Personalised Prediction of Disease Course in Ulcerative Colitis Using Multimodal Machine Learning - Part of the Presager Project

Copenhagen University Hospital, Hvidovre400 enrolled

Overview

In patients achieving clinical remission following a flare, artificial intelligence can reliably predict a new flare within the next 12 months utilizing clinical and objective information at day 0 and week 8. Secondary endpoints: * An artificial intelligence model's precision in predicting a new flare within 2 and 3 years * An artificial intelligence model's precision to rule out patients who will not experience a new flare within 1, 2 and 3 year

Study Type

OBSERVATIONAL

Enrollment

400

Conditions

Ulcerative Colitis

Interventions

Deep learning modelOTHER

Use of deep learning model to predict individual patients disease course

Eligibility

Sex: ALLMin age: 18 Years

Medical Language ↔ Plain English

Inclusion criteria: * Signed informed consent * Diagnosis of UC for at least 1 year * Relapse due to UC. * First endoscopic and histological evaluation of the flare Exclusion criteria * Well-founded doubt that the flare is due to other than the patients UC

Outcomes

Primary Outcomes

Flare I

Evaluate the accuracy in predicting a flare within 1 year after the patient's initial flare using machine learning methods.

Time frame: within 1 year

Secondary Outcomes

Flare II

Evaluate the accuracy in predicting a flare within 2 years, after the patient's initial flare, using machine learning methods.

Time frame: within 2 years

Flare III

Evaluate the accuracy in predicting a flare within 3 years, after the patient's initial flare, using machine learning methods.

Time frame: within 3 years