Objective To construct a global model of cardiac surface motion based on average intensity projection(AIP)image.Methods The cine magnetic resonance imaging data were divided into training set and test set for model construction and validation.The datum points were obtained on the AIP surface point cloud,and the corresponding points of each cardiac phase surface point cloud were found based on surface features.Principal component analysis was used to extract feature information,establish the mapping relationship between the datum points and the corresponding points,and construct a model for predicting each temporal phase surface point cloud from the AIP surface point cloud.Results The RMSE of the model on test set for corresponding point prediction ranged from(0.209±0.020)mm to(0.841±0.074)mm,while the Euclidean distance for each time phase surface point cloud prediction ranged from(1.399±0.029)mm to(1.658±0.100)mm.Conclusion The proposed global model exhibit high accuracy and can provide a reference for image segmentation and clinical treatments.

Objective To develop a combined model integrating radiomics and 3D deep learning features for improving the predictive efficacy of overall survival in non-small cell lung cancer(NSCLC)patients undergoing radiotherapy,thereby providing a foundation for optimizing individualized radiotherapy strategies.Methods A retrospective analysis was conducted on 522 NSCLC patients from 3 centers.Radiomics features were extracted from the tumor region of interest on radiotherapy planning CT scans,and a 3D-SE-ResNet was constructed to extract deep learning features.Following feature extraction,features were selected via univariate Cox analysis and Lasso-Cox regression,and a combined model was established by fusing the two feature types through principal component analysis.The discriminative ability of the model was evaluated using the concordance index(C-index)and the area under the receiver operating characteristic curve(AUC),while the risk stratification efficacy was verified by Kaplan-Meier survival analysis.Results The predictive performance of deep learning features was significantly superior to that of radiomics features(C-index:0.73 vs 0.65).The combined model achieved the highest predictive performance in the training set,internal test set,and external test set(C-index:0.74,0.69,0.72 respectively),with higher AUC values for predicting 1-year,2-year,and 3-year OS than either single model.Kaplan-Meier analysis showed significant differences in survival between the high-and low-risk groups(Log-rank test,P<0.001),and calibration curves indicated good consistency between predicted and actual survival outcomes.Conclusion The combined model integrating radiomics and 3D deep learning features can accurately predict survival outcomes in NSCLC patients undergoing radiotherapy.The multi-center validation results support its potential application in prognosis stratification for individualized radiotherapy.

Objective To develop a combined model integrating radiomics and 3D deep learning features for improving the predictive efficacy of overall survival in non-small cell lung cancer(NSCLC)patients undergoing radiotherapy,thereby providing a foundation for optimizing individualized radiotherapy strategies.Methods A retrospective analysis was conducted on 522 NSCLC patients from 3 centers.Radiomics features were extracted from the tumor region of interest on radiotherapy planning CT scans,and a 3D-SE-ResNet was constructed to extract deep learning features.Following feature extraction,features were selected via univariate Cox analysis and Lasso-Cox regression,and a combined model was established by fusing the two feature types through principal component analysis.The discriminative ability of the model was evaluated using the concordance index(C-index)and the area under the receiver operating characteristic curve(AUC),while the risk stratification efficacy was verified by Kaplan-Meier survival analysis.Results The predictive performance of deep learning features was significantly superior to that of radiomics features(C-index:0.73 vs 0.65).The combined model achieved the highest predictive performance in the training set,internal test set,and external test set(C-index:0.74,0.69,0.72 respectively),with higher AUC values for predicting 1-year,2-year,and 3-year OS than either single model.Kaplan-Meier analysis showed significant differences in survival between the high-and low-risk groups(Log-rank test,P<0.001),and calibration curves indicated good consistency between predicted and actual survival outcomes.Conclusion The combined model integrating radiomics and 3D deep learning features can accurately predict survival outcomes in NSCLC patients undergoing radiotherapy.The multi-center validation results support its potential application in prognosis stratification for individualized radiotherapy.

