Objective To explore the value of radiomics model and deep learning model based on unenhanced MRI in predicting microvascular invasion(MVI)of hepatocellular carcinoma(HCC)preoperatively.Methods A total of 189 patients with postopera-tive pathologically confirmed HCC from two centers were retrospectively selected,of which 119 cases from Zhujiang Hospital of Southern Medical University were used as the training set[60 cases with negative MVI,59 cases with positive MVI],and 70 cases from Shenzhen People's Hospital were used as the external test set[38 cases with negative MVI and 32 cases with positive MVI].Clinical indicators were analyzed by univariate and multivariate logistic regression analysis and the independent predictors of positive MVI were screened.Deep transfer learning(DTL)and traditional radiomics methods were used to construct radiomics model and deep learning model based on unenhanced MRI.The predictive performances of each model were compared using receiver operating charac-teristic(ROC)curves and area under the curve(AUC).DeLong test was employed to compare statistical differences in performance of the models.Results Alkaline phosphatase(ALP)and prothrombin time(PT)were independent predictors of positive MVI(P<0.05).The deep learning model based on T2WI had the best predictive efficacy,with AUC of 0.779[95%confidence interval(CI)0.696-0.863]and 0.741(95%CI 0.620-0.861)in the training set and external test set,respectively,and there were statistically significant differences compared with the radiomics model and the clinical model based on T1WI(P<0.05).Conclusion Deep learning model based on T2WI has a certain application value in preoperative noninvasive prediction of MVI status in HCC patients.

Objective To explore the value of radiomics model and deep learning model based on unenhanced MRI in predicting microvascular invasion(MVI)of hepatocellular carcinoma(HCC)preoperatively.Methods A total of 189 patients with postopera-tive pathologically confirmed HCC from two centers were retrospectively selected,of which 119 cases from Zhujiang Hospital of Southern Medical University were used as the training set[60 cases with negative MVI,59 cases with positive MVI],and 70 cases from Shenzhen People's Hospital were used as the external test set[38 cases with negative MVI and 32 cases with positive MVI].Clinical indicators were analyzed by univariate and multivariate logistic regression analysis and the independent predictors of positive MVI were screened.Deep transfer learning(DTL)and traditional radiomics methods were used to construct radiomics model and deep learning model based on unenhanced MRI.The predictive performances of each model were compared using receiver operating charac-teristic(ROC)curves and area under the curve(AUC).DeLong test was employed to compare statistical differences in performance of the models.Results Alkaline phosphatase(ALP)and prothrombin time(PT)were independent predictors of positive MVI(P<0.05).The deep learning model based on T2WI had the best predictive efficacy,with AUC of 0.779[95%confidence interval(CI)0.696-0.863]and 0.741(95%CI 0.620-0.861)in the training set and external test set,respectively,and there were statistically significant differences compared with the radiomics model and the clinical model based on T1WI(P<0.05).Conclusion Deep learning model based on T2WI has a certain application value in preoperative noninvasive prediction of MVI status in HCC patients.

Objective To propose a method for abdominal multi-organ segmentation assisted by multi-phase CT synthesis.Methods Multi-phase CT synthesis for synthesizing high-quality CT images was used to increase the information details for image segmentation.A transformer block was introduced to help to capture long-range semantic information in cooperation with perceptual loss to minimize the differences between the real image and synthesized image.Results The model was trained using multi-phase CT dataset of 526 total cases from Nanfang Hospital.The mean maximum absolute error(MAE)of the synthesized non-contrast CT,venous phase contrast-enhanced CT(CECT),and delay phase CECT images from arterial phase CECT was 19.192±3.381,20.140±2.676 and 22.538±2.874,respectively,which were better than those of images synthesized using other methods.Validation of the multi-phase CT synthesis-assisted abdominal multi-organ segmentation method showed an average dice coefficient of 0.847 for the internal validation set and 0.823 for the external validation set.Conclusion The propose method is capable of synthesizing high-quality multi-phase CT images to effectively reduce the errors in registration between different phase CT images and improve the performance for segmentation of 13 abdominal organs.

