Cure8 research brief
People with Crohn's disease or ulcerative colitis might see more AI-based prognostic tools in the literature, but this review shows most models are not yet proven ready for clinical use because they lack key validation and calibration steps.
Clinicians and researchers developing or evaluating prognostic models for IBD; patients and caregivers curious about emerging AI tools for predicting treatment response or disease course.
The review pooled and summarized published ML prognostic studies from 2012–2026, finding treatment-response prediction was a common target and that laboratory and EHR data were frequent model inputs.
Median reported discrimination across studies was high (median AUC 0.85), and externally validated models showed slightly higher AUCs than internally validated ones. However, external validation was present in under a third of studies, calibration was rarely reported, and analysis issues were the main source of bias identified by PROBAST + AI.
The practical takeaway is that while many ML models for IBD show promising discrimination on paper, the review judges most as not yet ready for clinical deployment without additional validation steps (external validation, calibration, utility analyses, and transparency of code/data).
The review assessed studies up to January 2026 using PROBAST + AI and reports on common methodological deficiencies (limited external validation, sparse calibration reporting, low code sharing). It summarizes study-level reporting rather than testing tools in clinical practice.
Review the original publication for the complete reporting, methods, and context.
This Cure8 brief is based on source text from the linked article. Cure8 is informational only and is not a substitute for professional medical advice, diagnosis, or treatment.
People with IBD face an elevated lifetime risk of colitis-associated cancer; improved detection and risk prediction could lead to earlier treatment and better outcomes. AI could make surveillance more accurate and...
Vitamin D supplementation is commonly considered in people with IBD because deficiency is frequent and low vitamin D may relate to inflammation.
Genetic ancestry can influence IBD risk and disease course, and most prior studies focused on European populations.
