Search

Abstract Background Iatrogenic mandibular nerve damage resulting from oral surgeries and dental procedures is painful and a formidable challenge for patients and oral surgeons alike, mainly because the absence of objective and quantitative methods for diagnosing nerve damage renders treatment and compensation ambiguous while often leading to medico-legal disputes. The aim of this study was to examine discriminating factors of traumatic mandibular nerve within a specific magnetic resonance imaging (MRI) protocol and to suggest tangible diagnostic criteria for peripheral trigeminal nerve injury. Methods Twenty-six patients with ipsilateral mandibular nerve trauma underwent T2 Flex water, 3D short tau inversion recovery (STIR), and diffusion-weighted imaging (DWI) acquired by periodically rotating overlapping parallel lines with enhanced reconstruction (PROPELLER) pulse sequences; 26 injured nerves were thus compared with contra-lateral healthy nerves at anatomically corresponding sites. T2 Flex apparent signal to noise ratio (FSNR), T2 Flex apparent nerve-muscle contrast to noise ratio (FNMCNR) 3D STIR apparent signal to noise ratio (SSNR), 3D STIR apparent nerve-muscle contrast to noise ratio (SNMCNR), apparent diffusion coefficient (ADC) and area of cross-sectional nerve (Area) were evaluated. Results Mixed model analysis revealed FSNR and FNMCNR to be the dual discriminators for traumatized mandibular nerve (p

Abstract In breast cancer radiation therapy, minimizing radiation-related risks and toxicity is vital for improving life expectancy. Tailoring radiotherapy techniques and treatment positions can reduce radiation doses to normal organs and mitigate treatment-related toxicity. This study entailed a dosimetric comparison of six different external beam whole-breast irradiation techniques in both supine and prone positions. We selected fourteen breast cancer patients, generating six treatment plans in both positions per patient. We assessed target coverage and organs at risk (OAR) doses to evaluate the impact of treatment techniques and positions. Excess absolute risk was calculated to estimate potential secondary cancer risk in the contralateral breast, ipsilateral lung, and contralateral lung. Additionally, we analyzed the distance between the target volume and OARs (heart and ipsilateral lung) while considering the treatment position. The results indicate that prone positioning lowers lung exposure in X-ray radiotherapy. However, particle beam therapies (PBTs) significantly reduce the dose to the heart and ipsilateral lung regardless of the patient’s position. Notably, negligible differences were observed between arc-delivery and static-delivery PBTs in terms of target conformity and OAR sparing. This study provides critical dosimetric evidence to facilitate informed decision-making regarding treatment techniques and positions.

Abstract The shielding parameters can vary depending on the geometrical structure of the linear accelerators (LINAC), treatment techniques, and beam energies. Recently, the introduction of O-ring type linear accelerators is increasing. The objective of this study is to evaluate the shielding parameters of new type of linac using a dedicated program developed by us named ORSE (O-ring type Radiation therapy equipment Shielding Evaluation). The shielding evaluation was conducted for a total of four treatment rooms including Elekta Unity, Varian Halcyon, and Accuray Tomotherapy. The developed program possesses the capability to calculate transmitted dose, maximum treatable patient capacity, and shielding wall thickness based on patient data. The doses were measured for five days using glass dosimeters to compare with the results of program. The IMRT factors and use factors obtained from patient data showed differences of up to 65.0% and 33.8%, respectively, compared to safety management report. The shielding evaluation conducted in each treatment room showed that the transmitted dose at every location was below 1% of the dose limit. The results of program and measurements showed a maximum difference of 0.003 mSv/week in transmitted dose. The ORSE program allows for the shielding evaluation results to the clinical environment of each institution based on patient data.

Abstract This work aims to investigate the clinical feasibility of deep learning-based synthetic CT images for cervix cancer, comparing them to MR for calculating attenuation (MRCAT). Patient cohort with 50 pairs of T2-weighted MR and CT images from cervical cancer patients was split into 40 for training and 10 for testing phases. We conducted deformable image registration and Nyul intensity normalization for MR images to maximize the similarity between MR and CT images as a preprocessing step. The processed images were plugged into a deep learning model, generative adversarial network. To prove clinical feasibility, we assessed the accuracy of synthetic CT images in image similarity using structural similarity (SSIM) and mean-absolute-error (MAE) and dosimetry similarity using gamma passing rate (GPR). Dose calculation was performed on the true and synthetic CT images with a commercial Monte Carlo algorithm. Synthetic CT images generated by deep learning outperformed MRCAT images in image similarity by 1.5% in SSIM, and 18.5 HU in MAE. In dosimetry, the DL-based synthetic CT images achieved 98.71% and 96.39% in the GPR at 1% and 1 mm criterion with 10% and 60% cut-off values of the prescription dose, which were 0.9% and 5.1% greater GPRs over MRCAT images.

