Comparison of interobserver variations in delineation of target volumes and organs-at-risk for intensity-modulated radiotherapy of nasopharyngeal carcinoma among physicians from different levels of cancer centers / 中国医学物理学杂志

Objective To assess inter-observer variations(IOV)in the delineation of target volumes and organs-at-risk(OAR)for intensity-modulated radiotherapy(IMRT)of nasopharyngeal carcinoma(NPC)among physicians from different levels of cancer centers,thereby providing a reference for quality control in multi-center clinical trials.Methods Twelve patients with NPC of different TMN stages were randomly selected.Three physicians from the same municipal cancer center manually delineated the target volume(GTVnx)and OAR for each patient.The manually modified and confirmed target volume(GTVnx)and OAR delineation structures by radiotherapy experts from the regional cancer center were used as the standard delineation.The absolute volume difference ratio(△V_diff),maximum/minimum volume ratio(MMR),coefficient of variation(CV),and Dice similarity coefficient(DSC)were used to compare the differences in organ delineation among physicians from different levels of cancer centers and among the 3 physicians from the same municipal cancer center.Furthermore,the IOV of GTVnx and OAR among physicians from different levels cancer centers were compared across different TMN stages.Results Significant differences in the delineation of GTVnx were observed among physicians from different levels of cancer centers.Among the 3 physicians,the maximum values of △V_diff,MMR,and CV were 97.23%±83.45%,2.19±0.75,and 0.31±0.14,respectively,with an average DSC of less than 0.7.Additionally,there were considerable differences in the delineation of small-volume OAR such as the left and right optic nerves,chiasm,and pituitary,with average MMR>2.8,CV>0.37,and DSC<0.51.However,relatively smaller differences were observed in the delineation of large-volume OAR such as the brainstem,spinal cord,left and right eyeballs,and left and right mandible,with average△V_diff<42%,MMR<1.55,and DSC>0.7.Compared with the differences among physicians from different levels cancer centers,the differences among the 3 physicians from the municipal cancer center were slightly reduced.Furthermore,there were also differences in the delineation of target volumes for NPC among physicians from different levels cancer centers,depending on the staging of the disease.Compared with the delineation of target volumes for earlier stage patients(stages I or II),the differences among physicians in the delineation of target volumes for advanced stage patients(stages III or IV)were smaller,with average △V_diff and DSC of 98.31%±67.36%vs 69.38%±72.61%(P<0.05)and 0.55±0.08 vs 0.72±0.12(P<0.05),respectively.Conclusion There are differences in the delineation of GTVnx and OAR in radiation therapy for NPC among physicians from different levels of cancer centers,especially in the delineation of target volume(GTVnx)and small-volume OAR for early-stage patients.To ensure the accuracy of multicenter clinical trials,it is recommended to provide unified training to physicians from different levels of cancer centers and review their delineation results to reduce the effect of differences on treatment outcomes.

A semi-supervised network-based tissue-aware contrast enhancement method for CT images / 南方医科大学学报

OBJECTIVE@#To propose a tissue- aware contrast enhancement network (T- ACEnet) for CT image enhancement and validate its accuracy in CT image organ segmentation tasks.@*METHODS@#The original CT images were mapped to generate low dynamic grayscale images with lung and soft tissue window contrasts, and the supervised sub-network learned to recognize the optimal window width and level setting of the lung and abdominal soft tissues via the lung mask. The self-supervised sub-network then used the extreme value suppression loss function to preserve more organ edge structure information. The images generated by the T-ACEnet were fed into the segmentation network to segment multiple abdominal organs.@*RESULTS@#The images obtained by T-ACEnet were capable of providing more window setting information in a single image, which allowed the physicians to conduct preliminary screening of the lesions. Compared with the suboptimal methods, T-ACE images achieved improvements by 0.51, 0.26, 0.10, and 14.14 in SSIM, QABF, VIFF, and PSNR metrics, respectively, with a reduced MSE by an order of magnitude. When T-ACE images were used as input for segmentation networks, the organ segmentation accuracy could be effectively improved without changing the model as compared with the original CT images. All the 5 segmentation quantitative indices were improved, with the maximum improvement of 4.16%.@*CONCLUSION@#The T-ACEnet can perceptually improve the contrast of organ tissues and provide more comprehensive and continuous diagnostic information, and the T-ACE images generated using this method can significantly improve the performance of organ segmentation tasks.

Multi-scale 3D convolutional neural network-based segmentation of head and neck organs at risk / 南方医科大学学报

OBJECTIVE@#To establish an algorithm based on 3D convolution neural network to segment the organs at risk (OARs) in the head and neck on CT images.@*METHODS@#We propose an automatic segmentation algorithm of head and neck OARs based on V-Net. To enhance the feature expression ability of the 3D neural network, we combined the squeeze and exception (SE) module with the residual convolution module in V-Net to increase the weight of the features that has greater contributions to the segmentation task. Using a multi-scale strategy, we completed organ segmentation using two cascade models for location and fine segmentation, and the input image was resampled to different resolutions during preprocessing to allow the two models to focus on the extraction of global location information and local detail features respectively.@*RESULTS@#Our experiments on segmentation of 22 OARs in the head and neck indicated that compared with the existing methods, the proposed method achieved better segmentation accuracy and efficiency, and the average segmentation accuracy was improved by 9%. At the same time, the average test time was reduced from 33.82 s to 2.79 s.@*CONCLUSIONS@#The 3D convolution neural network based on multi-scale strategy can effectively and efficiently improve the accuracy of organ segmentation and can be potentially used in clinical setting for segmentation of other organs to improve the efficiency of clinical treatment.