Preoperative MRI-based deep learning radiomics machine learning model for prediction of the histopathological grade of soft tissue sarcomas / 中华放射学杂志

Objective:To investigate the value of a preoperatively MRI-based deep learning (DL) radiomics machine learning model to distinguish low-grade and high-grade soft tissue sarcomas (STS).Methods:From November 2007 to May 2019, 151 patients with STS confirmed by pathology in the Affiliated Hospital of Qingdao University were enrolled as training sets, and 131 patients in the Affiliated Hospital of Shandong First Medical University and the Third Hospital of Hebei Medical University were enrolled as external validation sets. According to the French Federation Nationale des Centres de Lutte Contre le Cancer classification (FNCLCC) system, 161 patients with FNCLCC grades Ⅰ and Ⅱ were defined as low-grade and 121 patients with grade Ⅲ were defined as high-grade. The hand-crafted radiomic (HCR) and DL radiomic features of the lesions were extracted respectively. Based on HCR features, DL features, and HCR-DL combined features, respectively, three machine-learning models were established by decision tree, logistic regression, and support vector machine (SVM) classifiers. The area under the receiver operating characteristic curve (AUC) was used to evaluate the performance of each machine learning model and choose the best one. The univariate and multivariate logistic regression were used to establish a clinical-imaging factors model based on demographics and MRI findings. The nomogram was established by combining the optimal radiomics model and the clinical-imaging model. The AUC was used to evaluate the performance of each model and the DeLong test was used for comparison of AUC between every two models. The Kaplan-Meier survival curve and log-rank test were used to evaluate the performance of the optimal machine learning model in the risk stratification of progression free survival (PFS) in STS patients.Results:The SVM radiomics model based on HCR-DL combined features had the optimal predicting power with AUC values of 0.931(95%CI 0.889-0.973) in the training set and 0.951 (95%CI 0.904-0.997) in the validation set. The AUC values of the clinical-imaging model were 0.795 (95%CI 0.724-0.867) and 0.615 (95%CI 0.510-0.720), and of the nomogram was 0.875 (95%CI 0.818-0.932) and 0.786 (95%CI 0.701-0.872) in the training and validation sets, respectively. In validation set, the performance of SVM radiomics model was better than those of the nomogram and clinical-imaging models ( Z=3.16, 6.07; P=0.002,<0.001). Using the optimal radiomics model, there was statistically significant in PFS between the high and low risk groups of STS patients (training sets: χ2=43.50, P<0.001; validation sets: χ2=70.50, P<0.001). Conclusion:Preoperative MRI-based DL radiomics machine learning model has accurate prediction performance in differentiating the histopathological grading of STS. The SVM radiomics model based on HCR-DL combined features has the optimal predicting power and was expected to undergo risk stratification of prognosis in STS patients.

Differentiation of benign and malignant lesions of the parotid gland by MRI based imaging features and radiomics nomogram / 中华放射学杂志

Objective:To develop and validate a MRI-based radiomics nomogram combining with radiomics signature and clinical factors for the preoperative differentiation of benign parotid gland tumors (BPGT) and malignant parotid gland tumors (MPGT).Methods:From January 2015 to May 2020, 86 patients with parotid tumors confirmed by surgical pathology in the Affiliated Hospital of Qingdao University were enrolled as training sets, and 35 patients in the University of Hong Kong-Shenzhen Hospital from January 2013 to January 2020 were enrolled as independent external validation sets. The logistic regression was used to establish a clinical-factors model based on demographics and MRI findings. Radiomics features were extracted from preoperative T 1WI and fat-saturated T 2WI (fs-T 2WI), a radiomics signature model was constructed, and a radiomics score (Rad-Score) was calculated. A combined diagnostic model and nomogram combining with the Rad-score and independent clinical factors was constructed using multivariate logistic regression analysis. The receiver operating characteristic (ROC) analysis was used to evaluate the performance of each model and DeLong test was used for comparison of area under the ROC curve (AUC). Results:The logistic regression results showed that deep lobe involvement (OR=3.285, P=0.040) and surrounding tissue invasion (OR=15.919, P=0.013) were independent factors for MPGT and constructed the clinical-factors model. A total of 19 features were extracted from the joint T 1WI and fs-T 2WI to build the radiomics signature model. The combined diagnostic model and nomogram incorporating deep lobe involvement, surrounding tissue invasion and Rad-score were established. The AUCs of the clinical-factors model, radiomics signature model and combined diagnostic model for differentiating BPGT from MPGT for the training and validation sets were 0.758, 0.951, 0.953 and 0.752, 0.941 and 0.964 respectively. The AUCs of the radiomics signature model and the combined diagnostic model were significantly higher than those of the clinical-factors model for both training and validation sets (training set: Z=3.95, 4.31, both P<0.001; validation set: Z=2.16, 2.67, P=0.031, 0.008). There was no statistical difference in AUCs between the radiomics signature model and combined diagnostic model (training set: Z=0.39, P=0.697; validation set: Z=1.10, P=0.273). Conclusions:The MRI-based radiomics signature model and radiomics nomogram incorporating deep lobe involvement, surrounding tissue invasion, and Rad-score showed favorable predictive efficacy for differentiating BPGT from MPGT.