Abstract We aimed to investigate the deliverability of dynamic conformal arc therapy (DCAT) by gantry wobble owing to the intrinsic inter-segment break of the Elekta linear accelerator (LINAC) and its adverse influence on the dose to the patient. The deliverability of DCAT was evaluated according to the plan parameters, which affect the gantry rotation speed and resultant positional inaccuracies; the deliverability according to the number of control points and dose rates was investigated by using treatment machine log files and dosimetry devices, respectively. A non-negligible degradation in DCAT deliverability due to gantry wobble was observed in both the treatment machine log files and dosimetry devices. The resulting dose-delivery error occurred below a certain number of control points or above a certain dose rate. Dose simulations in the patient domain showed a similar impact on deteriorated deliverability. For targets located primarily in the isocenter, the dose differences were negligible, whereas for organs at risk located mainly off-isocenter, the dose differences were significant up to − 8.77%. To ensure safe and accurate radiotherapy, optimal plan parameters should be selected, and gantry angle-specific validations should be conducted before treatment.

Current elevator landing door safety retainers are able to prevent fall accidents from an inadvertent bump against the door, as they are designed to endure a kinetic energy of 450 J. However, the recent increase in fall accidents among people in electric wheelchairs indicates that retainers are required to sustain stronger external shocks to prevent fall accidents against elevator landing doors. The purpose of this study is to develop elevator landing door safety retainers capable of enduring 1,000 J to prevent fall accidents of electric wheelchairs at the landing doors of elevators. In this study, crash analysis simulations were conducted after designing new scenarios that are not stated in the current regulations, considering the collision behavior of an electric wheelchair against an elevator landing door. The side hook was confirmed to play an important role in withstanding a collisional energy of 1,000 J. Crash analysis yielded different results according to the shape of the guide shoe. The analysis simulations were verified by performing experiments on various established scenarios. This study also confirmed the reliability of the developed safety retainers by implementing more realistic experiments on the crash of electric wheelchairs instead of the impactor proposed in the current regulations.

Due to anatomical changes between pre-planning and implantation, there exists a need for tools that can streamline the adjustment of needle and seed configurations in low dose rate brachytherapy for prostate cancer. Specifically, upon taking a second ultrasound on the day of treatment, the distribution of seeds and needles will differ drastically from the original plan. Clinics that employ this method must then spend time and resources to generate a workflow to manipulate the original configuration to the new configuration. ONCOSEED extracts data from VariSeed treatment plans, calculating a labor score (LScore) to optimize adjustments to needle configurations. A case study of three simulated VariSeed treatment plans was used to compare the ONCOSEED software to the manual method of generating a workflow. In the same method that was used at the authors’ clinic, several assistants annotated by hand the original plan to convert it to the new plan. The time taken to do so was recorded and compared to the runtime of the software when generating a workflow for the same plan. Results showed that ONCOSEED was on average 28 times faster than generating a workflow by hand. ONCOSEED enhances the efficiency of seed replacement in LDR brachytherapy, promoting the adoption of adaptive brachytherapy practices.

Purpose: Herein, we developed a deep learning algorithm to improve the segmentation of the clinical target volume (CTV) on daily cone beam computed tomography (CBCT) scans in breast cancer radiation therapy. By leveraging the Intentional Deep Overfit Learning (IDOL) framework, we aimed to enhance personalized image-guided radiation therapy based on patient-specific learning. Methods and Materials: We used 240 CBCT scans from 100 breast cancer patients and employed a 2-stage training approach. The first stage involved training a novel general deep learning model (Swin UNETR, UNET, and SegResNET) on 90 patients. The second stage used intentional overfitting on the remaining 10 patients for patient-specific CBCT outputs. Quantitative evaluation was conducted using the Dice Similarity Coefficient (DSC), Hausdorff Distance (HD), mean surface distance (MSD), and independent samples t test with expert contours on CBCT scans from the first to 15th fractions. Results: IDOL integration significantly improved CTV segmentation, particularly with the Swin UNETR model (P values < .05). Using patient-specific data, IDOL enhanced the DSC, HD, and MSD metrics. The average DSC for the 15th fraction improved from 0.9611 to 0.9819, the average HD decreased from 4.0118 mm to 1.3935 mm, and the average MSD decreased from 0.8723 to 0.4603. Incorporating CBCT scans from the initial treatments and first to third fractions further improved results, with an average DSC of 0.9850, an average HD of 1.2707 mm, and an average MSD of 0.4076 for the 15th fraction, closely aligning with physician-drawn contours. Conclusion: Compared with a general model, our patient-specific deep learning-based training algorithm significantly improved CTV segmentation accuracy of CBCT scans in patients with breast cancer. This approach, coupled with continuous deep learning training using daily CBCT scans, demonstrated enhanced CTV delineation accuracy and efficiency. Future studies should explore the adaptability of the IDOL framework to diverse deep learning models, data sets, and cancer sites.

