Objective:To explore the differences in the spatial distributions of brain metastases (BMs) from adenocarcinoma and non-adenocarcinoma among non-small cell lung cancer (NSCLS), aiming to provide a reference for developing optimal treatment protocols.Methods:A retrospective analysis was performed on data from 283 patients with NSCLC who underwent radiotherapy at the Zhejiang Cancer Hospital from January 2020 to July 2023. These patients included 191 adenocarcinoma cases with an average age of 62.04 years and 92 non-adenocarcinoma cases with an average age of 63.85 years. CT images with detected BMs of these patients were synchronously deformed and registered into a standard brain template to determine the distribution of BMs in the template. The Dice coefficient was employed to analyze the similarity in the distribution of BMs from adenocarcinoma and non-adenocarcinoma, and two-sample t-test was performed respectively using SPM and Dpabi software to ensure the consistency of the result. For brain regions with the number of BMs greater than or equal to 4, the voxels with a size 2 mm × 2 mm × 2 mm were counted. Results:Non-adenocarcinoma tended to invade the posterior lobe of the cerebellum, precuneus, anterior lobe of the cerebellum, cuneus, middle occipital gyrus, and middle temporal gyrus, with corresponding voxel counts of 2 577, 2 291, 1 947, 1 550, 1 200, and 600, respectively. In contrast, adenocarcinoma was more commonly metastasized to the inferior parietal lobule, posterior lobe of the cerebellum, central frontal gyrus, precentral gyrus, middle temporal gyrus, and precuneus, with the corresponding voxel counts of 211, 201, 194, 186, 137, and 84, respectively.Conclusion:Brain metastases exhibit different distributions between adenocarcinoma and non-adenocarcinoma, and both subtypes prefer to invade specific brain regions.

Objective:To develop and evaluate an unsupervised intensity-modulated radiation therapy (IMRT) automated planning scheme for the Eclipse commercial treatment planning system (TPS), aiming to simulate the manual operation during the whole optimization process.Methods:A retrospective analysis was performed on 25 gastric cancer patients aged 40-60 years who had completed radiotherapy in Zhejiang Cancer Hospital from March 2022 to March 2023. All patients were divided into the training ( n=7) and test sets ( n=18). All patients were treated with the same clinically prescribed dose standard: 45 Gy/25 times. Abdominal CT scan was performed using Philips simulator with a thickness of 5 mm. Based on the deep reinforcement learning (DRL) framework, a multi-agent optimization policy network (MOPN) was proposed to simulate the process of clinical manual planning design and obtain high quality automatic planning according to adjusting multiple optimization objectives. The automatic plan for all cases was generated by code programming using the eclipse scripting application program interface (ESAPI). Wilcoxon signed rank test was used to investigate the significance of the difference between automatic planning and clinical manual planning. Results:After the initial optimization objectives were adjusted by MOPN, the average plan score of all automatic plans was increased from 576.1±221.2 to 1852.8±294.9. Compared with clinical manual plans, the average D max of the spinal cord, the average D mean and V 5 Gy of the liver in the MOPN plans were reduced by 21.4%, 9.8% and 11.5%, respectively. Conclusions:With the help of ESAPI tool, MOPN can realize data interaction with TPS and the automation of IMRT treatment plan for gastric cancer. The trained MOPN can mimic the manual operation of the planner to adjust multiple optimization objectives and gradually improve the plan quality.

Medical physicists play an important role in the delivery of radiotherapy. Compared with China′s mainland, Hong Kong has established a more mature training mode and a more complete management system for medical physics talents. In this article, the authors introduced the current state of medical physics talent training, as well as the recruitment, certification and promotion of medical physicist in Hong Kong by querying the official websites of medical physics organizations, reviewing related literature and interviewing senior medical physicists in Hong Kong. The authors also analyzed the shortcomings in the construction of medical physics talent system in China′s mainland and made valuable suggestions.

Objective:To study the improvement of normal tissue region of interest (ROI) segmentation based on clustering-based multi-Atlas segmentation method, thereby achieving better delineation of organs at risk.Methods:CT images of 100 patients with cervical cancer who had completed treatment in Zhejiang Cancer Hospital during 2019-2020 were selected as the Atlas database. According to the volume characteristic parameters of the organs at risk (bladder, rectum and outer contour), the Atlas database was divided into several subsets by k-means clustering algorithm. The image to be segmented was matched to the corresponding Atlas library for multi-Atlas segmentation. The dice similarity coefficient (DSC) was used to evaluate the segmentation results.Results:Using 30 patients as the test set, the sub-Atlas generated by different clustering methods were compared for the improvement of image segmentation results. Compared with general multi-Atlas segmentation methods, clustering-based multi-Atlas segmentation method significantly improve the segmentation accuracy for the bladder (DSC=0.83±0.09 vs. 0.69±0.15, P<0.001) and the rectum (0.7±0.07 vs. 0.56±0.16, P<0.001), but no statistical significance was observed for left and right femoral head (0.92±0.04, 0.91±0.02) and bone marrow (0.91±0.06). The average segmentation time of clustering-based multi-Atlas segmentation method was shorter than that of the general multi-Atlas segmentation method (2.7 min vs. 6.3 min). Conclusion:The clustering-based multi-Atlas segmentation method can not only reduce the number of Atlas images registered with the image to be segmented, but also can be expected to improve the segmentation effect and obtain higher accuracy.