CTdiagnosisofthepulmonarychondroma / 实用放射学杂志

Objective ToanalyzetheCTfeaturesandthediagnosticvalueofpulmonarychondroma.Methods Tencasesofpulmonary chondromaprovenbypathologywereretrospectivelyanalysed.Thenumber,location,size,shape,margin,calcificationpatternandCT valueofthelesions wereanalysedonnonGenhancedandenhanced CTscans.Results Allthe10casesofpulmonarychondroma showedsolitary,mildlylobulated,wellGcircumscribed masses.6lesionswerelocatedintherightlung,and4lesionswereintheleft lung.Thesizeofthelesionsrangedform1.3cm×0.8cmto10.7cm×9.8cm.OnplainCTimages,9lesions(90%)showedvaried calcification,withpunctatecalcificationin8lesionsandringcalcificationin1lesion.OncontrastGenhanced CTimages,6lesions showedslighthomogeneousenhancement(enhancedvalue≤14HU).Conclusion Pulmonarychondromaisusuallylocatedintheperiphery ofthelung.Thenodulehasasmoothboundary,withsignificantcalcificationandslightlyenhancement,whichcouldbehelpfulindiagnosis ofthedisease.

Correlation of radiological charicteristics of Coats’disease with pathology / 实用放射学杂志

Objective To investigate the CT and MRI findings of Coats’disease in comparison with pathology.Methods CT,MRI and ultrasonic features,FFA findings of eight patients of Coats’disease with histo-pathologically confirmed were analyzed retrospectively.CT scanning,routine MRI scanning and ultrasonic examination were performed in all eight patients.Results Unilateral eyeball was involved in all eight cases.On CT scanning,the density of the vitreous body was increased homogeneously which boundary was clear(n=8).The retina was thick(n=8).The anterior chamber depth was shallow(n=5).Multiple calcified foci occured in lens and vitreous body(n=1).The volume of affected eyeball increased(n=1).The affected eyeball shrinked(n=1).The difference of volume of bilateral eyeball was not obvious(n=6). Lens were thick and dislocation(n=5).On MRI scanning,the lesions in the vitreous body showed iso T1 and iso T2 signal (n=6),short T1 and long T2 signal(n=1),long T1 and long T2 signal(n=1).The retina showed short T2 signal(n=3).The vitreous body was filled with lesions(n=5).The lesions looked like‘V’sticked to retina(n= 3).On ultrasonic examination low echo was showed in the vitreous body(n=6),the ball wall bulged(n=8),retinal detachment(n=8).The echo of the ball wall was obviously enhanced,which indica-ted ossification(n=1).Strong echo calcified plaque was showed in one case.FFA showed retinal telangiectasia(n=8),retinal capillary zone(n=2),mutiple chestnut shaped aneurysms(n= 6),retinal neovascularization(n= 1 ).Pathological examination showed retinal telangiectasia with foam macrophages and lipid deposition.Conclusion Coats’disease carries some typical CT and MRI features.To summarize the radiologic features,the findings of FFA and ultrasonic inspection are helpful to diagnosis.