Abstract The purpose of our study is to establish an efficient quality assurance (QA) procedure using a transmission-type detector (IBA, Stealth chamber), a reference signal detector, as a field chamber. Relative dosimetry items, including monitor unit linearity, output constancy based on dose rate and field size, and output factor were measured and compared with results obtained from the Farmer-type chamber (IBA, Wellhofer, FC65-G). Moreover, output for each field size was measured to assess its applicability to small fields. Results using the Stealth chamber were in good agreement with the FC65-G within 1.0%, except for output constancy according to gantry angle, which had a 1.1% error rate for the Stealth chamber and 2.7% for the FC65-G. Differences of up to − 6.26% output factor were observed for the Stealth chamber and up to − 0.56% for the CC-13 ionization chamber (IBA) in the 3 × 3 cm2 field. Our study confirmed the possibility of using Stealth chambers for relative dosimetry measurement in QA.

Purpose: We aimed to develop Lyman–Kutcher–Burman (LKB) and multivariable normal tissue complication probability (NTCP) models to predict the risk of radiation-induced hypothyroidism (RIHT) in breast cancer patients. Materials and methods: A total of 1,063 breast cancer patients who underwent whole breast irradiation between 2009 and 2016 were analyzed. Individual dose-volume histograms were used to generate LKB and multivariable logistic regression models. LKB model was fit using the thyroid radiation dose-volume parameters. A multivariable model was constructed to identify potential dosimetric and clinical parameters associated with RIHT. Internal validation was conducted using bootstrapping techniques, and model performance was evaluated using the area under the curve (AUC) and Hosmer–Lemeshow (HL) goodness-of-fit test. Results: RIHT developed in 4 % of patients with a median follow-up of 77.7 months. LKB and multivariable NTCP models exhibited significant agreement between the predicted and observed results (HL P values > 0.05). The multivariable NTCP model outperformed the LKB model in predicting RIHT (AUC 0.62 vs. 0.54). In the multivariable model, systemic therapy, age, and percentage of thyroid volume receiving ≥ 10 Gy (V10) were significant prognostic factors for RIHT. The cumulative incidence of RIHT was significantly higher in patients who exceeded the cut-off values for all three risk predictors (systemic therapy, age ≥ 40 years, and thyroid V10 ≥ 26 %, P

Purpose: To quantify interobserver variation (IOV) in target volume and organs-at-risk (OAR) contouring across 31 institutions in breast cancer cases and to explore the clinical utility of deep learning (DL)-based auto-contouring in reducing potential IOV. Methods and materials: In phase 1, two breast cancer cases were randomly selected and distributed to multiple institutions for contouring six clinical target volumes (CTVs) and eight OAR. In Phase 2, auto-contour sets were generated using a previously published DL Breast segmentation model and were made available for all participants. The difference in IOV of submitted contours in phases 1 and 2 was investigated quantitatively using the Dice similarity coefficient (DSC) and Hausdorff distance (HD). The qualitative analysis involved using contour heat maps to visualize the extent and location of these variations and the required modification. Results: Over 800 pairwise comparisons were analysed for each structure in each case. Quantitative phase 2 metrics showed significant improvement in the mean DSC (from 0.69 to 0.77) and HD (from 34.9 to 17.9 mm). Quantitative analysis showed increased interobserver agreement in phase 2, specifically for CTV structures (5–19 %), leading to fewer manual adjustments. Underlying IOV differences causes were reported using a questionnaire and hierarchical clustering analysis based on the volume of CTVs. Conclusion: DL-based auto-contours improved the contour agreement for OARs and CTVs significantly, both qualitatively and quantitatively, suggesting its potential role in minimizing radiation therapy protocol deviation.

