Development of artificial intelligence based model to predict mortality in patients with acute-on-chronic liver failure

Recruiting

Conditions: Fibrosis and cirrhosis of liver,

Registration Number: CTRI/2021/11/038131

Lead Sponsor: IHUB Anubhuti IITD Foundation TIH

Brief Summary: Liver cirrhosis is the most common cause ofdeath among gastrointestinal diseases, with a global burden of 1.5 billionpersons being affected annually. It is the 4th and 10thmost common cause of disease in males and females, respectively.Acute-on-chronic liver failure (ACLF) development is the most common cause ofdeath in these patients. The mortality of this dynamic syndrome ranges from15-100% within 30-days and 90-100% within one year of presentation. Multiplelife-threatening events like sepsis, organ failures, and gastrointestinalbleeding may occur during its course. This proposal is aimed to develop,validate, and deploy a novel model to predict mortality among patients withacute-on-chronic liver failure (ACLF). The model will be derived from clinical,biochemical, radiological and plasma proteomics data obtained from patientswith ACLF. Principles of AI, machine learning, and bioinformatics will beutilized to build such a model. The model aims to incorporate the systemsapproach to medicine with clinical sciences and information technology sector.The proposal includes multiple innovations. It will develop a multimediaprocessing engine (converting the existing data in a ready format for ML and AImodeling), development, and deployment of AI-derived model for real-timepredictive analytics. The models’ deployment will contribute to the developmentof user-friendly data analytic apps. The models may be used to assess thefutility of care and allocate significant resources like ICU and mechanicalventilators to deserving patients. Importantly, patients’ lives can be savedwith a timely decision for definitive treatment (liver transplant) using thesemodels.

Thisproposal aligns with the mandate of TiH in terms of knowledge generation(development of clinical and proteomic database), technology productdevelopment (portable AI model), skill development (integration of skills suchas clinical, biochemical, bioinformatics and information technology),collaboration (between clinicians, basic scientist, computational biologist,and information technologist).

Detailed Description: Not available

Recruitment & Eligibility

Status: Open to Recruitment

Sex: All

Target Recruitment: 200

Inclusion Criteria

Patients with a confirmed diagnosis of ACLF either by EASL criteria or by APASL criteria.

Exclusion Criteria

Patients with HIV infection, pregnant or lactating women, patients having any active malignancy or previous organ-transplantation and those refusing to give a consent will be excluded.

Study & Design

Study Type: Observational

Study Design: Not specified

Primary Outcome Measures

Name	Time	Method
Derivation of AI/ML model for prediction of mortality in ACLF patients	After establishment of clinical and proteomic database, model derivation phase will began in first one and a half year

Secondary Outcome Measures

Name	Time	Method
Development of web application for AI model deployment	After internal validation of model, a validation cohort will be recruited and for model validation and refinement in next one and a half year

Trial Locations

Locations (1): Post Graduate Institute of Medical Education and Research, Chandigarh
🇮🇳
Chandigarh, CHANDIGARH, India
Post Graduate Institute of Medical Education and Research, Chandigarh
🇮🇳Chandigarh, CHANDIGARH, India
Dr Nipun Verma
Principal investigator
9914208562
nipun29j@gmail.com