Apollo’s AI-LF model offers precise clinical decision support for risk stratification in
low-cost settings, validated with an AUC of 0.83 against 98 liver biopsies. Its superiority
emphasizes potential for efficient, personalized healthcare.
Non-alcoholic fatty liver disease (NAFLD) poses a significant health concern, impacting 9% to
32% of India’s population. This health condition encompasses a spectrum from steatosis to
fibrosis. Recognizing the need for precise identification of advanced fibrosis, Apollo Hospitals
initiated a ground breaking endeavour to create a Machine Learning Model. With over 70 hospitals
catering to over 200 million patients across India, Apollo Hospitals is at the forefront of
healthcare. Presently, discussions are underway between Apollo Hospitals and a leading
Ultrasound Pioneer, exploring a potential collaboration to integrate advanced algorithms into
liver screenings, marking a crucial step towards enhancing AI-LF solutions.
Apollo Hospitals, pioneers a revolutionary approach to liver health with AI LF. Extracting
insights from 5150 NAFLD patients, data spanning 2011-2017 utilized machine learning through XGB
algorithm. The resulting AI-driven risk assessment tool, AI LF, excels in identifying advanced
fibrosis (F2-4) risk. Prospective validation from 2018-2020 solidifies its superior performance
over Fib4 Score, validated even by 98 liver biopsies. This initiative cements Apollo’s legacy as
a healthcare leader, pushing boundaries for precise and accessible liver health solutions.
Apollo’s AI-LF model promises a transformative impact, providing invaluable clinical decision
support for risk stratification in low-cost settings. Recognizing 11 significant variables,
including age and diabetes history, the model offers precise insights. Its validation against 98
liver biopsies yielded an AUC of 0.83, and the Net Reclassification Improvement (NRI) at 0.499
emphasizes its superiority over existing methodologies. This innovation holds immense potential
for personalized, efficient healthcare delivery. Apollo’s AI-driven diagnosis and risk
stratification tools, extensively reviewed and validated before use, are now being used in at
least eight countries and have been adapted for other non-communicable diseases, such as
diabetes, asthma and liver fibrosis.