In radiotherapy, when photon energy exceeding 8 MV is utilized, photoneutrons can activate the components within the gantry of the linear accelerator (linac). At the end of the linac's lifecycle, radiation workers are tasked with its dismantling and disposal, potentially exposing them to unintentional radiation. This study aims to identify and measure the radioisotopes generated by this activation through spectroscopy, and to evaluate the effective dose rate. We selected nine medical linacs, considering various factors such as manufacturer (Siemens, Varian, and Elekta), model, energy, period of operation, and workload. We identified the radionuclides in the linac head by employing an in situ high-purity germanium (HPGe) detector. Spectroscopy and dose-rate measurements were conducted post-shutdown. We also measured the dose rates at the beam-exit window following irradiation with 10 MV and 15 MV photon beams. As a result of the spectroscopy, we identified approximately 20 nuclides including those with half-lives of 100 days or longer, such as 54Mn, 60Co, 65Zn, 122Sb, and 198Au. The dose rate measurements after 10 MV irradiation decreased to the background level in 10 min. By contrast, on 15 MV irradiation, the dose rate was 628 nSv/h after 10 min and decreased to 268 nSv/h after 1.5 hours. It was confirmed that the difference in the level of radiation and the type of nuclide depends on the period of use, energy, and workload. However, the type of nuclide does not differ significantly between the linacs. It is necessary to propose appropriate guidelines for the safety of workers, and disposal/move-install should be planned while taking into consideration the equipment's energy usage rate.

Abstract This study aims to evaluate the specific characteristics of various multileaf collimator (MLC) position errors that are correlated with the indices using dose distribution. The dose distribution was investigated using the gamma, structural similarity, and dosiomics indices. Cases from the American Association of Physicists in Medicine Task Group 119 were planned, and systematic and random MLC position errors were simulated. The indices were obtained from distribution maps and statistically significant indices were selected. The final model was determined when all values of the area under the curve, accuracy, precision, sensitivity, and specificity were higher than 0.8 (p 0.9. Furthermore, the results of the DVH were related to dosiomics analysis in that it reflects the characteristics of the MLC position error. It was also shown that dosiomics analysis could provide important information on localized dose-distribution differences in addition to DVH information.

The planning accuracy of charged particle therapy (CPT) is subject to the accuracy of stopping power (SP) estimation. In this study, we propose a method of deriving a pseudo-triple-energy CT (pTECT) that can be achievable in the existing dual-energy CT (DECT) systems for better SP estimation. In order to remove the direct effect of errors in CT values, relative CT values according to three scanning voltage settings were used. CT values of each tissue substitute phantom were measured to show the non-linearity of the values thereby suggesting the absolute difference and ratio of CT values as parameters for SP estimation. Electron density, effective atomic number (EAN), mean excitation energy and SP were calculated based on these parameters. Two of conventional methods were implemented and compared to the proposed pTECT method in terms of residuals, absolute error and root-mean-square-error (RMSE). The proposed method outperformed the comparison methods in every evaluation metrics. Especially, the estimation error for EAN and mean excitation using pTECT were converging to zero. In this proof-of-concept study, we showed the feasibility of using three CT values for accurate SP estimation. Our suggested pTECT method indicates potential clinical utility of spectral CT imaging for CPT planning.

This paper proposes a spatiotemporal architecture with a deep neural network (DNN) for road surface conditions and types classification using LiDAR for autonomous emergency braking (AEB) systems. It is known that LiDAR provides information on the reflectivity and number of point clouds depending on a road surface. Thus, this paper utilizes the characteristics of LiDAR and the vehicle’s speed obtained from in-vehicle sensors for classifying front road surface conditions for the AEB system application. First, we divided the front road area into four subregions to use spatial information from the vehicle’s speed. Second, we constructed feature vectors using each subregion’s reflectivity, number of point clouds, and in-vehicle information. Third, the DNN classifies road surface conditions and types for each subregion. Finally, the output of the DNN feeds into the spatiotemporal process to make the final classification reflecting vehicle speed and probability given by the outcomes of softmax functions of the DNN output layer. To validate the effectiveness of the proposed method, we performed a comparative study with five other algorithms. With the proposed DNN, we obtained the highest accuracy of 98.0% and 98.6% for two subregions near the vehicle. In addition, we implemented the proposed method on the NVIDIA Jetson TX2 board to confirm that it is applicable in real-time.

