Objective:To effectively quantify and evaluate the quality of different deformation registration algorithms, in order to enhance the possibility of implementing deformation registration in clinical practice.Methods:The Jacobian determinant mean (JDM) is proposed based on the Jacobian determinant (JD) of displacement vector field (DVF), and the Jacobian determinant error (DJDE) is introduced by incorporating the JD of the inverse DVF. The optical flow method (OF-DIR) and fast demons method with elastic regularization (FD-DIR) were tested on nasopharyngeal and lung cancer datasets. Finally, JDM and DJDE with the Jacobian determinant negative percentage (JDNP), inverse consistency error (ICE) and normalized mean square error (NMSE) were used to evaluate the registration algorithms and compare the differences evaluation indicators in different tumor images and different algorithms, and the receiver operating curve (ROC) was analyzed in evaluation.Results:In lung cancer, OF-DIR outperformed FD-DIR in terms of JDM, NMSE, DJDE and ICE, and the difference was statistically significant( z = -2.24, -4.84, t = 4.01, 6.54, P<0.05). In nasopharyngeal carcinoma, DJDE, ICE and NMSE of OF-DIR were superior to FD-DIR, and the difference was statistically significant ( t = 4.46, -7.49, z = -2.22, P<0.05), but there was no significant difference in JDM ( P>0.05). In lung cancer and nasopharyngeal carcinoma, JDNP of OF-DIR was worse than that of FD-DIR, and the difference was statistically significant ( z = -4.29, -4.02, P<0.01). In addition, DJDE is more specific and sensitive on ROC curve (AUC=0.77), and has different performance result for tumor images at different sites. Conclusions:The JDM and DJDE evaluation metrics proposed are effective for deformation registration algorithms. OF-DIR is suitable for both lung cancer and nasopharyngeal carcinoma, while the influence of organ motion on the registration effect should be considered when using FD-DIR.

Objective:To investigate the necessity of adaptive re-planning during radiotherapy for nasopharyngeal carcinoma (NPC) and its impact on dose improvement.Methods:Clinical data of 89 NPC patients admitted to Sun Yat-sen University Cancer Center from July 2014 to December 2017 were retrospectively analyzed. All patients received 25+7 rounds of adaptive re-planning during radiotherapy. Plan-A was defined as the initial CT scan-based 25-fraction radiotherapy plan, while plan-B was defined as the re-planned 7-fraction radiotherapy plan based on a subsequent CT scan. The changes in the target and parotid gland volumes were compared between plan-A and plan-B. Plan-I was a one-time simulation of plan-A extended to 32 fraction radiotherapy plan, and plan-II was generated through registration and fusion of the plan-A and plan-B for adaptive re-planning. The differences in dose metrics, homogeneity index (HI), conformity index (CI), and dose to organs at risk (OAR) were compared between plan-I and plan-II. Statistical analysis was performed by using paired t-test. Results:Compared with plan-A, the gross tumor volume of massive bleeding lesions (GTV nx) and parotid gland volume of plan-B were decreased by 13.14% and 11.12%, respectively (both P<0.001). While planning clinical target volume of metastatic lymph nodes (PCTV nd) of plan-B was increased by 7.75%( P<0.001). There were significant changes in the lymph nodes of plan-A and plan-B. The D mean, D 5%, D 95% of massive bleeding lesions planning target volume (PTV nx) and D 5% of high risk planning target volume (PTV1) in plan-II were all significantly higher than those in plan-I (all P<0.05). The CI of PTV nx and PTV1 in plan-II was closer to 1 than that in plan-I. In all assessed OAR, the D mean, D 50%, and D max of plan-II were significantly lower than those of plan-I (all P<0.05). Conclusions:During radiotherapy, NPC patients may experience varying degrees of primary tumor shrinkage, parotid gland atrophy, and lymph node changes. It is necessary to deliver re-planning and significantly improve the dose of target areas and OAR.