Objective To observe the value of radiomics models based on gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid(Gd-EOB-DTPA)enhanced hepatobiliary phase(HBP)MRI for assessing clinical pathological stage of hepatic fibrosis(HF).Methods Data of 240 patients with pathologically/clinically diagnosed and clinical pathological staged HF who underwent Gd-EOB-DTPA enhanced MR examination were retrospectively analyzed.The liver-to-muscle signal intensity ratio(SIR1)and liver-to-spleen signal intensity ratio(SIR2)were measured based on HBP images.Radiomics features of HBP images were extracted and screened to construct radiomics models.The signal intensity ratio(SIR)-radiomics combined models were constructed based on SIR and radiomics signatures.Receiver operating characteristic(ROC)curves were drawn to evaluate the efficacy of each model for assessing clinical pathological stage of HF.Results The area under the curve(AUC)of SIR1 and SIR2 models for assessing clinical pathological stage of HF were 0.63-0.70 and 0.65-0.71,respectively.The most effective radiomics model for assessing HF,significant HF,advanced HF and early cirrhosis was support vector machine(SVM),SVM,light gradient boosting machine and K-nearest neighbor model,respectively,with the AUC in validation set of 0.87,0.82,0.81 and 0.80,respectively,while the AUC of SIR-radiomics combined models in validation set of 0.88,0.82,0.82 and 0.81,respectively.Conclusion The radiomics models based on Gd-EOB-DTPA enhanced HBP MRI were helpful for assessing clinical pathological stage of HF.Combining with HBP SIR could improve their efficacy.

Objective To propose a method for abdominal multi-organ segmentation assisted by multi-phase CT synthesis.Methods Multi-phase CT synthesis for synthesizing high-quality CT images was used to increase the information details for image segmentation.A transformer block was introduced to help to capture long-range semantic information in cooperation with perceptual loss to minimize the differences between the real image and synthesized image.Results The model was trained using multi-phase CT dataset of 526 total cases from Nanfang Hospital.The mean maximum absolute error(MAE)of the synthesized non-contrast CT,venous phase contrast-enhanced CT(CECT),and delay phase CECT images from arterial phase CECT was 19.192±3.381,20.140±2.676 and 22.538±2.874,respectively,which were better than those of images synthesized using other methods.Validation of the multi-phase CT synthesis-assisted abdominal multi-organ segmentation method showed an average dice coefficient of 0.847 for the internal validation set and 0.823 for the external validation set.Conclusion The propose method is capable of synthesizing high-quality multi-phase CT images to effectively reduce the errors in registration between different phase CT images and improve the performance for segmentation of 13 abdominal organs.

Objective:To explore the value of a radiomics nomogram based on T 1WI for prediction of the relapse of osteosarcoma after surgery within 1 year from multicenter data. Methods:The imaging and clinical data of 107 patients with pathologica1ly confirmed osteosarcoma who received neoadjuvant chemotherapy before surgery from 6 hospitals from January 2009 to October 2017 were retrospectively analyzed. A training cohort consisted of 75 patients from firstly enrolled 4 hospitals and an independent validation cohort of 32 patients from other 2 hospitals. Pretreatment T 1WI was used to extract radiomics features. Least absolute shrinkage and selection operator (LASSO) regression was applied to reduce the dimension and then the radiomics signature was constructed to predict the relapse of osteosarcoma after surgery within 1 year in training cohort. Independent clinical risk factors were screened using one-way logistic regression, and then a radiomics nomogram incorporated the radiomics signature and MRI characteristics was developed by multivariate logistic regression. The predictive nomogram was evaluated using receiver operating characteristic (ROC) curve in the training cohort, and validated in the independent validation cohort. The calibration curve was used to evaluate the agreement between prediction and actual observation and the decision curve was used to demonstrate the clinical usefulness. Results:Based on T 1WI from multicenter institutions, the radiomics signature was built using 2 valuable selected features that were significantly associated with relapse within 1 year. Two selected features included 1 gray-level co-occurrence matrices (GLCM) feature (L_G_1.0_GLCM_homogeneity1, LASSO coefficient 3.122) and 1 gray-level run length matrix (GLRLM) feature (GLRLM_RP, LASSO coefficient -2.474). The prediction nomogram including radiomics signature and MRI characteristics (joint invasion and perivascular involvement) showed good discrimination with the area under the ROC curve of 0.884 and 0.821 in the training and validation cohorts, respectively. The calibration curve showed that the nomogram achieved good agreement between prediction and actual observation. Decision curve analysis demonstrated that the radiomics nomogram was clinically useful when the threshold probability was greater than 21%. Conclusion:The radiomics nomogram based on T 1WI can be used as a non-invasive quantitative tool to predict relapse of osteosarcoma within 1 year before treatment, which provides support for clinical decision-making in osteosarcoma.

