AI-based Hepatic Steatosis Detection and Integrated Hepatic Assessment from Cardiac CT Attenuation Scans Enhances...

Area of Science:

• Cardiology
• Radiology
• Artificial Intelligence
• Hepatology

Background:

• Hepatic steatosis (HS) is a common cardiometabolic risk factor often under-diagnosed in patients with coronary artery disease (CAD).
• Accurate identification of HS is crucial for risk stratification in this patient population.

Purpose of the Study:

• To utilize artificial intelligence (AI) for automated quantification of hepatic tissue measures from CT attenuation correction (CTAC) scans during myocardial perfusion imaging (MPI).
• To evaluate the prognostic value of AI-derived HS detection and a novel AI liver risk index (LIRI) for predicting all-cause mortality in patients undergoing MPI.

Main Methods:

• An AI model was developed to segment liver and spleen on CTAC scans from 27,039 patients, quantifying liver and liver minus spleen (LmS) attenuation.
• HS was defined using thresholds of mean liver attenuation (<40 HU) or LmS attenuation (<-10 HU).
• A liver risk index (LIRI) was developed integrating hepatic measures and validated for prognostic value in predicting all-cause mortality.

Main Results:

• The AI algorithm identified HS in 24% of patients. During follow-up, HS was associated with increased mortality risk (adjusted HR: 1.14).
• The developed AI liver risk index (LIRI) demonstrated superior prognostic value compared to HS alone (adjusted HR: 1.5 vs 1.16).
• AI-based hepatic measures were automatically quantified without additional radiation or physician interaction.

Conclusions:

• AI-based hepatic measures can effectively identify HS from CTAC scans in patients undergoing MPI.
• An integrated AI liver risk assessment provides enhanced risk stratification for all-cause mortality, outperforming traditional HS detection.