OBJECTIVE: To identify the problems in clinical radiotherapy planning for cervical cancer through quantitative evaluation of the radiotherapy plans to improve the quality of the plans and the radiotherapy process. METHODS: We selected the clinically approved and administered radiotherapy plans for 227 cervical cancer patients undergoing external radiotherapy at Sun Yat-sen University Cancer Center from May, 2019 to January, 2022. These plans were transferred from the treatment planning system to the Plan IQ^TM workstation. The plan quality metrics were determined based on the guidelines of ICRU83 report, the GEC-ESTRO Working Group, and the clinical requirements of our center and were approved by a senior clinician. The problems in the radiotherapy plans were summarized and documented, and those with low scores were re-planned and the differences were analyzed. RESULTS: We identified several problems in the 277 plans by quantitative evaluation. Inappropriate target volume selection (with scores < 60) in terms of GTV, PGTV (CI) and PGTV (V_{66 Gy}) was found in 10.6%, 65.2%, and 1% of the plans, respectively; and the PGTV (CI), GTV, and PCTV (D_98%, HI) had a score of 0 in 0.4%, 10.1%, 0.4%, 0.4% of the plans, respectively. The problems in the organs at risk (OARs) involved mainly the intestines (the rectum, small intestine, and colon), found in 20.7% of the plans, and in occasional cases, the rectum, small intestine, colon, kidney, and the femoral head had a score of 0. Senior planners showed significantly better performance than junior planners in PGTV (V_{60 Gy}, D_98%), PCTV (CI), and CTV (D_98%) (P≤0.046) especially in terms of spinal cord and small intestine protection (P≤0.034). The bowel (the rectum, small intestine and colon) dose was significantly lower in the prone plans than supine plans (P < 0.05), and targets coverage all met clinical requirements. Twenty radiotherapy plans with low scores were selected for re-planning. The re-planned plans had significantly higher GTV (D_min) and PTV (V_{45 Gy}, D_98%) (P < 0.05) with significantly reduced doses of the small intestines (V_{40 Gy} vs V_{30 Gy}), the colon (V_{40 Gy} vs V_{30 Gy}), and the bladder (D_35%) (P < 0.05). CONCLUSION Quantitative evaluation of the radiotherapy plans can not only improve the quality of radiotherapy plan, but also facilitate risk management of the radiotherapy process.
Humans ; Female ; Uterine Cervical Neoplasms/radiotherapy* ; Rectum ; Colon ; Kidney ; Organs at Risk

Objective:To analyze the clinically acceptable and reproducible bladder and rectum volumes of prostate cancer patients during radiotherapy under bladder and bowel preparation, aiming to provide quantitative indicators for bowel and bladder preparation before and after radiotherapy.Methods:Clinical data of 275 prostate cancer patients with strict bladder and bowel preparation and completion of whole course radical radiotherapy at Sun Yat-sen University Cancer Center from April 2015 to December 2020 were retrospectively analyzed. Patients were scanned with cone beam CT (CBCT) before each treatment and the setup error was recorded. Sixty-six patients were selected by simple random sampling and the bladder and rectum on daily CBCT was outlined using MIM software. The relationship between the ratio of daily bladder or rectum volume to the planned bladder or rectum volume (relative value of volume) and setup error was analyzed. Quantitative data were expressed as mean±SD. Normally distributed data were analyzed by paired t-test while non-normally distributed data were assessed by Kruskal-Wallis test.Results:The bladder and rectum volume on planning CT were (370.87±110.04) ml and (59.94±25.07) ml of 275 patients. The bladder and rectum volumes on planning CT were (357.51±107.38) ml and (65.28±35.37) ml respectively of the 66 selected patients with 1611 sets of CBCT images. And the bladder and rectum volumes on daily CBCT were (258.96±120.23) ml and (59.95 ± 30.40) ml. The bladder volume of patients was decreased by 3.59 ml per day on average during the treatment and 0.37 ml for the rectum volume. According to the bladder volume on planning CT, all patients were divided into three groups: <250 ml, 250-450 ml and >450 ml groups. The relative value of volume in the 250-450 ml group during the course of radiotherapy was the smallest. And the setup error in the superior and inferior (SI) direction was (0.28±0.24) cm and (0.19±0.17) cm in the left and right (LR) direction, significantly lower than those in the other two groups (both P≤0.027). According to the rectum volume on planning CT, all patients were divided into four groups: <50 ml, 50-<80 ml, 80-120 ml and >120 ml groups. The <50 ml group had the smallest relative value of volume during radiotherapy, and the setup error in the SI direction was (0.26±0.22) cm and (0.24±0.22) cm in the anterior and posterior (AP) direction, significantly smaller than those in the other groups (both P≤0.003). The setup errors in the SI, LR, AP directions of the enrolled 66 patients were (0.30±0.25) cm, (0.20±0.18) cm and (0.28±0.27) cm, respectively. Among them, the relative value of bladder volume in the AP direction was (0.73±0.37) in the setup error <0.3 cm group, which was statistically different from those in the setup error 0.3-0.5 cm and >0.5 cm groups (both P<0.05). Conclusion:Under the bladder and bowel preparation before planning CT, the appropriate bladder and rectum volumes are in the range of 250-450 ml and <50 ml, which yields higher reproducibility and smaller setup error.

