At the 2024 AGA conference, researchers presented a breakthrough in ulcerative colitis (UC) diagnostics: a new AI-powered histology model that promises to simplify and standardize how remission is assessed in clinical trials and real-world care. This work represents a significant step forward for sponsors, CROs, and sites striving to reduce variability and accelerate drug development in inflammatory bowel disease (IBD).
Read the full abstract on GastroJournal.org
Why Histological Remission Matters in UC Trials
Histological remission (HR) is increasingly recognized as a key treatment target in UC, offering a more sensitive measure of mucosal healing than endoscopy alone. However, traditional histology scoring systems are complex, time-consuming, and subject to high interobserver variability, limiting their use outside of select research settings.
This variability poses real challenges for clinical trial sponsors and investigators who need consistent, reproducible assessments of disease activity to evaluate therapeutic response, inform endpoints, and support regulatory submissions.
Introducing the PICASSO Histologic Remission Index (PHRI)
To address these barriers, the study team developed the PICASSO Histologic Remission Index (PHRI), a simplified scoring system that uses just one feature: the presence or absence of neutrophils. This streamlines assessment while aligning closely with endoscopic findings such as Mayo and UCEIS scores.
The PHRI was validated on 614 digitized biopsy samples from over 300 UC patients across multiple centers. It demonstrated:
- Strong correlation with endoscopic scores (Spearman’s ρ = 0.55 to 0.78)
- Near-perfect agreement between pathologists (ICC = 0.84)
- Alignment with key clinical endpoints relevant to both trial design and patient management
Embedding PHRI into an AI System
Using a semi-supervised deep learning strategy, researchers trained a convolutional neural network (CNN) to detect neutrophils in biopsy patches and classify disease activity. The model achieved:
- Up to 94% specificity and 90% PPV for remission prediction
- 83% overall accuracy in test cohorts
- Reliable AI performance using only a binary presence/absence of neutrophils
This AI-powered approach offers a scalable, objective solution for histological grading in UC that can significantly reduce variability and time-to-insight.
Why This Matters for Sponsors and Sites
As AI becomes more integrated into clinical workflows, this model represents a compelling tool for:
- Standardizing endpoint assessments across trial sites
- Accelerating pathology review timelines
- Improving confidence in efficacy signals tied to histologic response
- Supporting regulatory-grade evidence generation
By embedding AI-driven histology into UC trial protocols, sponsors can more efficiently stratify patients, monitor disease activity, and validate endpoints with increased objectivity and reproducibility.
“This tool can effectively expedite, support, and standardize the histological assessment of UC in clinical practice.” — AGA 2024 Abstract Authors
Read the full abstract on GastroJournal.org.
This research is a milestone in bridging the gap between traditional pathology and modern machine learning. It adds to a growing body of evidence that AI can support more efficient, scalable, and patient-centric GI trials.
About Iterative Health
Iterative Health is a healthcare technology and services company powering the acceleration of clinical research to transform patient outcomes. By combining deep expertise in clinical trials with cutting-edge AI, we empower research teams and study sponsors to expand and expedite access to novel therapeutics for patients in need. Today, Iterative Health is based in Cambridge, Massachusetts, and New York City with 250+ employees world-wide.
To learn more or explore a partnership, please visit our Contact page.