Objective:To evaluate the value of the proposed interpretable denoising deep learning model-grouped sharing convolutional attention-visual transformer（GSCAViT）for classifying normal fetal echocardiograms.Methods:A total of 2 501 images from 358 fetuses who underwent cardiac ultrasound examinations at Xiangya Second Hospital of Central South University from January to November 2024 were retrospectively analyzed. GSCAViT was constructed based on fetal echocardiograms from the three-vessel and trachea view，apical four-chamber view，long-axis view of the aortic arch，bicaval view，left ventricular outflow tract view，three-vessel view and right ventricular outflow tract view were compared with both baseline and improved models in the validation set to assess the performance of the classification echocardiography in terms of accuracy，precision，recall and F1-score. Its generalizability across test sets was assessed using the area under the ROC curve（AUC），sensitivity，specificity and F1-score. The impact of image features was interpreted using SHapley Additive exPlanations（SHAP）.The effectiveness of the GSCA module was compared through visual analysis，image parameter metrics and classification performance.Results:The GSCAViT model achieved classification performance for fetal echocardiograms second only to MaxViT，with an accuracy of 97.1%，precision of 97.1%，recall of 97.0%，and an F1-score of 97.0%. In the E10，E20 and E8 test sets，the AUCs of GSCAViT for the prediction of 7 types of fetal echocardiograms were 0.994，0.928 and 0.932，the sensitivities were 99.4%，81.3% and 72.9%，the specificities were 99.7%，96.8% and 94.8%，the F1-scores were 99.4%，81.3% and 67.6%，respectively. SHAP visualization indicated that the model could identify key structural features within the images. Images processed by the denoising-guided group-sharing convolutional attention module best captured and enhanced important regional features，achieving the highest contrast-to-noise ratio，peak signal-to-noise ratio and optimal classification performance，which demonstrated the module's effectiveness.Conclusions:The proposed GSCAViT model exhibits superior performance in classifying seven types of normal fetal echocardiograms compared to the baseline and some improved models. Furthermore，SHAP visualization enhances the interpretability of the classification results，and visual comparisons，image parameter analyses，as well as classification performance metrics confirming the effectiveness of the denoising-guided group-sharing convolutional attention module in the visual transformer model.

Objective:To predict the occurrence of chronic radiation enteritis (CRE) in cervical cancer patients by developing a prediction model based on the combination of radiomic features derived from magnetic resonance imaging (MRI) scans and clinical parameters, in order to provide a reference for clinicians to determine the prognosis of these patients and offer them individualized diagnosis and treatment.Methods:A retrospective analysis was conducted on 111 cervical cancer patients who received radical radiotherapy at the First Affiliated Hospital of Anhui Medical University. Radiological features were extracted from the T1-weighted MRI images of local lesions of cervical cancer obtained before the radiotherapy. Features were selected using the least absolute shrinkage and selection operator (LASSO) to obtain the radiomics score. The radiomics scores and clinical parameters were assessed using univariate and multivariate logistic regression analyses, followed by the establishment of nomograms. The ability of radiomics to achieve CRE prediction was assessed using the area under the curve (AUC) and the calibration and decision curves.Results:Multivariate logistic regression analysis result revealed that the independent risk factors for identifying CRE in patients included radiomics score ( HR: 17.457, 95% CI: 5.540-55.009, P<0.001), tumor volume ( HR: 3.617, 95% CI: 1.293-10.115, P=0.014), and pelvic lymph node metastasis ( HR: 3.559, 95% CI: 1.013-12.501, P=0.048). The model combining radiomics and clinical data demonstrated high performance, with its AUCs of the training and validation groups (0.888 and 0.870, respectively) higher than those of the radiomics model (0.842 and 0.804, respectively) and the clinical data model (0.721 and 0.704, respectively). The analyses of calibration and decision curves confirmed the application value of clinical radiomic nomograms. Conclusions:The model combining radiomics and clinical data allows for accurate CRE prediction. Therefore, radiomic features have the potential to serve as a promising imaging biomarker for CRE.

