Abstract The power of computed tomography (CT) radiomics for preoperative prediction of microvascular invasion (MVI) in hepatocellular carcinoma (HCC) demonstrated in current research is variable. This systematic review and meta-analysis aim to evaluate the value of CT radiomics for MVI prediction in HCC, and to investigate the methodologic quality in the workflow of radiomics research. Databases of PubMed, Embase, Web of Science, and Cochrane Library were systematically searched. The methodologic quality of included studies was assessed. Validation data from studies with Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) statement type 2a or above were extracted for meta-analysis. Eleven studies were included, among which nine were eligible for meta-analysis. Radiomics quality scores of the enrolled eleven studies varied from 6 to 17, accounting for 16.7%-47.2% of the total points, with an average score of 14. Pooled sensitivity, specificity, and Area Under the summary receiver operator Characteristic Curve (AUC) were 0.82 (95% CI 0.77-0.86), 0.79 (95% CI 0.75-0.83), and 0.87 (95% CI 0.84-0.91) for the predictive performance of CT radiomics, respectively. Meta-regression and subgroup analyses showed radiomics model based on 3D tumor segmentation, and deep learning model achieved superior performances compared to 2D segmentation and non-deep learning model, respectively (AUC: 0.93 vs. 0.83, and 0.97 vs. 0.83, respectively). This study proves that CT radiomics could predict MVI in HCC. The heterogeneity of the included studies precludes a definition of the role of CT radiomics in predicting MVI, but methodology warrants uniformization in the radiology community regarding radiomics in HCC.

OBJECTIVES: This study aimed to determine the concordance between CT and nucleic acid testing in diagnosing coronavirus disease (COVID-19) outside its district of origin (Wuhan, China). METHODS: Twenty-three consecutive patients with COVID-19, confirmed by nucleic acid testing, were enrolled from two designated hospitals outside the district of disease origin. We collected clinical, laboratory, and CT data and assessed the concordance between CT manifestations and nucleic acid test results by comparing the percentage of patients with and without abnormal CT findings. Furthermore, using Chi-square tests, we analyzed the differences in CT manifestations between patients with and without an exposure history or symptoms. RESULTS: Multiple ground-glass opacities (GGOs), with or without consolidation, were observed on the initial CT scans of 19 patients (82.6%), whereas the remaining 4 (17.4%) showed no CT abnormalities, indicating that the initial chest CT findings were not entirely concordant with the nucleic acid test results in diagnosing COVID-19. Among the latter 4 patients, we observed multiple GGOs with and without consolidation in 2 patients on the follow-up chest CT scans taken on days 7 and 14 after admission, respectively. The remaining 2 patients showed no abnormalities on the follow-up CT scans. Furthermore, abnormal CT findings were found more frequently in patients who had been exposed to COVID-19 in its district of origin than in those who had not been exposed and in symptomatic patients than in asymptomatic patients (all p<0.05). CONCLUSIONS: Patients with positive results on nucleic acid testing may or may not have the abnormal CT manifestations that are frequently found in symptomatic patients with a history of exposure to the district of COVID-19 origin.

OBJECTIVE: To determine whether the gross tumor volume of resectable gastric adenocarcinoma on multidetector computed tomography could predict the presence of regional lymph node metastasis and could determine N categories. MATERIALS AND METHODS: A total of 202 consecutive patients with gastric adenocarcinoma who had undergone gastrectomy 1 week after contrast-enhanced multidetector computed tomography were retrospectively identified. The gross tumor volume was evaluated on multidetector computed tomography images. Univariate and multivariate analyses were performed to determine whether the gross tumor volume could predict regional lymph node metastasis, and the Mann-Whitney U test was performed to compare the gross tumor volume among N categories. Additionally, a receiver operating characteristic analysis was performed to identify the accuracy of the gross tumor volume in differentiating N categories. RESULTS: The gross tumor volume could predict regional lymph node metastasis (p<0.0001) in the univariate analysis, and the multivariate analyses indicated that the gross tumor volume was an independent risk factor for regional lymph node metastasis (p=0.005, odds ratio=1.364). The Mann-Whitney U test showed that the gross tumor volume could distinguish N0 from the N1-N3 categories, N0-N1 from N2-N3, and N0-N2 from N3 (all p<0.0001). In the T1-T4a categories, the gross tumor volume could differentiate N0 from the N1-N3 categories (cutoff, 12.3 cm3), N0-N1 from N2-N3 (cutoff, 16.6 cm3), and N0-N2 from N3 (cutoff, 24.6 cm3). In the T4a category, the gross tumor volume could differentiate N0 from the N1-N3 categories (cutoff, 15.8 cm3), N0-N1 from N2-N3 (cutoff, 17.8 cm3), and N0-N2 from N3 (cutoff, 24 cm3). CONCLUSION: The gross tumor volume of resectable gastric adenocarcinoma on multidetector computed tomography could predict regional lymph node metastasis and N categories.

