Zhang S, Chen Z, Wei J, et al. - In chronic hepatitis B (CHB) patients, the clinical decision-making could be done under the guidance of accurate liver necroinflammatory activity diagnosis. In this work, a non-invasive diagnostic model was developed for liver necroinflammatory activity by incorporating deep learning features and clinico-biochemical characteristics in CHB patients. Among a total of 239 CHB patients who underwent liver biopsy, the training cohort (n = 179) and an independent validation cohort (n = 60) were formed. Independent factors were immunoglobulin M, platelets, laminin, type IV collagen, gamma-glutamyl transferase, alanine aminotransferase, aspartate transaminase, alkaline phosphatase, direct bilirubin, and total bilirubin. Findings revealed better performance of the combined model vs models based on clinico-biochemical characteristics alone. Its clinical usefulness was confirmed in the decision curve. The combined model was identified as capable of providing an accurate non-invasive prediction of liver necroinflammatory activity, which might aid in clinical decision-making in CHB patients.