Objective:To predict the occurrence of chronic radiation enteritis (CRE) in cervical cancer patients by developing a prediction model based on the combination of radiomic features derived from magnetic resonance imaging (MRI) scans and clinical parameters, in order to provide a reference for clinicians to determine the prognosis of these patients and offer them individualized diagnosis and treatment.Methods:A retrospective analysis was conducted on 111 cervical cancer patients who received radical radiotherapy at the First Affiliated Hospital of Anhui Medical University. Radiological features were extracted from the T1-weighted MRI images of local lesions of cervical cancer obtained before the radiotherapy. Features were selected using the least absolute shrinkage and selection operator (LASSO) to obtain the radiomics score. The radiomics scores and clinical parameters were assessed using univariate and multivariate logistic regression analyses, followed by the establishment of nomograms. The ability of radiomics to achieve CRE prediction was assessed using the area under the curve (AUC) and the calibration and decision curves.Results:Multivariate logistic regression analysis result revealed that the independent risk factors for identifying CRE in patients included radiomics score ( HR: 17.457, 95% CI: 5.540-55.009, P<0.001), tumor volume ( HR: 3.617, 95% CI: 1.293-10.115, P=0.014), and pelvic lymph node metastasis ( HR: 3.559, 95% CI: 1.013-12.501, P=0.048). The model combining radiomics and clinical data demonstrated high performance, with its AUCs of the training and validation groups (0.888 and 0.870, respectively) higher than those of the radiomics model (0.842 and 0.804, respectively) and the clinical data model (0.721 and 0.704, respectively). The analyses of calibration and decision curves confirmed the application value of clinical radiomic nomograms. Conclusions:The model combining radiomics and clinical data allows for accurate CRE prediction. Therefore, radiomic features have the potential to serve as a promising imaging biomarker for CRE.

Objective:To evaluate the value of the proposed interpretable denoising deep learning model-grouped sharing convolutional attention-visual transformer（GSCAViT）for classifying normal fetal echocardiograms.Methods:A total of 2 501 images from 358 fetuses who underwent cardiac ultrasound examinations at Xiangya Second Hospital of Central South University from January to November 2024 were retrospectively analyzed. GSCAViT was constructed based on fetal echocardiograms from the three-vessel and trachea view，apical four-chamber view，long-axis view of the aortic arch，bicaval view，left ventricular outflow tract view，three-vessel view and right ventricular outflow tract view were compared with both baseline and improved models in the validation set to assess the performance of the classification echocardiography in terms of accuracy，precision，recall and F1-score. Its generalizability across test sets was assessed using the area under the ROC curve（AUC），sensitivity，specificity and F1-score. The impact of image features was interpreted using SHapley Additive exPlanations（SHAP）.The effectiveness of the GSCA module was compared through visual analysis，image parameter metrics and classification performance.Results:The GSCAViT model achieved classification performance for fetal echocardiograms second only to MaxViT，with an accuracy of 97.1%，precision of 97.1%，recall of 97.0%，and an F1-score of 97.0%. In the E10，E20 and E8 test sets，the AUCs of GSCAViT for the prediction of 7 types of fetal echocardiograms were 0.994，0.928 and 0.932，the sensitivities were 99.4%，81.3% and 72.9%，the specificities were 99.7%，96.8% and 94.8%，the F1-scores were 99.4%，81.3% and 67.6%，respectively. SHAP visualization indicated that the model could identify key structural features within the images. Images processed by the denoising-guided group-sharing convolutional attention module best captured and enhanced important regional features，achieving the highest contrast-to-noise ratio，peak signal-to-noise ratio and optimal classification performance，which demonstrated the module's effectiveness.Conclusions:The proposed GSCAViT model exhibits superior performance in classifying seven types of normal fetal echocardiograms compared to the baseline and some improved models. Furthermore，SHAP visualization enhances the interpretability of the classification results，and visual comparisons，image parameter analyses，as well as classification performance metrics confirming the effectiveness of the denoising-guided group-sharing convolutional attention module in the visual transformer model.

Objective To explore the distributions of dose differences and their correlations with tumor types in independent three-dimensional dose verification for intensity-modulated radiotherapy(IMRT)plan,and to establish prediction models.Methods The fixed field IMRT plans of 180 patients with head and neck,chest,and abdominal tumors were collected.Independent sample t-test was used to analyze the consistency between the initial dose calculated in treatment planning system and the verified dose for evaluating the feasibility of independent validation.The distributions of planning target volume(PTV)dose differences among different tumor types were analyzed.The correlations of PTV dose differences with conformity index(CI)and homogeneity index were analyzed using correlation analysis and multiple linear regression method.Results The PTV dose differences for head and neck,chest,and abdominal tumors were within±1.2%,and the average 3%/3 mm gamma passing rate between the initial dose and the verified dose was higher than 99.5%,indicating good consistency in dose distribution between independent validation and treatment planning system.The distributions of PTV dose differences were different in head and neck,chest,and abdominal plans,and were significantly correlated with CI which could be used to preliminarily determine the PTV dose differences in independent validation.The multiple regression equation based on CI and homogeneity index could be used to estimate the dose verification differences of IMRT plan.Conclusion Independent verification can quickly achieve pre-treatment validation for radiotherapy plans,improving the efficiency of quality control of radiotherapy plans.The estimation based on dose differences provides guidance for optimizing radiotherapy plans.