OBJECTIVE: To determine the associations of liver lobe-based magnetic resonance diffusion-weighted imaging findings using multiple b values with the presence and Child-Pugh class of cirrhosis in patients with hepatitis B. METHODS: Seventy-four cirrhotic patients with hepatitis B and 25 healthy volunteers underwent diffusion-weighted imaging using b values of 0, 500, 800 and 1000 sec/mm2. The apparent diffusion coefficients of individual liver lobes for b(0,500), b(0,800) and b(0,1000) were derived from the signal intensity averaged across images obtained using b values of 0 and 500 sec/mm2, 0 and 800 sec/mm2, or 0 and 1000 sec/mm2, respectively, and were statistically analyzed to evaluate cirrhosis. RESULTS: The apparent diffusion coefficients for b(0,500), b(0,800) and b(0,1000) inversely correlated with the Child-Pugh class in the left lateral liver lobe, the left medial liver lobe, the right liver lobe and the caudate lobe (r=-0.35 to -0.60, all p<0.05), except for the apparent diffusion coefficient for b(0,1000) in the left medial liver lobe (r=-0.17, p>0.05). Among these parameters, the apparent diffusion coefficient for b(0,500) in the left lateral liver lobe best differentiated normal from cirrhotic liver, with an area under the receiver operating characteristic curve of 0.989. The apparent diffusion coefficient for b(0,800) in the right liver lobe best distinguished Child-Pugh class A from B-C and A-B from C, with areas under the receiver operating characteristic curve of 0.732 and 0.747, respectively. CONCLUSION: Liver lobe-based apparent diffusion coefficients for b(0,500) and b(0,800) appear to be associated with the presence and Child-Pugh class of liver cirrhosis.

OBJECTIVE: The aim of this study was to determine whether and how the diameter of the vein that gives rise to the inflowing vein of the esophageal and gastric fundic varices secondary to posthepatitic cirrhosis, as measured with multidetector-row computed tomography, could predict the varices and their patterns. METHODS: A total of 106 patients with posthepatitic cirrhosis underwent multidetector-row computed tomography. Patients with and without esophageal and gastric fundic varices were enrolled in Group 1 and Group 2, respectively. Group 1 was composed of Subgroup A, consisting of patients with varices, and Subgroup B consisted of patients with varices in combination with portal vein-inferior vena cava shunts. The diameters of the originating veins of veins entering the varices were reviewed and statistically analyzed. RESULTS: The originating veins were the portal vein in 8% (6/75) of patients, the splenic vein in 65.3% (49/75) of patients, and both the portal and splenic veins in 26.7% (20/75) of patients. The splenic vein diameter in Group 1 was larger than that in Group 2, whereas no differences in portal vein diameters were found between groups. In Group 1, the splenic vein diameter in Subgroup A was larger than that in Subgroup B. A cut-off splenic vein diameter of 8.5 mm achieved a sensitivity of 83.3% and specificity of 58.1% for predicting the varices. For discrimination of the varices in combination with and without portal vein-inferior vena cava shunts, a cut-off diameter of 9.5 mm achieved a sensitivity of 66.7% and specificity of 60.0%. CONCLUSION: The diameter of the splenic vein can be used to predict esophageal and gastric fundic varices and their patterns.

OBJECTIVE: The features of earthquake-related head injuries may be different from those of injuries obtained in daily life because of differences in circumstances. We aim to compare the features of head traumas caused by the Sichuan earthquake with those of other common head traumas using multidetector computed tomography. METHODS: In total, 221 patients with earthquake-related head traumas (the earthquake group) and 221 patients with other common head traumas (the non-earthquake group) were enrolled in our study, and their computed tomographic findings were compared. We focused the differences between fractures and intracranial injuries and the relationships between extracranial and intracranial injuries. RESULTS: More earthquake-related cases had only extracranial soft tissue injuries (50.7% vs. 26.2%, RR=1.9), and fewer cases had intracranial injuries (17.2% vs. 50.7%, RR = 0.3) compared with the non-earthquake group. For patients with fractures and intracranial injuries, there were fewer cases with craniocerebral injuries in the earthquake group (60.6% vs. 77.9%, RR = 0.8), and the earthquake-injured patients had fewer fractures and intracranial injuries overall (1.5 + 0.9 vs. 2.5 +1.8; 1.3 + 0.5 vs. 2.1 + 1.1). Compared with the non-earthquake group, the incidences of soft tissue injuries and cranial fractures combined with intracranial injuries in the earthquake group were significantly lower (9.8% vs. 43.7%, RR = 0.2; 35.1% vs. 82.2%, RR = 0.4). CONCLUSION: As depicted with computed tomography, the severity of earthquake-related head traumas in survivors was milder, and isolated extracranial injuries were more common in earthquake-related head traumas than in non-earthquake-related injuries, which may have been the result of different injury causes, mechanisms and settings.

