Diagnosis and risk stratification of liver disease using deep learning on routine clinical data

Background: Liver diseases are widespread and can have life-threatening consequences, requiring new approaches to diagnosis and treatment.

Aim:  Develop and validate computer-based methods for diagnosis and prediction of unfavourable courses in liver diseases.

Methods: Evaluation of histological images from liver biopsies and clinical data by means of deep learning with artificial neural networks.

Application: Develop and test a prototype of an online platform for decentralised provision of deep learning methods.

Longterm goal: Offer a tool that will improve individual risk prediction and accelerate the diagnosis of rare diseases.