Objective To investigate the impact of spatial distribution of emphysema on pulmonary function and clinical severity in patients with chronic obstructive pulmonary disease (COPD).Methods Chest MSCT was performed in 84 patients with COPD.The percentage of low attenuation areas were automatically calculated for each lobe with the COPD analysis software.The correlations between low-attenuation area percent(LAA％) of each lobe and the parameters of pulmonary function were analyzed.According to the pulmonary emphysema heterogeneity index,the patients were divided into non-emphysema group and emphysema group,and the latter were divided into the predominantly upper lobe subgroup and the predominantly lower-lobe subgroup.The differences of pulmonary function parameters between the 2 groups and clinical severity between the 2 subgroups in emphysema group were compared.Results LAA％ of each lobe was negatively correlated with single breath carbon monoxide diffusing capacity to predicted value ratio (DLcoSB％pred).Except for the right middle lobe,LAA％ of the remaining lobes were negatively correlated with forced expiratory volume in one second/forced vital capacity (FEV1/FVC) and FEV1％pred,while positively correlated with resonance frequency (Fres).The differences of pulmonary function parameters between the 2 groups were statistically significant (all P ＜ 0.05).The differences of FEV1％pred and FEV1/FVC between the 2 subgroups in emphysema group were statistically significant (all P ＜0.05),and the differences of Fres and DLcoSB％pred between the 2 subgroups were not statistically significant (all P＞0.05).The difference of clinical severity between the 2 subgroups was statistically significant (x2 =4.17,P=0.041).Conclusion The spatial distribution of pulmonary emphysema affects pulmonary function and clinical severity in COPD patients.The patients with the predominantly lower-lobe emphysema are relatively worse in lung function,and more serious in the clinical severity.

Objective To investigate the MRI features of combined hepatocellular-cholangiocarcinoma(cHCC-CC).Methods The MR imaging features of 8 patients with cHCC-CC confirmed by histopathology were analyzed retrospectively.And 3 patients among them performed MR dynamic contrast scanning with gadoxetic acid(Gd-EOB-DTPA)-a hepatocyte-specific MRI contrast agent.Results MRI showed hypointensity on T1 WI and hyperintensity on T2 WI of tumor's solid components.Dynamic MRI enhancement demonstrated peripheral hyperenhancement of the lesion (n =7) and global hyperenhancement(n=1) in arterial phase,progressively delayed or continued enhancement of the central area in portal venous phase and equilibrium phase,meanwhile the peripheral hyperenhancement of the lesion in arterial phase washed out(n=6),edge enhancement continued(n=1),and enhancing area extended progressively(n=1) which the peripheral hyperenhancement in arterial phase remained.Enhancing pseudocapsule were observed in equilibrium phase(n=6).3 patients injected with Gd-EOB-DTPA presented a target-like appearance in hepatobiliary phase,defined as a peripheral hypointense rim compared to the central area with delayed enhancement of the lesion.Conclusion Dynamic MRI can be a certain value in the preoperative dagnosis of cHCC CC.

Objective To analyze the correlation between quantitative evaluation of high-resolution CT (HRCT) and pulmonary function tests (PFT) in chronic obstructive pulmonary disease (COPD), and evaluate the characteristics of PFT in different CT phenotypes of COPD. Methods 65 patients with COPD underwent CT chest scan and PFT. The dimensions of subsubsegmental apical bronchi of right upper lobes were measured , which included the following parameters:airway wall thickness (T), total airway diameter (D), and pulmonary artery (PA), ratio of airway wall thickness and pulmonary artery (T/PA), thickness diameter ratio (TDR), percentage of wall area (WA%) and percentage of airway luminal area (Ai%). The percentage of total lung area occupied by low attenuation area (LAA%) was measured by lung density software. The relationship between the measurements and PFT parameters were analyzed. Patients were classified into three phenotypes based on the presence of emphysema or bronchial wall thickening evaluated by HRCT. Phenotype A: without or with little emphysema (LAA% < 25%), with or without bronchial wall thickening; Phenotype E: emphysema without bronchial wall thickening (LAA%≥25%, T/PA < 30%); Phenotype M: emphysema with bronchial wall thickening (LAA%≥25%, T/PA≥30%). The characteristics of different CT phenotypes were analyzed. Results LAA% had a good correlation with PFT parameters. PA, T/PA, WA% and Ai% also had correlation with PFT parameters. There was 53 patients of Phenotype A, 12 patients of Phenotype M. PA, T/PA, LAA%and PFT were different in two groups. Conclusions LAA%and airway quantitative measurements in HRCT have certain correlation with PFT parameters. Patients with different CT phenotypes differs in PFT parameters , which can provide reference for clinical diagnosis and treatment.

Various primary and secondary tumors are commonly seen in the liver and have different histological types, but some types of tumors are rare. Computed tomography (CT) and magnetic resonance imaging (MRI) can show the size, morphology, and density or signal of intrahepatic lesions, their association with surrounding tissues, and tumor blood supply. Meanwhile, MRI has a high resolution for soft tissue. This article summarizes the CT and MRI findings of rare liver tumors, in order to improve the diagnostic level of rare liver tumors and give full play to the role of imaging examination in the diagnosis and treatment of liver tumors.