Objective To construct a global model of cardiac surface motion based on average intensity projection(AIP)image.Methods The cine magnetic resonance imaging data were divided into training set and test set for model construction and validation.The datum points were obtained on the AIP surface point cloud,and the corresponding points of each cardiac phase surface point cloud were found based on surface features.Principal component analysis was used to extract feature information,establish the mapping relationship between the datum points and the corresponding points,and construct a model for predicting each temporal phase surface point cloud from the AIP surface point cloud.Results The RMSE of the model on test set for corresponding point prediction ranged from(0.209±0.020)mm to(0.841±0.074)mm,while the Euclidean distance for each time phase surface point cloud prediction ranged from(1.399±0.029)mm to(1.658±0.100)mm.Conclusion The proposed global model exhibit high accuracy and can provide a reference for image segmentation and clinical treatments.

Objective:To evaluate the feasibility of predicting lung cancer target position by online optical surface motion monitoring.Methods:CT images obtained in different ways of stereotactic body radiotherapy (SBRT) plans from 16 lung cancer cases were selected for experimental simulation. The planned CT and the original target position were taken as the reference, and the 10 phases of CT in four dimension CT and each cone beam (CBCT) were taken as the floating objects, on which the floating target location was delineated. The binocular visual surface imaging method was used to obtain point cloud data of reference and floating image body surface, while the point cloud feature information was extracted for comparison. Based on the random forest algorithm, the feature information difference and the corresponding target area position difference were fitted, and an online prediction model of the target area position was constructed.Results:The model had a high prediction success rate for the target position. The variance explainded and root mean squared error ( RMSE) of left-right, superior-inferior, anterior-posterior directions were 99.76%, 99.25%, 99.58%, and 0.0447 mm, 0.0837 mm, 0.0616 mm, respectively. Conclusion:The online monitoring of lung SBRT target position proposed in this study is feasible, which can provide reference for online monitoring and verification of target position and dose evaluation in clinical radiotherapy.

Objective:To propose a markerless beam's eye view (BEV) tumor tracking algorithm, which can be applied to megavolt (MV) images with poor image quality, multi-leaf collimator (MLC) occlusion and non-rigid deformation.Methods:Window template matching, image structure transformation and demons non-rigid registration method were used to solve the registration problem in MV images. The quality assurance (QA) plan was generated in the phantom and executed after manually setting the treatment offset on the accelerator, and 682 electronic portal imaging device (EPID) images in the treatment process were collected as fixed images. Meanwhile, the digitally reconstructured radiograph (DRR) images corresponding to the field angle in the planning system were collected as floating images to verify the accuracy of the algorithm. In addition, a total of 533 images were collected from 21 cases of lung tumor treatment data for tumor tracking study, providing quantitative results of tumor location changes during treatment. Image similarity was used for third-party verification of tracking results.Results:The algorithm could cope with different degrees (10%-80%) of image missing. In the phantom verification, 86.8% of the tracking errors were less than 3 mm, and 80% were less than 2 mm. Normalized mutual information (NMI) varied from 1.182±0.026 to 1.202±0.027 ( P<0.005) before and after registration and the change of Hausdorff distance (HD) was from 57.767±6.474 to 56.664±6.733 ( P<0.005). The case results were predominantly translational (-6.0 mm to 6.2 mm), but non-rigid deformation still existed. NMI varied from 1.216±0.031 to 1.225±0.031 ( P<0.005) before and after registration and the change of HD was from 46.384±7.698 to 45.691±8.089 ( P<0.005). Conclusions:The proposed algorithm can cope with different degrees of image missing and performs well in non-rigid registration with data missing images which can be applied in different radiotherapy technologies. It provides a reference idea for processing MV images with multi-modality, partial data and poor image quality.