Background: A relationship between the axillary-lateral thoracic vessel juncture (ALTJ) dose and lymphedema rate has been reported in patients with breast cancer. The purpose of this study was to validate this relationship and explore whether incorporation of the ALTJ dose-distribution parameters improves the prediction model’s accuracy. Methods: A total of 1,449 women with breast cancer who were treated with multimodal therapies from two institutions were analyzed. We categorized regional nodal irradiation (RNI) as limited RNI, which excluded level I/II, vs extensive RNI, which included level I/II. The ALTJ was delineated retrospectively, and dosimetric and clinical parameters were analyzed to determine the accuracy of predicting the development of lymphedema. Decision tree and random forest algorithms were used to construct the prediction models of the obtained dataset. We used Harrell’s C-index to assess discrimination. Results: The median follow-up time was 77.3 months, and the 5-year lymphedema rate was 6.8 %. According to the decision tree analysis, the lowest lymphedema rate (5-year, 1.2 %) was observed in patients with ≤ six removed lymph nodes and ≤ 66 % ALTJ V35Gy. The highest lymphedema rate was observed in patients with > 15 removed lymph nodes and an ALTJ maximum dose (Dmax) of > 53 Gy (5-year, 71.4 %). Patients with > 15 removed lymph nodes and an ALTJ Dmax ≤ 53 Gy had the second highest rate (5-year, 21.5 %). All other patients had relatively minor differences, with a rate of 9.5 % at 5 years. Random forest analysis revealed that the model’s C-index increased from 0.84 to 0.90 if dosimetric parameters were included instead of RNI (P

Abstract This work attempted to construct a new metal artifact reduction (MAR) framework in kilo-voltage (kV) computed tomography (CT) images by combining (1) deep learning and (2) multi-modal imaging, defined as MARTIAN (Metal Artifact Reduction throughout Two-step sequentIAl deep convolutional neural Networks). Most CNNs under supervised learning require artifact-free images to artifact-contaminated images for artifact correction. Mega-voltage (MV) CT is insensitive to metal artifacts, unlike kV CT due to different physical characteristics, which can facilitate the generation of artifact-free synthetic kV CT images throughout the first network (Network 1). The pairs of true kV CT and artifact-free kV CT images after post-processing constructed a subsequent network (Network 2) to conduct the actual MAR process. The proposed framework was implemented by GAN from 90 scans for head-and-neck and brain radiotherapy and validated with 10 independent cases against commercial MAR software. The artifact-free kV CT images following Network 1 and post-processing led to structural similarity (SSIM) of 0.997, and mean-absolute-error (MAE) of 10.2 HU, relative to true kV CT. Network 2 in charge of actual MAR successfully suppressed metal artifacts, relative to commercial MAR, while retaining the detailed imaging information, yielding the SSIM of 0.995 against 0.997 from the commercial MAR.

Abstract Background Based on a previous study on the feasibility of TomoEQA, an exit detector-based patient-specific pre-treatment quality assurance (QA) method for helical tomotherapy, an in-depth clinical evaluation was conducted. Methods Data of one hundred patients were analyzed to evaluate the clinical usefulness of TomoEQA for patient-specific pre-treatment QA in comparison with the conventional phantom-based method. Additional investigations were also performed under unusual measurement conditions to validate the off-axis region. In addition to the clinical evaluation of TomoEQA, a statistical analysis was conducted to determine the plan parameters that affect the pass/failure results of pre-treatment QA. Results The average and standard deviations of the gamma passing rate and point dose error for TomoEQA were comparable to those of the conventional QA method. For TomoEQA, the average values of the gamma passing rate and point dose error were 96.32% (standard deviation (1 sigma) = 3.94; 95% confidence interval (CI), 95.55 to 97.09) and − 1.12% (standard deviation (1 sigma) = 1.04; CI, − 1.32 to − 0.92), respectively. For the conventional QA method, the average values of the gamma passing rate and point dose error were 95.95% (standard deviation (1 sigma) = 4.35; 95% confidence interval (CI), 95.10 to 96.80) and − 1.20% (standard deviation (1 sigma) = 1.61; CI, − 1.52 to − 0.88), respectively. Further experiments on the off-axis region demonstrated that TomoEQA can provide accurate results for 3D dose analysis, which is inherently difficult in the conventional QA method. Through a statistical analysis based on the results of TomoEQA, it was validated that the total fraction (Total Fx), monitor units, beam-on-time, leaf-of-time below 100 ms, and planning target volume diameter were statistically significant for the pass/failure of the pre-treatment QA results. Conclusions TomoEQA is a clinically beneficial alternative to the conventional phantom-based QA method.