OBJECTIVE: To determine the features of earthquake-related pelvic crush fractures versus non-earthquake fractures with digital radiography and multidetector row computed tomography. METHODS: One hundred and sixty-seven survivors with pelvic crush fractures in the 2008 Sichuan earthquake were entered in our study as the earthquake-related group (139 underwent digital radiography, 28 underwent multidetector row computed tomography); 70 victims with non-earthquake pelvic fractures were enrolled into this study as the non-earthquake group (54 underwent digital radiography, 16 underwent multidetector row computed tomography). Data were reviewed retrospectively between groups, focusing on anatomic distributions, status of pelvic bone fractures, numbers of pelvic bones involved, and classification of pelvic ring fractures according to the Tile classification system. RESULTS: Pelvic fractures occurred more frequently in the pubis in the earthquake-related group than in the non-earthquake group (135/167, 81% vs. 48/70, 69%). In addition, comminuted fractures were more common in the earthquake-related group than in the non-earthquake group (55/167, 33% vs. 10/70, 14%). Multiple fractures were less common in the earthquake-related group than in the non-earthquake group (81/167, 49% vs. 46/70, 66%). Regarding the classification of pelvic ring fractures, Type C predominantly composed of subtype C3 occurred more frequently (64/167, 38% vs. 12/70, 17%), and Type A was less common in the earthquake-related group than in the non-earthquake group (31/167, 19% vs. 23/70, 32%). All differences were statistically significant (p<0.05). No difference was found in Type B fractures between the groups (72/167, 43% vs. 35/70, 50%). CONCLUSION: Earthquake-related pelvic crush fractures can be characterized by a high incidence of pelvic fractures occurring in the pubis, comminuted fractures, and Type C fractures predominantly composed by subtype C3, despite a low incidence of multiple fractures.

PURPOSE: Massive earthquakes are harmful to humankind. This study of a historical cohort aimed to investigate the difference between earthquake-related crush thoracic traumas and thoracic traumas unrelated to earthquakes using a multi-detector Computed Tomography (CT). METHODS: We retrospectively compared an earthquake-exposed cohort of 215 thoracic trauma crush victims of the Sichuan earthquake to a cohort of 215 non-earthquake-related thoracic trauma patients, focusing on the lesions and coexisting injuries to the thoracic cage and the pulmonary parenchyma and pleura using a multi-detector CT. RESULTS: The incidence of rib fracture was elevated in the earthquake-exposed cohort (143 vs. 66 patients in the non-earthquake-exposed cohort, Risk Ratio (RR) = 2.2; p<0.001). Among these patients, those with more than 3 fractured ribs (106/143 vs. 41/66 patients, RR=1.2; p<0.05) or flail chest (45/143 vs. 11/66 patients, RR=1.9; p<0.05) were more frequently seen in the earthquake cohort. Earthquake-related crush injuries more frequently resulted in bilateral rib fractures (66/143 vs. 18/66 patients, RR= 1.7; p<0.01). Additionally, the incidence of non-rib fracture was higher in the earthquake cohort (85 vs. 60 patients, RR= 1.4; p<0.01). Pulmonary parenchymal and pleural injuries were more frequently seen in earthquake-related crush injuries (117 vs. 80 patients, RR=1.5 for parenchymal and 146 vs. 74 patients, RR = 2.0 for pleural injuries; p<0.001). Non-rib fractures, pulmonary parenchymal and pleural injuries had significant positive correlation with rib fractures in these two cohorts. CONCLUSIONS: Thoracic crush traumas resulting from the earthquake were life threatening with a high incidence of bony thoracic fractures. The ribs were frequently involved in bilateral and severe types of fractures, which were accompanied by non-rib fractures, pulmonary parenchymal and pleural injuries.