Objective:Topredict the three-dimensional dose distribution of regions of interest (ROI) with brachytherapy for cervical cancer based on U-Net fully convolutional network, and evaluate the accuracy of prediction model.Methods:First, 100 cases of cervical cancer intracavity combined with interstitial implantation were selected as the entire research data set, and divided into the training set ( n=72), validation set ( n=8), and test set ( n=20). Then the U-Net was used to construct two models based on whether the uterine tandem and the implantation needles were included as the distinguishing factors. Finally, dose distribution of 20 cases in the test set were predicted using the trained model, and comparative analysis was performed. The performance of the model was jointly evaluated by , and the mean absolute deviation (MAD). Results:Compared with the model without the uterine tandem and the implantation needles, the of the rectum was increased by (16.83±1.82) cGy ( P<0.05), and the or of the other ROI were not different significantly (all P>0.05). The MAD of the high-risk clinical target volume, rectum, sigmoid, small bowel, and bladder was increased by (11.96±3.78) cGy, (11.43±0.54) cGy, (24.08±1.65) cGy, (17.04±7.17) cGy and (9.52±4.35) cGy, respectively (all P<0.05). The MAD of the intermediate-risk clinical target volume was decreased by (120.85±29.78) cGy ( P<0.05). The mean value of MAD for all ROI was decreased by (7.8±53) cGy ( P<0.05), which was closer to the actual plan. Conclusions:U-Net fully convolutional network can be used to predict three-dimensional dose distribution of patients with cervical cancer undergoing brachytherapy. Combining the uterine tube with the implantation needles as the input parameters yields more accurate predictions than a single use of the ROI structure as the input.

Objective:To compare the positioning errors of individual head-rest combined with thermoplastic fixation mask versus thermoplastic fixation mask alone in patients with head and neck tumors. Methods:Twenty-eight patients who received irradiation with helical tomotherapy in Cancer Hospital of the University of Chinese Academy of Sciences (Zhejiang Cancer Hospital) between October 2019 and April 2020 were included in this study. They were randomly assigned to receive position fixation with either individual head-rest combined with thermoplastic fixation mask (N1 group, n = 14) or thermoplastic fixation mask alone (N2 group, n = 14). Megavoltage computed tomography (MVCT) scanning registration was used to obtain the positioning errors in translation and rotation (ROLL) in the left-right (X), head-food (Y), and belly-back (Z) directions. There were a total of 841 CT scans, consisting of 425 scans in group N1 and 416 scans in group N2. Results:The positioning errors in X, Y, Z and ROLL directions in the N1 group were (1.37 ± 1.04) mm, (1.38 ± 1.12) mm, (1.47 ± 1.62) mm and (1.47 ± 1.62) ° respectively, and they were (1.57 ± 1.21) mm, (2.10 ± 1.51) mm, (1.61 ± 1.50) mm and (1.40 ± 1.30) ° respectively in the N2 group. There was significant difference in positioning errors in the Y direction between N1 and N2 groups ( P = 0.013). In the N1 group, the outward expansion boundaries in X, Y and Z directions was 4.15 mm, 4.23 mm and 4.81 mm respectively, and it was 4.77, 6.31 and 5.08 mm, respectively in the N2 group. In the X direction, there was significant difference in positioning errors taking 3 mm as the dividing point between N1 and N2 groups ( χ2 = 10.516, P < 0.001). In the Y direction, there was significant difference in positioning errors taking 1, 2 and 3 mm as the dividing points between N1 and N2 groups ( χ2 = 24.889, P < 0.001; χ2 = 42.202, P < 0.001; χ2 = 46.204, P < 0.001). In the Z direction, there was significant difference in positioning errors taking 2 mm as the dividing point between N1 and N2 groups ( χ2 = 7.335, P = 0.007). In the N1 group, the percentage of positioning errors < 3 mm in the X, Y and Z directions was 92%, 90% and 92%, respectively. Conclusion:Compared with thermoplastic fixation mask alone, individual head-rest combined with thermoplastic fixation mask can better effectively improve the positioning stability and reduce positioning errors in patients receiving irradiation with helical tomotherapy for head and neck tumors. The combined method is of certain innovation.

The training program of radiation therapists in the United States has been established early, and the mode of training, qualification and continuing education are relatively complete. Literature review was conducted at home and abroad and United States Department of Labor, American Registry of Radiologic Technologists, American Society of Radiologic Technologists as well as Joint Review Committee on Education in Radiologic Technology websites were reviewed. The training mode, qualification, work content, continuing education and employment situation of American radiotherapists were analyzed, aiming to provide some reference and enlightenment for the establishment of a new model for the training of professional radiologists suitable for the national conditions of China Mainland.