BackgroundOligometastatic disease (OMD) represents an indolent cancer status characterized by slow tumor growth and limited metastatic potential. The use of local therapy in the management of the condition continues to rise. This study aimed to investigate the advantage of pretreatment tumor growth rate in addition to baseline disease burden in characterizing OMDs, generally defined by the presence of ≤ 5 metastatic lesions.MethodsThe study included patients with metastatic melanoma treated with pembrolizumab. Gross tumor volume of all metastases was contoured on imaging before (TP-1) and at the initiation of pembrolizumab (TP0). Pretreatment tumor growth rate was calculated by an exponential ordinary differential equation model using the sum of tumor volumes at TP-1 and TP0 and the time interval between TP-1. and TP0. Patients were divided into interquartile groups based on pretreatment growth rate. Overall survival, progression-free survival, and subsequent progression-free survival were the study outcomes.ResultsAt baseline, median cumulative volume and number of metastases were 28.4 cc (range, 0.4-1194.8 cc) and 7 (range, 1-73), respectively. The median interval between TP-1 and TP0 was -90 days and pretreatment tumor growth rate (×10-2 days-1) was median 4.71 (range -0.62 to 44.1). The slow-paced group (pretreatment tumor growth rate ≤ 7.6 ×10-2 days-1, the upper quartile) had a significantly higher overall survival rate, progression-free survival, and subsequent progression-free survival compared to those of the fast-paced group (pretreatment tumor growth rate > 7.6 ×10-2 days-1). Notably, these differences were prominent in the subgroup with >5 metastases.ConclusionPretreatment tumor growth rate is a novel prognostic metric associated with overall survival, progression-free survival, and subsequent progression-free survival among metastatic melanoma patients, especially patients with >5 metastases. Future prospective studies should validate the advantage of disease growth rate plus disease burden in better defining OMDs.

The use of adjuvant Immune Checkpoint Inhibitors (ICI) after concurrent chemo-radiation therapy (CCRT) has become the standard of care for locally advanced non-small cell lung cancer (LA-NSCLC). However, prolonged radiotherapy regimens are known to cause radiation-induced lymphopenia (RIL), a long-neglected toxicity that has been shown to correlate with response to ICIs and survival of patients treated with adjuvant ICI after CCRT.In this study, we aim to develop a novel neural network (NN) approach that integrates patient characteristics, treatment related variables, and differential dose volume histograms (dDVH) of lung and heart to predict the incidence of RIL at the end of treatment. Multi-institutional data of 139 LA-NSCLC patients from two hospitals were collected for training and validation of our suggested model. Ensemble learning was combined with a bootstrap strategy to stabilize the model, which was evaluated internally using repeated cross validation.The performance of our proposed model was compared to conventional models using the same input features, such as Logistic Regression (LR) and Random Forests (RF), using the Area Under the Curve (AUC) of Receiver Operating Characteristics (ROC) curves. Our suggested model (AUC=0.77) outperformed the comparison models (AUC=0.72, 0.74) in terms of absolute performance, indicating that the convolutional structure of the network successfully abstracts additional information from the differential DVHs, which we studied using Gradient-weighted Class Activation Map.This study shows that clinical factors combined with dDVHs can be used to predict the risk of RIL for an individual patient and shows a path toward preventing lymphopenia using patient-specific modifications of the radiotherapy plan.

U-Net, based on a deep convolutional network (CNN), has been clinically used to auto-segment normal organs, while still being limited to the planning target volume (PTV) segmentation. This work aims to address the problems in two aspects: 1) apply one of the newest network architectures such as vision transformers other than the CNN-based networks, and 2) find an appropriate combination of network hyper-parameters with reference to recently proposed nnU-Net (“no-new-Net”). VT U-Net was adopted for auto-segmenting the whole pelvis prostate PTV as it consisted of fully transformer architecture. The upgraded version (v.2) applied the nnU-Net-like hyper-parameter optimizations, which did not fully cover the transformer-oriented hyper-parameters. Thus, we tried to find a suitable combination of two key hyper-parameters (patch size and embedded dimension) for 140 CT scans throughout 4-fold cross validation. The VT U-Net v.2 with hyper-parameter tuning yielded the highest dice similarity coefficient (DSC) of 82.5 and the lowest 95% Haussdorff distance (HD95) of 3.5 on average among the seven recently proposed deep learning networks. Importantly, the nnU-Net with hyper-parameter optimization achieved competitive performance, although this was based on the convolution layers. The network hyper-parameter tuning was demonstrated to be necessary even for the newly developed architecture of vision transformers.

