Objective: Artificial intelligence (AI) has begun to play a pivotal role in hepatic imaging. This systematic scoping review summarizes the latest progress of AI in evaluating hepatic diseases based on computed tomography (CT) and magnetic resonance (MR) imaging. Methods: We searched PubMed and Web of Science for publications, using terms related to deep learning, radiomics, imaging methods (CT or MR), and the liver. Two reviewers independently selected articles and extracted data from each eligible article. The Quality Assessment of Diagnostic Accuracy Studies-AI (QUADAS-AI) tool was used to assess the risk of bias and concerns regarding applicability. Results: The screening identified 45 high-quality publications from 235 candidates, including 8 on diffuse liver diseases and 37 on focal liver lesions. Nine studies used deep learning and 36 studies used radiomics. All 45 studies were rated as low risk of bias in patient selection and workflow, but 36 (80%) were rated as high risk of bias in the index test because they lacked external validation. In terms of concerns regarding applicability, all 45 studies were rated as low concerns. These studies demonstrated that deep learning and radiomics can evaluate liver fibrosis, cirrhosis, portal hypertension, and a series of complications caused by cirrhosis, predict the prognosis of malignant hepatic tumors, and differentiate focal hepatic lesions. Conclusions: The latest studies have shown that deep learning and radiomics based on hepatic CT and MR imaging have potential application value in the diagnosis, treatment evaluation, and prognosis prediction of common liver diseases. The AI methods may become useful tools to support clinical decision-making in the future. Advances in knowledge: Deep learning and radiomics have shown their potential in the diagnosis, treatment evaluation, and prognosis prediction of a series of common diffuse liver diseases and focal liver lesions.

Objective: The purpose is to establish and validate a machine-learning-derived radiomics approach to determine the existence of atrial fibrillation (AF) by analyzing epicardial adipose tissue (EAT) in CT images. Methods: Patients with AF based on electrocardiographic tracing who underwent contrast-enhanced (n = 200) or non-enhanced (n = 300) chest CT scans were analyzed retrospectively. After EAT segmentation and radiomics feature extraction, the segmented EAT yielded 1691 radiomics features. The most contributive features to AF were selected by the Boruta algorithm and machine-learning-based random forest algorithm, and combined to construct a radiomics signature (EAT-score). Multivariate logistic regression was used to build clinical factor and nested models. Results: In the test cohort of contrast-enhanced scanning (n = 60/200), the AUC of EAT-score for identifying patients with AF was 0.92 (95%CI: 0.84–1.00), higher than 0.71 (0.58–0.85) of the clinical factor model (total cholesterol and body mass index) (DeLong’s p = 0.01), and higher than 0.73 (0.61–0.86) of the EAT volume model (p = 0.01). In the test cohort of non-enhanced scanning (n = 100/300), the AUC of EAT-score was 0.85 (0.77–0.92), higher than that of the CT attenuation model (p < 0.001). The two nested models (EAT-score+volume and EAT-score+volume+clinical factors) for contrast-enhanced scan and one (EAT-score+CT attenuation) for non-enhanced scan showed similar AUCs with that of EAT-score (all p > 0.05). Conclusion: EAT-score generated by machine-learning-based radiomics achieved high performance in identifying patients with AF. Advances in knowledge: A radiomics analysis based on machine learning allows for the identification of AF on the EAT in contrast-enhanced and non-enhanced chest CT.

With the continuous expansion of the disease scope of chest CT and cardiac CT, the number of these CT examinations has increased rapidly. In addition to their common indications, many incidental cardiac findings can be observed when carefully evaluating the coronary arteries, valves, pericardium, ventricles, and large vessels. These findings may have clinical significance or risk of complications, but they are sometimes overlooked or may not be described in the final reports. Although most of the incidental findings are benign, timely detection and treatment can improve the management of chronic diseases or reduce the possibility of severe complications. In this review, we summarized the imaging findings, incidence rate, and clinical relevance of some benign cardiac findings such as coronary artery calcification, aortic and mitral valve calcification, aortic calcification, cardiac thrombus, myocardial bridge, aortic dilation, cardiac myxoma, pericardial cyst, and coronary artery fistula. Reporting incidental cardiac findings will help reduce the risk of severe complications or disease deterioration and contribute to the recovery of patients.

Objective: The objective of this study was to construct a large language model (LLM) for the Response Evaluation Criteria in Solid Tumors (RECIST) assessment using exclusively longitudinal radiology report text. Methods: This study included 258 patients with solid tumors, encompassing 2065 longitudinal CT/MRI examination time points. We developed WiNGPT-32B, an open-source and locally deployable LLM, by infusing it with domain-specific medical knowledge and optimizing it via knowledge distillation, using GPT-4 as the teacher model. Central to its architecture is the Chained Task Execution (CTE) framework, which structures RECIST assessment into four modular components: lesion diameter extraction, sum of longest diameter computation, tumor response classification, and report generation. Model performance (accuracy, recall, precision, and F1 score) was benchmarked against GPT-4 and a single radiologist, utilizing the consensus of three independent radiologists as the reference standard. Results: The number of patients with imaging time points was 212 (82.2%) with 4–10, 36 (13.9%) with 11–20, and 10 (3.9%) with >20 time points. For target lesions, the successful extraction rate of WiNGPT-32B was 0.934 (95% CI: 0.922–0.944), which was slightly higher than that of GPT-4 0.920 (0.907–0.931; p = 0.083). In five-category RECIST classification (complete response, partial response, stable disease, progressive disease, and not evaluable), WiNGPT-32B achieved an overall accuracy of 0.805 (0.786–0.823), significantly higher than GPT-4 (0.699, 0.678–0.720; p < 0.001) but lower than the radiologist (0.915, 0.901–0.928; p < 0.001). For progressive disease, WiNGPT-32B had an F1 score of 0.841 (0.813–0.870), significantly outperforming GPT-4’s 0.755 (0.720–0.790), and approaching the radiologist’s 0.922 (0.902–0.942). Conclusions: WiNGPT-32B demonstrates the feasibility of a text-only, open-source LLM with the CTE framework for longitudinal RECIST assessment, with promising performance in detecting disease progression.