Commentary
Artificial intelligence uses computational methods that enable machines to learn from patterns and relationships in data in order to make predictions. Natural language processing can be used to extract knowledge from text, data can be modelled for the prediction of outcomes, and machine vision can analyse and recognise images.
Various demographic, laboratory, and histopathologic factors can be combined to determine which patients have the greatest risk of NASH.
Although non-invasive tests have shown promise and potential to replace liver biopsy, liver histology likely will remain the primary end point for phase 2b/3 trials and a key regulatory approval criterion for NASH drugs in the next few years. Unfortunately, histological evaluation of liver tissue has well-known limitations — sample variability, lack of patient acceptance, low reproducibility among readers, and variable placebo effects — that have been troublesome in NASH trials. Artificial intelligence can be used in order to reduce the shortcomings of this imperfect gold standard.
