AI-Based Stool Image Analysis for Colorectal Neoplasia Risk Assessment

Official title

FECAL-AI: Prospective Observational Validation of AI-Based Stool Image Analysis Against Quantitative Fecal Immunochemical Testing for Colorectal Neoplasia Risk Assessment

Overview

This prospective observational substudy evaluates the association between artificial intelligence-derived features from stool images analyzed using the FAEX Health digital platform and fecal immunochemical test results in adults undergoing colorectal cancer screening or diagnostic evaluation. Participants will capture stool images using a mobile application. The primary analysis will compare AI-derived image outputs with quantitative FIT values and FIT positivity. Secondary exploratory analyses will assess associations with colonoscopy and histopathological findings when these results are available. The platform will be used exclusively for research and will not provide diagnoses, replace clinical evaluation, or influence medical decisions.

Detailed description

This is a substudy of the project "Estrategia de prevención secundaria de cáncer colorrectal en personas mayores a 18 años." The substudy evaluates the FAEX Health digital platform, which applies artificial intelligence algorithms to stool images for non-diagnostic research and validation.

Participants will capture images of their stools using the FAEX Health mobile application. The images will be coded and analyzed using computational algorithms designed to identify visual characteristics such as color, consistency, and possible visible blood.

The primary objective is to evaluate the association and discriminatory performance of AI-derived stool image features for quantitative fecal immunochemical test results and FIT positivity. Secondary exploratory objectives are to assess associations between these image features and colonoscopic and histopathological findings among participants for whom these results are available.

The AI-derived results will not be returned to participants or treating clinicians and will not be used to determine whether colonoscopy or any other clinical procedure is performed. The findings may inform future studies evaluating stool image analysis as a potential triage strategy when FIT is unavailable or declined; however, the present study does not evaluate the platform as a replacement for FIT.

Personal identifiers will not be stored together with stool images. Access to coded study information will be restricted to authorized researchers, and study data will be managed according to applicable ethical, legal, and confidentiality requirements.

Interventions

• Diagnostic test AI-Based Stool Image Analysis
  Participants capture stool images using the FAEX Health mobile application. Coded images are analyzed using artificial intelligence algorithms to derive visual features and a prespecified patient-level output or score. The AI-derived output is used exclusively for research and is compared primarily with quantitative FIT results and FIT positivity, with secondary comparisons against colonoscopy and histopathology when available. The output is not used to provide a diagnosis or guide clinical mana

Primary outcome measures

• Correlation Between AI-Derived Stool Image Score and Quantitative FIT (faecal immunochemical test) Concentration [Time frame: Within 90 days of stool image submission]

Secondary outcome measures (3)

• Area Under the ROC Curve for FIT Positivity [Time frame: Within 90 days of stool image submission]
• Sensitivity and Specificity of the AI-Derived Stool Image Score for FIT Positivity/Negativity [Time frame: Within 90 days of stool image submission]
• Area Under the ROC Curve for Colonoscopy-Detected Colorectal Neoplasia [Time frame: Within 90 days of stool image submission]

Eligibility criteria

Inclusion criteria

• Age 18 years or older.
• Referred for screening or diagnostic colonoscopy at Hospital Dr. Sótero del Río.
• Quantitative fecal immunochemical testing planned or completed within 30 days before or after stool image submission.
• Able to submit at least one stool image using the FAEX Health mobile application, independently or with assistance.
• Able and willing to provide written informed consent.

Exclusion criteria

• Unable or unwilling to provide written informed consent.
• Previous enrollment in the study.
• Unable to complete stool-image capture, even with assistance.
• Study images and clinical data cannot be reliably linked using the assigned study code.

Criteria are shown verbatim from the registry (in English). Final eligibility is always assessed by the study center.