Objective:To analyze the dosimetric differences of volumetric modulated arc therapy (VMAT) plans for lung cancer caused by different dose calculation algorithms and radiation field settings and thus to provide a reference for designing clinical VMAT plans for lung cancer.Methods:This study randomly selected 20 patients with lung cancer and divided them into four groups of VMAT plans, namely, a group adopting two fields and two arcs based on the AAA algorithm (2F2A_AAA), a group employing two fields and two arcs based on the AXB algorithm (2F2A_AXB), a group using two fields and two arcs based on the MC algorithm (2F2A_MC), and a group adopting one field and two arcs based on the MC algorithm (1F2A_MC). Then, this study evaluated the target coverage, high-dose control, dose homogeneity index (HI), conformity index (CI), and organs at risk (OARs) of the plans using different algorithms and radiation field settings.Results:The planning target volume (PTV) results of two fields combined with two arcs (2F2A) of three groups using different algorithms are as follows. 2F2A_MC achieved better results in both D1% and V 95% (the relative volume of the target volume surrounded by 95% of the prescribed dose) of planning gross target volume (PGTV) than 2F2A_AAA (D1%: t=-2.44, P=0.03; V95%:z=-2.04, P=0.04) and 2F2A_AXB (D1%: t=2.34, P=0.03; z=-3.21, P < 0.01). 2F2A_AXB outperformed 2F2A_AAA ( z=-3.66, P < 0.01) and was comparable to 2F2A_MC in terms of the CI of PGTV. Regarding OARs, 2F2A_AXB and 2F2A_MC decreased the V5 Gy of the whole lung by 0.68% ( z=-2.69, P=0.01) and 3.05% ( z=-3.52, P < 0.01), respectively compared to 2F2A_AAA. 2F2A_AXB achieved a whole-lung Dmean of 1776.44 cGy, which was superior to that of 2F2A_MC ( t=2.67, P=0.02) and 2F2A_AAA ( t=8.62, P < 0.01). Compared to 2F2A_AAA and 2F2A_MC, 2F2A_AXB decreased the V20 Gy of Body_5 mm by 1.45% ( z=-3.88, P < 0.01) and 2.01% ( z=-3.66, P < 0.01), respectively. The results of the two groups with different field settings showed that 1F2A_MC was superior to 2F2A_MC in both the CI of PTV1 and the HI of PTV2 (CI: t=2.61, P=0.02; HI: z=-2.20, P=0.03). Moreover, 1F2A_MC increased the Dmean of the whole lung by 26.29 cGy compared to 2F2A_MC ( t=2.28, P=0.04). Conclusions:Regarding the design of VMAT plans for lung cancer, the MC algorithm is suitable for the target priority and the AXB algorithm is suitable for the OAR priority. When only the MC algorithm is available, it is recommended to choose 1F2A in the case of target priority and select 2F2A in the case of OAR priority.

