials with cutting-edge AI, we empower research teams to expand access to novel therapies and accelerate patient insights. Headquarters are in Cambridge, Massachusetts, and New York, New York. For more information, visit www.iterative.health and follow us on LinkedIn for updates.
To explore research partnerships, visit our Contact page.