Deep learning-based differentiation of peripheral high-flow and low-flow vascular malformations in T2-weighted short tau inversion recovery MRI.

BACKGROUND: Differentiation of high-flow from low-flow vascular malformations (VMs) is crucial for therapeutic management of this orphan disease. OBJECTIVE: A convolutional neural network (CNN) was evaluated for differentiation of peripheral vascular malformations (VMs) on T2-weighted short tau inversion recovery (STIR) MRI. METHODS: 527 MRIs (386 low-flow and 141 high-flow VMs) were randomly divided into training, validation and test set for this single-center study. 1) Results of the CNN's diagnostic performance were compared with that of two expert and four junior radiologists. 2) The influence of CNN's prediction on the radiologists' performance and diagnostic certainty was evaluated. 3) Junior radiologists' performance after self-training was compared with that of the CNN. RESULTS: Compared with the expert radiologists the CNN achieved similar accuracy (92% vs. 97%, p = 0.11), sensitivity (80% vs. 93%, p = 0.16) and specificity (97% vs. 100%, p = 0.50). In comparison to the junior radiologists, the CNN had a higher specificity and accuracy (97% vs. 80%, p <  0.001; 92% vs. 77%, p <  0.001). CNN assistance had no significant influence on their diagnostic performance and certainty. After self-training, the junior radiologists' specificity and accuracy improved and were comparable to that of the CNN. CONCLUSIONS: Diagnostic performance of the CNN for differentiating high-flow from low-flow VM was comparable to that of expert radiologists. CNN did not significantly improve the simulated daily practice of junior radiologists, self-training was more effective.

Diagnostic Accuracy of Indocyanine Green Clearance Test for Different Stages of Liver Fibrosis and Cirrhosis.

(1) Background: This study aimed to correlate the indocyanine green clearance (ICG) test with histopathological grades of liver fibrosis and liver cirrhosis to assess its diagnostic accuracy in differentiating normal liver parenchyma from liver fibrosis and liver cirrhosis. (2) Methods: A total of 82 patients who received a histopathological liver examination, imaging, and ICG test within three months were included in this retrospective study. The histopathological level of fibrosis was graded using the Ishak scoring system, and the patients were divided into five categories: no liver fibrosis (NLF), mild liver fibrosis (MLF), advanced liver fibrosis (ALF), severe liver fibrosis (SLF), and liver cirrhosis (LC). The non-parametric Kruskal-Wallis test with post hoc pairwise comparison utilizing Mann-Whitney U tests and Bonferroni adjustment was used to analyze differences in the ICG test results between the patient groups. Cross correlation between the individual fibrosis/cirrhosis stages and the score of the ICG test was performed, and the sensitivity, specificity, and positive and negative predictive values were calculated for each model predicting liver fibrosis/cirrhosis. (3) Results: A significant difference (p ≤ 0.001) between stages of NLF, LF, and LC was found for the ICG parameters (ICG plasma disappearance rate (ICG-PDR) and ICG retention percentage at 15 min (ICG-R15)). The post hoc analysis revealed that NLF significantly differed from SLF (ICG-PDR: p = 0.001; ICG-R15: p = 0.001) and LC (ICG-PDR: p = 0.001; ICG-R15: p = 0.001). ALF also significantly differed from SLF (ICG-PDR: p = 0.033; ICG-R15: p = 0.034) and LC (ICG-PDR: p = 0.014; ICG-R15: p = 0.014). The sensitivity for detection of an initial stage of liver fibrosis compared to no liver fibrosis (Ishak ≥ 1) was 0.40; the corresponding specificity was 0.80. The differentiation of advanced liver fibrosis or cirrhosis (Ishak ≥ 4) compared to other stages of liver fibrosis was 0.75, with a specificity of 0.81. (4) Conclusions: This study shows that the ICG test, as a non-invasive diagnostic test, is able to differentiate patients with no liver fibrosis from patients with advanced liver fibrosis and liver cirrhosis. The ICG test seems to be helpful in monitoring patients with liver fibrosis regarding compensation levels, thus potentially enabling physicians to both detect progression from compensated liver fibrosis to advanced liver fibrosis and cirrhosis and to initiate antifibrotic treatment at an earlier stage.

Application of A U-Net for Map-like Segmentation and Classification of Discontinuous Fibrosis Distribution in Gd-EOB-DTPA-Enhanced Liver MRI.