Study design

Observational model

Cohort

Study locations

Chile · 1 center

• Hospital Sotero del Rio — Santiago

Publications

• Lee JW, Woo D, Kim KO, Kim ES, Kim SK, Lee HS, Kang B, Lee YJ, Kim J, Jang BI, Kim EY, Jo HH, Chung YJ, Ryu H, Park SK, Park DI, Yu H, Jeong S; IBD Research Group of KASID and Crohn's and Colitis Association in Daegu-Gyeongbuk (CCAiD). Deep Learning Model Using Stool Pictures for Predicting Endoscopic Mucosal Inflammation in Patients With Ulcerative Colitis. Am J Gastroenterol. 2025 Jan 1;120(1):2 PMID 39051648
• Collins GS, Moons KGM, Dhiman P, Riley RD, Beam AL, Van Calster B, Ghassemi M, Liu X, Reitsma JB, van Smeden M, Boulesteix AL, Camaradou JC, Celi LA, Denaxas S, Denniston AK, Glocker B, Golub RM, Harvey H, Heinze G, Hoffman MM, Kengne AP, Lam E, Lee N, Loder EW, Maier-Hein L, Mateen BA, McCradden MD, Oakden-Rayner L, Ordish J, Parnell R, Rose S, Singh K, Wynants L, Logullo P. TRIPOD+AI statement: PMID 38626948
• Sounderajah V, Guni A, Liu X, Collins GS, Karthikesalingam A, Markar SR, Golub RM, Denniston AK, Shetty S, Moher D, Bossuyt PM, Darzi A, Ashrafian H; STARD-AI Steering Committee. The STARD-AI reporting guideline for diagnostic accuracy studies using artificial intelligence. Nat Med. 2025 Oct;31(10):3283-3289. doi: 10.1038/s41591-025-03953-8. Epub 2025 Sep 15. PMID 40954311
• Zhong H, Hou C, Huang Z, Chen X, Zou Y, Zhang H, Wang T, Wang L, Huang X, Xiang Y, Zhong M, Hu M, Xiong D, Wang L, Zhang Y, Luo Y, Guan Y, Xia M, Liu X, Yang J, Gan T, Wei W, Chen H, Gong H. A clinical pilot trial of an artificial intelligence-driven smart phone application of bowel preparation for colonoscopy: a randomized clinical trial. Scand J Gastroenterol. 2025 Jan;60(1):116-121. doi: 10.108 PMID 39709551
• Ramprasad C, Saini D, Del Carmen H, Krasnovsky L, Chandra R, Mcgregor R, Shinohara RT, Eaton E, Gummadi M, Mehta S, Lewis JD. Text Message System for the Prediction of Colonoscopy Bowel Preparation Adequacy Before Colonoscopy: An Artificial Intelligence Image Classification Algorithm Based on Images of Stool Output. Gastro Hep Adv. 2024 Sep 19;4(2):100556. doi: 10.1016/j.gastha.2024.09.011. eColle PMID 39866713
• Katsoula A, Paschos P, Haidich AB, Tsapas A, Giouleme O. Diagnostic Accuracy of Fecal Immunochemical Test in Patients at Increased Risk for Colorectal Cancer: A Meta-analysis. JAMA Intern Med. 2017 Aug 1;177(8):1110-1118. doi: 10.1001/jamainternmed.2017.2309. PMID 28628706
• Rahman F, Trivedy M, Rao C, Akinlade F, Mansuri A, Aggarwal A, Laskaratos FM, Rajendran N, Banerjee S. Faecal Immunochemical Testing to Detect Colorectal Cancer in Symptomatic Patients: A Diagnostic Accuracy Study. Diagnostics (Basel). 2023 Jul 10;13(14):2332. doi: 10.3390/diagnostics13142332. PMID 37510076
• Bailey JA, Weller J, Chapman CJ, Ford A, Hardy K, Oliver S, Morling JR, Simpson JA, Humes DJ, Banerjea A. Faecal immunochemical testing and blood tests for prioritization of urgent colorectal cancer referrals in symptomatic patients: a 2-year evaluation. BJS Open. 2021 Mar 5;5(2):zraa056. doi: 10.1093/bjsopen/zraa056. PMID 33693553

Identifiers

NCT: NCT07740122 · HSR-FAEX-CRC-FIT-2026

Primary sources (government registries)

• ClinicalTrials.gov (US NIH)

View this study on ClinicalTrials.gov ↗