Abstract Background Adjuvant radiation therapy improves the overall survival and loco-regional control in patients with breast cancer. However, radiation-induced heart disease, which occurs after treatment from incidental radiation exposure to the cardiac organ, is an emerging challenge. This study aimed to generate synthetic contrast-enhanced computed tomography (SCECT) from non-contrast CT (NCT) using deep learning (DL) and investigate its role in contouring cardiac substructures. We also aimed to determine its applicability for a retrospective study on the substructure volume-dose relationship for predicting radiation-induced heart disease. Methods We prepared NCT-CECT cardiac scan pairs of 59 patients. Of these, 35, 4, and 20 pairs were used for training, validation, and testing, respectively. We adopted conditional generative adversarial network as a framework to generate SCECT. SCECT was validated in the following three stages: (1) The similarity between SCECT and CECT was evaluated; (2) Manual contouring was performed on SCECT and CECT with sufficient intervals and based on this, the geometric similarity of cardiac substructures was measured between them; (3) The treatment plan was quantitatively analyzed based on the contours of SCECT and CECT. Results While the mean values (± standard deviation) of the mean absolute error, peak signal-to-noise ratio, and structural similarity index measure between SCECT and CECT were 20.66 ± 5.29, 21.57 ± 1.85, and 0.77 ± 0.06, those were 23.95 ± 6.98, 20.67 ± 2.34, and 0.76 ± 0.07 between NCT and CECT, respectively. The Dice similarity coefficients and mean surface distance between the contours of SCECT and CECT were 0.81 ± 0.06 and 2.44 ± 0.72, respectively. The dosimetry analysis displayed error rates of 0.13 ± 0.27 Gy and 0.71 ± 1.34% for the mean heart dose and V5Gy, respectively. Conclusion Our findings displayed the feasibility of SCECT generation from NCT and its potential for cardiac substructure delineation in patients who underwent breast radiation therapy.

The objective of this study is to compare the plan robustness at various beam angles. Hence, the influence of the beam angles on robustness and linear energy transfer (LET) was evaluated in gantry-based carbon-ion radiation therapy (CIRT) for prostate cancer. 10 patients with prostate cancer were considered, and a total dose of 51.6 Gy (Relative biological effectiveness (RBE) was prescribed for the target volume in 12 fractions. Five beam field plans comprising two opposed fields with different angle pairs were characterized. Further, dose parameters were extracted, and the RBE-weighted dose and LET values for all angle pairs were compared. All plans considering the setup uncertainty satisfied the dose regimen. When a parallel beam pair was used for perturbed scenarios to take into account set-up uncertainty in the anterior direction, the LET clinical treatment volume (CTV) D95% standard deviation was 1.5 times higher, and the standard deviation of RBE-weighted CTV D95% was 7.9 times higher compared to an oblique pair. The oblique beam fields were superior in terms of dose sparing for the rectum compared to the dose distribution using two conventional lateral opposed fields for prostate cancer.

Abstract Background There is an increasing interest in HER2-low breast cancer with promising data from clinical trials using novel anti-HER2 antibody–drug conjugates. We explored the differences in clinicopathological characteristics and survival outcomes between HER2-low and HER2-IHC 0 breast cancer. Methods Using nationwide data from the Korean Breast Cancer Registry between 2006 and 2011, 30,491 patients with stages I to III breast cancer were included in the analysis: 9,506 (31.2%) in the HER2-low group and 20,985 (68.8%) in the HER2-IHC 0 group. Kaplan–Meier and Cox proportional hazards regression survival analysis were used to compare breast cancer-specific survival between the two groups. Results HER2-low breast cancer was more frequent in patients with hormone receptor-positive breast cancer than in those with triple-negative breast cancer. In patients with hormone receptor-positive breast cancer, HER2-low breast cancer was associated with fewer T4 tumors, higher histological grade, and a negative lymphatic invasion. In patients with triple-negative breast cancer, HER2-low breast cancer was associated with a high lymph node ratio and positive lymphatic invasion. HER2-low breast cancer was significantly associated with a lower Ki-67 labeling index. No significant difference was observed in overall survival between the two groups. HER2-low breast cancer showed significantly better breast cancer-specific survival than HER2-IHC 0 breast cancer, regardless of the hormone receptor status. In multivariate analysis, the impact of low HER2 expression on breast cancer-specific survival was significant only in triple-negative breast cancer (HRs, 0.68; 95% CI, 0.49–0.93; P = 0.019). Conclusions These findings suggest that the biology and clinical impact of low HER2 expression can differ according to the hormone receptor status and support the need for further investigation on the understanding of the biology of HER2-low breast cancer.

The rapid rise in the application of novel treatment techniques, such as intensity-modulated radiotherapy (IMRT), motivated us to survey the status of Korea's radiation safety management and the shielding designs of facilities employing medical linear accelerators (LINACs). To this end, a questionnaire was used to collect information on LINAC facilities and treatments, workload, shielding design, shielding management, and path of obtaining shielding information. Out of 100 domestic institutions, 52 responded to the survey. Approximately 70% of the institutions utilized IMRT for more than 60% of their cases, and an IMRT factor of 5 was adopted by 75% of these institutions. Over 80% of the institutions accounted for the applied time-averaged dose rate per week and instantaneous dose equivalent rates in their shielding designs. Approximately 45% of the institutions obtained important shielding information via a radiation shielding design company and the NCRP-151 report. Overall, most facilities were shown to follow the standards recommended by the relevant international agencies. However, the requirement to establish standardized shielding design information and clarify ambiguous paths for information acquisition was also highlighted. Therefore, the study's results can be used as a foundation for establishing a safety control system and for creating adequate shielding designs.