Objective:To evaluate the application of a multi-task learning-based light-weight convolution neural network (MTLW-CNN) for the automatic segmentation of organs at risk (OARs) in thorax.Methods:MTLW-CNN consisted of several layers for sharing features and 3 branches for segmenting 3 OARs. 497 cases with thoracic tumors were collected. Among them, the computed tomography (CT) images encompassing lung, heart and spinal cord were included in this study. The corresponding contours delineated by experienced radiation oncologists were ground truth. All cases were randomly categorized into the training and validation set ( n=300) and test set ( n=197). By applying MTLW-CNN on the test set, the Dice similarity coefficients (DSCs) of 3 OARs, training and testing time and space complexity (S) were calculated and compared with those of Unet and DeepLabv3+ . To evaluate the effect of multi-task learning on the generalization performance of the model, 3 single-task light-weight CNNs (STLW-CNNs) were built. Their structures were totally the same as the corresponding branches in MTLW-CNN. After using the same data and algorithm to train STLW-CNN, the DSCs were statistically compared with MTLW-CNN on the testing set. Results:For MTLW-CNN, the averages (μ) of lung, heart and spinal cord DSCs were 0.954, 0.921 and 0.904, respectively. The differences of μ between MTLW-CNN and other two models (Unet and DeepLabv3+ ) were less than 0.020. The training and testing time of MTLW-CNN were 1/3 to 1/30 of that of Unet and DeepLabv3+ . S of MTLW-CNN was 1/42 of that of Unet and 1/1 220 of that of DeepLabv3+ . The differences of μ and standard deviation (σ) of lung and heart between MTLW-CNN and STLW-CNN were approximately 0.005 and 0.002. The difference of μ of spinal cord was 0.001, but σof STLW-CNN was 0.014 higher than that of MTLW-CNN.Conclusions:MTLW-CNN spends less time and space on high-precision automatic segmentation of thoracic OARs. It can improve the application efficiency and generalization performance of the models.

Objective:To evaluate the clinical efficacy and safety of stereotactic body radiation therapy (SBRT) for stage Ⅰ-Ⅱ non-small cell lung cancer.Methods:Retrospective analysis of patients with early stage lung cancer who received SBRT in Zhejiang Cancer Hospital from 2012 to 2018 was conducted. The Kaplan-Meier method was used for survival analysis. The main endpoints of the study were locoregional control (LRC) and cancer specific survival (CSS).Results:A total of 142 eligible cases were included, with a median BED10100Gy (100-132Gy). The median age was 75.6 years (47.2-89.0 years), among which 75 patients were aged (greater than or equal to 75 years old). The median follow-up time was 31.0 months, for patients< 75 years old and patients ≥ 75 years old. The 5-year LRC were 84.5% and 95.8% respectively, 5-year CSS were 72.4% and 78.6% respectively, for patients< 75 years old and elderly patients. The systemic response was mild during treatment, no grade 4-5 adverse events occurred in all patients. The main acute side effect was radiation pneumonitis (RP) below grade 3. Grade 2 RP appeared in 14 patients (9.9%) after SBRT where grade 3 RP occurred in 2(1.4%). There was no treatment-related mortality in the SBRT group.Conclusions:SBRT is a safe and effective treatment for early primary lung cancer with satisfactory rates of LRC and CSS in 5 years and mild complication, which is similar to previous reports.

Objective:To explore a three-dimensional dose distribution prediction method for the left breast cancer radiotherapy planning based on full convolution network (FCN), and to evaluate the accuracy of the prediction model.Methods:FCN was utilized to achieve three-dimensional dose distribution prediction. First, a volumetric modulated arc therapy (VMAT) plan dataset with 60 cases of left breast cancer was built. Ten cases were randomly chosen from the dataset as the test set, and the remaining 50 cases were used as the training set. Then, a U-Net model was built with the organ structure matrix as inputs and dose distribution matrix as outputs. Finally, the model was adopted to predict the dose distribution of the cases in the test set, and the predicted 3D doses were compared with actual planned results.Results:The mean absolute differences of PTV, ipsilateral lung, heart, whole lung and spinal cord for 10 cases were (119.95±9.04) cGy, (214.02±9.04) cGy, (116.23±30.96) cGy, (127.67±69.19) cGy, and (37.28±18.66) cGy, respectively. The Dice similarity coefficient (DSC) of the prediction dose and the planned dose in the 80% and 100% prescription dose range were 0.92±0.01 and 0.92±0.01. The γ rate of 3 mm/3% in the area of 80% and 10% prescription dose range were 0.85±0.03 and 0.84±0.02. Conclusion:FCN can be used to predict the three-dimensional dose distribution of left breast cancer patients undergoing VMAT.