ObjectiveBy exploring the volatile components， polysaccharide composition and changes in the contents of five carbohydrate components of Polygonatum cyrtonema rhizoma before and after processing， and then the effect of yellow rice wine on the odour formation of P. cyrtonema rhizoma was investigated. MethodThe volatile components of P. cyrtonema rhizoma before and after processing were detected by headspace gas chromatography-mass spectrometry（HS-GC-MS）， and sample data were subjected to principal component analysis（PCA） and orthogonal partial least squares-discriminant analysis（OPLS-DA） using SIMCA 14.1， then the differences between these components of P. cyrtonema rhizoma before and after processing were screened according to the principle of variable importance in the projection（VIP） value>1. Crude carbohydrate components in raw and wine-processed P. cyrtonema rhizoma were subjected to oxime and silylation， the carbohydrate components were analyzed by gas chromatography-mass spectrometry（GC-MS/MS）， and the relative contents of various components were calculated by peak area normalization， then quantitative analysis of four carbohydrate components was also carried out. ResultA total of 23 volatile components were identified from the raw products and the wine-processed products， including 15 components in raw products and 20 components in wine-processed products. Among them， 2-methylbutyraldehyde and isovaleraldehyde had a sweet odor and their contents increased after processing， but the contents of hexanal and caproic acid decreased， new components such as 2-acetylfuran and 5-methylfuranal were produced after processing. PCA and OPLS-DA results showed that there were significant differences between raw products and the wine-processed products， a total of 13 differential compounds were screened out， of which 7 showed an upward trend in relative content and 6 showed a downward trend. A total of 7 carbohydrate components， including 5 monosaccharides and 2 disaccharides， were identified in raw products and the wine-processed products. The results of determination showed that the contents of fructose， glucose， mannose and sucrose in P. cyrtonema rhizoma increased after wine-processing， and their increases were 4.54， 1.51， 2.93， 3.66 times， respectively. ConclusionAfter processing， the increase of aromatic flavor of P. cyrtonema rhizoma may be related to the increase of the contents of aldehydes such as 2-methylbutyraldehyde and isovaleraldehyde， while the decrease of raw flavor may be related to the decrease of the contents of volatile components such as hexanal and hexanoic acid， the increase of sweet flavor may be related to the increase of the contents of monosaccharides and oligosaccharides such as fructose and sucrose.

ObjectiveBy comparing the difference of volatile components of the decoction pieces before and after being processed by braising method of Jianchangbang and steaming method included in the 2020 edition of Chinese Pharmacopoeia， the influence of processing methods on the flavor formation of Polygoni Multiflori Radix （PMR） was compared. MethodHeadspace-gas chromatography-mass spectrometry （HS-GC-MS） was used to detect the volatile components of 30 batches of PMR samples from 3 origins with 3 processing methods. The GC was performed under programmed temperature （starting temperature of 40 ℃， rising to 150 ℃ at 5 ℃·min^-1， and then rising to 195 ℃ at 10 ℃·min^-1） with high purity helium as carrier gas and the split ratio of 10∶1. Mass spectrometry conditions were electron impact ion source （EI） and the detection range of m/z 50-650， the peak area normalization method was used to calculate the relative mass fraction of each component. The chromaticity values of different processed products were measured by a precision colorimeter， the relationship between chromaticity values and relative contents of volatile components was investigated by OriginPro 2021， principal component analysis （PCA） and orthogonal partial least squares-discriminant analysis （OPLS-DA） were performed on the sample data by SIMCA14.1. The differential components of different processed products of PMR were screened according to the principle of variable importance in the projection （VIP） value>1.5， and the material basis of different odor formation of PMR and its processed products was explored. ResultA total of 59 volatile components were identified， among which 34 were raw products， 33 were braised products， and 27 were steamed products. PCA and OPLS-DA results showed that there were significant differences between the three， but there was no significant difference between samples from different origins of the same processing method. Color parameters of a^*， b^*， E^*ab had no significant correlation with contents of volatile components， while L^* was negatively correlated with contents of 2-methyl-2-butenal， 2-methyltetrahydrofuran-3-one and 2，3-dihydro-3，5-dihydroxy-6-methyl-4（H）-pyran-4-one （P<0.05）. The contents of pungent odor components such as caproic acid， nonanoic acid and synthetic camphor decreased after processing， while the contents of sweet flavor components such as 2-methyl-2-butenal， furfural and 5-hydroxymethylfurfural increased after processing， and the contents of furfural， 5-methyl-2-furanmethanol， 5-hydroxymethylfurfural and other aroma components in the braised products were significantly higher than that in the steamed products. ConclusionHS-GC-MS can quickly identify the volatile substance basis that causes the different odors of PMR and its processed products. The effect of processing methods on the odor is greater than that of origin. There is a significant correlation between the color parameter of L^* and contents of volatile components， the "raw" taste of PMR may be related to volatile components such as caproic acid， pelargonic acid and synthetic camphor， the "flavor" after processing may be related to the increase of the contents of 2-methyl-2-butenal， furfural， 5-hydroxymethylfurfural， methyl maltol and furfuryl alcohol.

