This study aimed to develop a deep learning-based model to simultaneously perform the objective response (OR) and tumor segmentation for hepatocellular carcinoma (HCC) patients who underwent transarterial chemoembolization (TACE) treatment. A total of 248 patients from two hospitals were retrospectively included and divided into the training, internal validation, and external testing cohort. A network consisting of an encoder pathway, a prediction pathway, and a segmentation pathway was developed, and named multi-DL (multi-task deep learning), using contrast-enhanced CT images as input. We compared multi-DL with other deep learning-based OR prediction and tumor segmentation methods to explore the incremental value of introducing the interconnected task into a unified network. Additionally, the clinical model was developed using multivariate logistic regression to predict OR. Results showed that multi-DL could achieve the highest AUC of 0.871 in OR prediction and the highest dice coefficient of 73.6% in tumor segmentation. Furthermore, multi-DL can successfully perform the risk stratification that the low-risk and high-risk patients showed a significant difference in survival (p = 0.006). In conclusion, the proposed method may provide a useful tool for therapeutic regime selection in clinical practice.

INTRODUCTION: Blood–brain barrier (BBB) breakdown is essential in Alzheimer's disease (AD) and dementia with Lewy bodies (DLB), whereas the variability in BBB permeability to water and contrast agent is less clear. METHODS: We investigated BBB permeability to water and contrast agent simultaneously using a novel tracer kinetic model for dynamic contrast‐enhanced magnetic resonance imaging (DCE‐MRI) in 42 AD patients, 22 DLB patients, and 22 healthy controls. All participants underwent clinical, cognitive, and MRI assessments. RESULTS: AD patients exhibited a significant reduction in the water exchange rates across the BBB, whereas DLB patients showed a significant increase in BBB permeability to contrast agent. Moreover, BBB permeability to both water and contrast agent in multiple brain regions demonstrated correlations with clinical severity. DISCUSSION: The simultaneous evaluation of BBB permeability to water and contrast agent based on the proposed model highlights the heterogeneous patterns of BBB breakdown in AD and DLB. Highlights: We measured blood–brain barrier (BBB) permeability to water and contrast agent based on dynamic contrast‐enhanced magnetic resonance imaging. Alzheimer's disease (AD) is characterized by lower water exchange rates across the BBB. Dementia with Lewy bodies exhibits higher BBB permeability to contrast agent. BBB permeability is related to cognitive impairment and disease burden. BBB permeability to water was negatively associated with that to contrast agent.

Background: Congenital biliary dilatation (CBD) necessitates the timely removal of dilated bile ducts. Accurate differentiation between CBD and secondary biliary dilatation (SBD) is crucial for treatment decisions, and identification of CBD with intrahepatic involvement is vital for surgical planning and supportive care. This study aimed to develop quantitative models based on bile duct morphology to distinguish CBD from SBD and further identify CBD with intrahepatic involvement. Materials and methods: The retrospective study included 131 CBD and 209 SBD patients between December 2014 and December 2021 for model development, internal validation, and testing. A separate cohort of 15 CBD and 34 SBD patients between January 2022 and December 2022 was recruited for temporally-independent validation. Quantitative shape-based (Shape) and diameter-based (Diam) morphological characteristics of bile ducts were extracted to build a CBD diagnosis model to distinguish CBD from SBD and an intrahepatic involvement identification model to classify CBD with/without intrahepatic involvement. The diagnostic performance of the models was compared with that of experienced hepatobiliary surgeons. Results: The CBD diagnosis model using clinical, Shape, and Diam characteristics showed good performance with an AUROC of 0.942 (95% CI: 0.890–0.994), AUPRC of 0.917 (0.855–0.979), accuracy of 0.891, sensitivity of 0.950, and F1-score of 0.864. The model outperformed two experienced surgeons in accuracy, sensitivity, and F1-score. The intrahepatic involvement identification model using clinical, Shape, and Diam characteristics yielded outstanding performance with an AUROC of 0.944 (0.879–1.000), AUPRC of 0.982 (0.947–1.000), accuracy of 0.932, sensitivity of 0.971, and F1-score of 0.957. The models demonstrated generalizable performance on the temporally-independent validation cohort. Conclusions: This study developed two robust quantitative models for distinguishing CBD from SBD and identifying CBD with intrahepatic involvement, respectively, based on morphological characteristics of the bile ducts, showing great potential in risk stratification and surgical planning of CBD.

Vertebrobasilar dissecting aneurysm (VBDA) has high morbidity and mortality rates, and it is an important cause of stroke in young and middle-aged adults.