Objective:To provide evidence for the selection of fixation devices and CTV to PTV margins (M ptv) in precision radiotherapy for pelvic tumors by analyzing three fixation devices in precision radiotherapy for prostate cancer. Methods:From April 2015 to December 2020, 133 prostate cancer patients treated with pelvic drainage area irradiation in our center were retrospectively analyzed. The patients were fixed with 1.2m vacuum bag (n=39), 1.8m vacuum bag (n=44) and personalized prone plate by our center (n=50). Each patient was asked to complete our bowel and bladder preparation process before positioning and radiotherapy. The registration of CBCT to planned CT before each treatment adopted the same registration box and algorithm. Setup errors in the SI, LR and AP directions under qualified bowel and bladder conditions were recorded. Setup errors in three directions under three fixation devices and corresponding M ptv values were analyzed. The correlation between setup errors with age and body mass index (BMI) was analyzed. Results:Analysis of 3333 setup errors data showed: in the SI and LR directions, the mean setup errors of 1.2m vacuum bag (3.26mm, 2.34mm) were greater than those of 1.8m vacuum bag (2.51mm, P<0.001; 1.90mm, P<0.001), and personalized prone plate (3.07mm, P=0.066; 2.10 mm, P=0.009). In the AP direction, the mean setup errors of 1.2m vacuum bag (supine)(2.20mm) were smaller than those of 1.8m vacuum bag (3.33mm, P<0.001) and personalized prone plate (3.61mm, P<0.001). The setup errors of 1.8m vacuum bag in all directions were smaller than those of personalized prone plate (P≤0.028). According to Van Herk's expansion formula, the M ptv of 1.2m vacuum bag in three directions was approximately 4 mm. The M ptv of 1.8m vacuum bag and personalized prone plate in the SI and LR directions was approximately 3 mm, and more than 5 mm in the AP direction. The setup errors were not correlated with age or BMI. Conclusions:From the setup errors results of three devices, 1.8m vacuum bag is the best, followed by personalized prone plate. And supine position is better than prone position in the AP direction.

Objective:To standardize the naming of organ at risk (OAR) and target area during cervical cancer radiotherapy based on AAPM TG-263.Methods:After self-programming of Matlab software to implement the reading and resolution of radiotherapy structure files, the naming of each substructure was automatically output, recorded and restored. After naming all substructures, the structure names were classified by keywords. According to TG-263, a standard naming conversion table of OAR and target area was developed, and the classified structure names were standardized through procedures. Finally, the standardized named radiotherapy structure files were output and imported into the treatment planning system (TPS).Results:The radiation structure of 144 patients with cervical cancer was successfully transformed and displayed correctly in TPS. Before the transformation, the naming of OAR and target area lacked of uniform norms and standards, and the naming of the same structure significantly differed. After the transformation, 43 naming methods of OAR and 74 naming methods of the target area were unified into 20 and 8 naming methods, which were more convenient for staff understanding and communication.Conclusion:The standardization of cervical cancer radiotherapy structure naming can reduce the inconsistency of naming and provide reference for the standardized naming of pelvic tumors.

Objective:Based on the AAPM TG-263, a Content-Based Standardizing Nomenclatures (CBSN) was proposed to explore the feasibility of its standardization verification for organs at risk (OAR) of nasopharyngeal carcinoma (NPC).Methods:The radiotherapy structure files of 855 patients with nasopharyngeal carcinoma (NPC) receiving intensity-modulated radiotherapy (IMRT) from 2017 to 2019(15 of whom showed clinical anomalous structures) were retrospectively collected and processed. The Matlab self-developed software was used to obtain the image position, geometric features, first-order gray histogram, and the Gray-level Co-occurrence Matrix′s texture features of the OAR contour outlined by the doctor to establish the CBSN Location Verification model and CBSN Knowledge Library. Fisher discriminant analysis was employed to establish a CBSN OAR classification model, which was evaluated using self-validation, cross-validation, and external validation, respectively.Results:99%(69/70) of the simulated anomalous structures were outside the 90% reference range of the CBSN Knowledge Library and the characteristic parameters significantly differed among different OARs (all P<0.001). The accuracy rates of self-validation, cross-validation and external verification of the CBSN OAR classification model were 92.1%, 92.0% and 91.8%, respectively. Fourteen cases of clinical abnormal structures were successfully detected by CBSN with an accuracy rate of 93%(14/15). In the simulation test, the accuracy of the left and right location verification reached 100%, such as detecting the right eye lens named Len_L. Conclusion:CBSN can be used for OAR verification of NPC, providing reference for multi-center cooperation and standardized radiotherapy of NPC patients.