We aimed to evaluate whether U-shaped convolutional neuronal networks can be used to segment liver parenchyma and indicate the degree of liver fibrosis/cirrhosis at the voxel level using contrast-enhanced magnetic resonance imaging. This retrospective study included 112 examinations with histologically determined liver fibrosis/cirrhosis grade (Ishak score) as the ground truth. The T1-weighted volume-interpolated breath-hold examination sequences of native, arterial, late arterial, portal venous, and hepatobiliary phases were semi-automatically segmented and co-registered. The segmentations were assigned the corresponding Ishak score. In a nested cross-validation procedure, five models of a convolutional neural network with U-Net architecture (nnU-Net) were trained, with the dataset being divided into stratified training/validation (n = 89/90) and holdout test datasets (n = 23/22). The trained models precisely segmented the test data (mean dice similarity coefficient = 0.938) and assigned separate fibrosis scores to each voxel, allowing localization-dependent determination of the degree of fibrosis. The per voxel results were evaluated by the histologically determined fibrosis score. The micro-average area under the receiver operating characteristic curve of this seven-class classification problem (Ishak score 0 to 6) was 0.752 for the test data. The top-three-accuracy-score was 0.750. We conclude that determining fibrosis grade or cirrhosis based on multiphase Gd-EOB-DTPA-enhanced liver MRI seems feasible using a 2D U-Net. Prospective studies with localized biopsies are needed to evaluate the reliability of this model in a clinical setting.

Comparison of sequential CT arterioportography-arteriosplenography with standard cross-sectional imaging and endoscopy in children with portal hypertension.

In this study the diagnostic capability and additional value of sequential CT arterioportography-arteriosplenography (CT AP-AS) in comparison to standard cross-sectional imaging and upper gastrointestinal endoscopy (UGE) in pediatric portal hypertension (PH) was analyzed. Patients with clinical signs of PH who underwent CT AP-AS in combination with additional contrast-enhanced magnetic resonance imaging (CE-MR) and/or contrast-enhanced computed tomography (CE-CT) were included. Two radiologists reviewed independently imaging regarding the capability to prove patency of (1) extrahepatic and intrahepatic main stem portal vein (PV), (2) intrahepatic PV system and (3) splenomesenteric venous axis. Imaging was reviewed for detection of abdominal varices and results were compared to UGE. Main venous supply of varices (PV and/or splenic vein system) and splenorenal shunting were evaluated. 47 imaging studies (20 CT AP-AS, 16 CE-MR, 11 CE-CT) and 12 UGE records of 20 patients were analyzed. CT AP-AS detected significantly more splenorenal shunts (p = 0.008) and allowed more confident characterization of the extra-/intrahepatic PV-system and splenomesenteric veins in comparison to CE-MR (p < 0.001). Extra- and intrahepatic PV-system were significantly more confidently assessed in CT AP-AS than in CE-CT (p = 0.008 and < 0.001 respectively). CT AP-AS was the only modality that detected supply of varices and additional gastric/duodenal varices. In this retrospective study CT AP-AS was superior to standard cross-sectional imaging concerning confident assessment of the venous portosplenomesenteric axis in pediatric patients. CT AP-AS detected additional varices, splenorenal shunting and supply of varices.

Stereotactic Microwave Ablation of Hepatocellular Carcinoma: The Impact of Tumor Size and Minimal Ablative Margin on Therapeutic Success.

BACKGROUND: Microwave ablation (MWA) has gained relevance in the treatment of hepatic malignancies and especially in hepatocellular carcinoma (HCC), and it is an important alternative to surgery. The purpose of the study was to evaluate whether the minimal ablative margin (MAM) or the initial tumor size has a greater effect on the success of stereotactic MWA of HCC regarding the time to local tumor progression (LTP) and overall survival (OS). METHODS: 88 patients, who received stereotactic MWA of 127 tumor lesions with a curative intention were included in this single-center, retrospective study. The MAM was evaluated in a side-by-side comparison of pre- and post-ablative, contrast-enhanced slice imaging. A Cox proportional hazard model with a frailty term was computed to assess the influence of the MAM and the maximum tumor diameter on the time to LTP and the OS. RESULTS: The maximum tumor diameter was identified as a significant positive predictor for LTP (hazard ratio 1.04, 95% CI 1.00-1.08, p = 0.03), but it was not a significant positive predictor for the OS (p = 0.20). The MAM did not have a significant influence on LTP-free survival (p = 0.23) and OS (p = 0.67). CONCLUSION: For the successful stereotactic MWA of HCC, the MAM and maximum tumor diameter might not have an influence on the OS, but the maximum tumor diameter seems to be an independent predictor of the time to LTP.