Optimizing noninvasive assessment Of Dysmetabolic compensated advanced Liver disease by integration of artificial intelligence model and omics data (MODELS)

Globally, nonalcoholic fatty liver disease (NAFLD) and its severe form, nonalcoholic steatohepatitis (NASH), have a high prevalence and are becoming leading causes of cirrhosis, hepatocellular carcinoma (HCC), and the need for liver transplantation. By 2030, NAFLD's overall prevalence is projected to reach 33.5 %. In 2015, 20% of NAFLD cases were classified as NASH, but this is expected to increase to 27% by 2030. The incidence of decompensated cirrhosis is estimated to rise by 168% to 105,430 cases by 2030, and HCC incidence is predicted to increase by 137%. Liver-related deaths are projected to increase by 178%.
To effectively prevent liver disease, early identification before cirrhosis is crucial. In 2015, the term "compensated advanced chronic liver disease" (CACLD) was introduced to encompass advanced fibrosis and cirrhosis in asymptomatic individuals. The diagnosis or exclusion of cACLD was proposed to rely on specific liver stiffness values measured non-invasively using transient elastography. However, in NAFLD, there is a risk of misclassification.
Our project aims to address the need for accurate classification of patients labeled as "unclassified" by the Baveno VII criteria for CACLD. Specifically, our study focuses on (1) evaluating the effectiveness of innovative non-invasive tests (NITS), new ultrasound techniques, and elastography of liver stiffness (LSM) and spleen stiffness (SSM) for CACLD, (2) assessing whether the combination of LSM, SSM, and NITS can facilitate non-invasive diagnosis of CACLD in this group, and (3) determining whether incorporating non-invasive tests and omic data improves the detection strategy for CACLD compared to the Baveno VII criteria, using an artificial intelligence approach.
To achieve these objectives, our research team consists of experts in clinical hepatology, NAFLD, and advanced liver disease, along with facilities for large-volume outpatient clinics, ultrasound sciences, laboratory sciences, and a secondary team specializing in statistical analysis, data management, and artificial intelligence (AI) models.
By utilizing cutting-edge methods, our project aims to enhance the diagnostic pathway for CACLD, addressing the urgent clinical need resulting from its high prevalence and the potential overuse of diagnostic resources. We anticipate generating valuable insights and evidence to improve clinical practice through optimized diagnostic pathways using artificial intelligence models. Additionally, our results can contribute to refining referral paths and assessing the effectiveness of innovative diagnostic approaches by sharing a comprehensive dataset with the scientific community.
Our approach incorporates health technology assessment (HTA) to distinguish between valuable and unwarranted applications. HTA is a comprehensive evaluation process that considers medical, economic, ethical, and social aspects of health technologies. Through HTA, we can systematically assess the benefits, risks, and cost-effectiveness of different applications, ensuring that only those with proven value and efficacy are recommended for implementation. This approach enables informed decision-making based on evidence, maximizing the benefits of healthcare technologies while minimizing unnecessary expenses and potential harm. By distinguishing valuable from unwarranted applications, we can optimize resource allocation, promote patient-centered care, and contribute to the overall improvement of the healthcare system.