OBJECTIVE: To investigate the accuracy of automatic segmentation of organs at risk (OARs) in radiotherapy for nasopharyngeal carcinoma (NPC). METHODS: The CT image data of 147 NPC patients with manual segmentation of the OARs were randomized into the training set (115 cases), validation set (12 cases), and the test set (20 cases). An improved network based on three-dimensional (3D) Unet was established (named as AUnet) and its efficiency was improved through end-to-end training. Organ size was introduced as a priori knowledge to improve the performance of the model in convolution kernel size design, which enabled the network to better extract the features of different organs of different sizes. The adaptive histogram equalization algorithm was used to preprocess the input CT images to facilitate contour recognition. The similarity evaluation indexes, including Dice Similarity Coefficient (DSC) and Hausdorff Distance (HD), were calculated to verify the validity of segmentation. RESULTS: DSC and HD of the test dataset were 0.86±0.02 and 4.0±2.0 mm, respectively. No significant difference was found between the results of AUnet and manual segmentation of the OARs ( CONCLUSIONS AUnet, an improved deep learning neural network, is capable of automatic segmentation of the OARs in radiotherapy for NPC based on CT images, and for most organs, the results are comparable to those of manual segmentation.
Databases, Factual ; Humans ; Image Processing, Computer-Assisted ; Nasopharyngeal Carcinoma/radiotherapy* ; Nasopharyngeal Neoplasms/radiotherapy* ; Organs at Risk ; Tomography, X-Ray Computed

Objective:The aim of the study is to investigate the inter-observer and intra-observer precision in manually segmentation of organs-at-risk(OARs) for cervical cancer on the basis of MR image, and to preliminarily explore sequence selection designed for radiotherapy planning.Methods:Thirty cervical cancer patients scanned by MR-sim from 2016 to 2018 in the department of radiotherapy of Sun Yat-sen university cancer center were retrospectively analyzed. T1WI, T1dixonc and T2WI sequence from MR-simulator were selected and imported into Monaco planning system. Manual segmentation of 5 organs-at-risk (bladder, rectum, anal canal, and left/ right femoral head) was done by 2 independent experienced physicians on three sequences acquired from these patients. A month later, the second segmentation of the OARs in the T1WI sequence was done by one of the two physicians. Dice similarity coefficient (DSC), Hausdorff distance (HD) and position difference(Δ x, Δ y, Δ z) of each OAR were used to analyze the robustness of inter-observer and intra-observer segmentation OARs. Results:The HD values of five OARs segmentation by the two physicians in T1WI, T1dixonc and T2WI sequences and the same physician in T1WI at different time were all less than 2 mm; the position differences were less than 5 mm. The DSC values, HD values and difference position values of the two physician and the same physician at different time was positively correlated with the volume of OARs ( R=0.178-0.582, P<0.05). Due to the small volume of the anal canal (7.385±1.555) cm 3, the DSC values were less than 0.7 and the performance was slightly worse. The average DSC values of other OARs were all greater than 0.82. Comparing the DSC, HD and position differences of OARs in the three sequences, DSC values of rectum, left / right femoral head and bladder, HD values of left/right femoral head and rectum, and Δ z axis difference of anal canal and right femoral head of T1WI sequence were better than T1dixonc ( t=-3.116-3.604, P<0.05); DSC value of rectum and HD value of anal canal in T1WI sequence were better than T2WI sequence( t= 2.934, 3.677, P<0.05 ); T1dixonc sequence rectal DSC, right femoral head Δ z axis difference were slightly better than T2WI( t=6.806, 2.130, P<0.05 ). T2WI sequence bone tissue (left/right femoral head) stability was better than T1WI, T1dixonc, and the difference was statistically significant ( t=-6.580-6.542, P<0.05). Conclusions:From three index of DSC, HD and position difference, the robustness of inter-observer and intra-observer segmentation of bladder, rectum and femoral head are fine based on MR sequence, followed by anal canal. In addition, the robustness of OARs segmentation by T1WI sequence is better than that of T1dixonc and T2WI sequence.