Objective:To propose a machine learning-based markerless beam′s eye view (BEV) tumor tracking algorithm that can be applied to low-quality megavolt (MV) images with multileaf collimator (MLC)-induced occlusion and non-rigid deformation.Methods:This study processed the registration of MV images using the window template matching method and end-to-end unsupervised network Voxelmorph and verified the accuracy of the tumor tracking algorithm using dynamic chest models. Phantom QA plans were executed after the treatment offset was manually set on the accelerator, and 682 electronic portal imaging device (EPID) images obtained during the treatment were collected as fixed images. Moreover, the digitally reconstructed radiography (DRR) images corresponding to the portal angles in the planning system were collected as floating images for the study of target volume tracking. In addition, 533 pairs of EPID and DRR images of 21 lung tumor patients treated with radiotherapy were collected to conduct the study of tumor tracking and provide quantitative result of changes in tumor locations during the treatment. Image similarity was used for third-party validation of the algorithm.Results:The algorithm could process images with different degrees (10%-80%) of data missing and performed well in non-rigid registration of images with data missing. As shown by the phantom verification, 86.8% and 80% of the tracking errors were less than 3 mm and less than 2 mm, respectively, and the normalized mutual information (NMI) varied from 1.18 ± 0.02 to 1.20 ± 0.02 after registration ( t = -6.78, P = 0.001). The tumor motion of the clinical cases was dominated by translation, with an average displacement of 3.78 mm and a maximum displacement of 7.46 mm. The registration result of the cases showed the presence of non-rigid deformations, and the corresponding NMI varied from 1.21 ± 0.03 before registration to 1.22 ± 0.03 after registration ( t = -2.91, P = 0.001). Conclusions:The tumor tracking algorithm proposed in this study has reliable tracking accuracy and high robustness and can be used for non-invasive and real-time tumor tracking requiring no additional equipment and radiation dose.

Objective:To propose a method of image similarity measurement based on structure information and intuitionistic fuzzy set and measure the similarity between CT image and CBCT image of radiotherapy plan positioning, aiming to objectively measure the setup errors.Methods:A total of four pre-registration images of a nasopharyngeal carcinoma patient on the cross-sectional and sagittal planes and a pelvic tumor patient on the cross-sectional and coronal planes were randomly selected. Five methods were used to quantify the setup errors, including correlation coefficient, mean square error, image joint entropy, mutual information and similarity measure method.Results:All five methods could describe the deviation to a certain extent. Compared with other methods, the similarity measure method showed a stronger upward trend with the increase of errors. After normalization, the results of five types of error increase on the cross-sectional plane of the nasopharyngeal carcinoma patient were 0.553, 0.683, 1.055, 1.995, 5.151, and 1.171, 1.618, 1.962, 1.790, 3.572 on the sagittal plane, respectively. The results of other methods were between 0 and 2 after normalization, and the results of different errors of the same method slightly changed. In addition, the method was more sensitive to the soft tissue errors.Conclusions:The image similarity measurement method based on structure information and intuitionistic fuzzy set is more consistent with human eye perception than the existing evaluation methods. The errors between bone markers and soft tissues can be objectively quantified to certain extent. The soft tissue deviation reflected by the setup errors is of significance for individualized precision radiotherapy.