INTErPRET-NAFLD project: Research team has demonstrated that a Transformer Neural Network can be trained on 12 years of clinical history to achieve an accurate prediction of risk of all-cause mortality in patients with MASLD up to three years in advance..!
Can all-cause mortality in patients with Metabolic dysfunction-associated steatotic liver disease (MASLD) be predicted in advance using data from previous hospital stays? That was the question we set out to answer five years ago with Tim Kendall and Jonathan Fallowfield through the INTErPRET-NAFLD project. It turns out, yes it can. With our collaborators, the Bering team has demonstrated that a Transformer Neural Network can be trained on 12 years of clinical history to achieve an accurate prediction of risk of all-cause mortality in patients with MASLD up to three years in advance. We hope that extrapolation of this technique to population-level data can enable scalable and accurate risk stratification to identify people most likely to benefit from anticipatory health care and personalized interventions. You can read the full findings through the Annals of Hepatology: https://lnkd.in/edNj4wiF I'm incredibly greatful to the SteatoSITE, Precision Medicine Scotland Innovation Centre, and the West of Scotland Safe Haven teams for their support. And of course special thanks to Lynn McMahon, Marian McNeil, and Charlie Mayor for putting up with me over the last few years! #MASLD #MachineLearning #ClinicalArtificialIntelligence #BraveAI