Abstract Predicting the radiation dose‒toxicity relationship is important for local tumor control and patients’ quality of life. We developed a first intuitive evaluation system that directly matches the three-dimensional (3D) dose distribution with the skin surface image of patients with radiation dermatitis (RD) to predict RD in patients undergoing radiotherapy. Using an RGB-D camera, 82 3D skin surface images (3DSSIs) were acquired from 19 patients who underwent radiotherapy. 3DSSI data acquired included 3D skin surface shape and optical imaging of the area where RD occurs. Surface registration between 3D skin dose (3DSD) and 3DSSI is performed using the iterative closest point algorithm, then reconstructed as a two-dimensional color image. The developed system successfully matched 3DSSI and 3DSD, and visualized the planned dose distribution onto the patient's RD image. The dose distribution pattern was consistent with the occurrence pattern of RD. This new approach facilitated the evaluation of the direct correlation between skin-dose distribution and RD and, therefore, provides a potential to predict the probability of RD and thereby decrease RD severity by enabling informed treatment decision making by physicians. However, the results need to be interpreted with caution due to the small sample size.

A 3D-printed applicator for intraoperative radiation therapy (IORT) using the INTRABEAM system has been introduced, designed to protect normal tissue, expand indications, and assess its effectiveness in various intraoperative settings. The study involved designing and creating inclined applicators at 0° and 45° angles, which were capable of producing flattened and unflattened beams and shielding non-target areas from radiation. The applicator was evaluated based on dose uniformity, percentage depth dose, leakage, and dose rate. The findings showed that the applicator provided a uniform beam at various depths (1.23 to 1.30 for the flattened beam and 1.25 to 1.32 for the unflattened beam) and effectively shielded against leakage, with no leakage dose measured at 1 mm from the surface. Additionally, the dose rates after 3 min of irradiation for flattened and unflattened beams were 51.0 and 52.9 cGy for 0° and 48.5 and 54.8 cGy for 45° applicators, respectively. This suggests that the applicator can efficiently deliver IORT while minimizing exposure to normal tissues, particularly in treating chest and abdominal lesions near or on critical organs. The applicator has undergone dose calibration and has moved forward to clinical application.

Purpose We aimed to investigate the benefit of a screening tool in selecting head-injured children who can potentially cooperate with computed tomography (CT). Methods The study population consisted of head-injured children aged 36 months or younger who visited the emergency department (ED) and underwent CT from January 2013 through December 2020. Procedural sedation and analgesia (PSA) using per os chloral hydrate or per rectal thiopental was implemented to children presumed less cooperative for CT as per a clinical screening tool for cooperative candidates for CT (e.g., ability to lie still on bed for 10 seconds without a guardian). According to the PSA and the first attempt success of CT, we compared baseline characteristics, CT findings, clinically important traumatic brain injury, ED length of stay (EDLOS), and ED disposition. Results Among the 247 children, PSA was used in 102 (41.3%). The PSA group showed a higher proportion of 3-36 months of age (PSA, 96.1% vs. non-PSA, 82.8%; P = 0.001) and longer median EDLOS (PSA, 127.0 [interquartile range, 101.0-172.0] vs. non-PSA, 85.0 minutes [63.0-130.0]; P < 0.001). The number of children undergoing the first attempt success was 213 (86.2%) without a difference in both groups (PSA, 84.3% vs. non-PSA, 87.6%; P = 0.645). In the 213 children, the implementation of PSA was associated with longer median EDLOS (PSA, 121.0 [99.0-156.0] vs. non-PSA, 77.0 minutes [60.0-122.0]; P < 0.001). In the non-PSA group (n = 145), the first attempt failure was associated with the presence of skull fracture (success, 7.1% vs. failure, 27.8%; P = 0.009). Conclusion This study suggests a potential benefit of the screening tool in selecting cooperative candidates for CT, i.e., those who could forgo PSA using oral chloral hydrate or per rectal thiopental, in young children with head injury.

Abstract Purpose To study the performance of a proposed deep learning-based autocontouring system in delineating organs at risk (OARs) in breast radiotherapy with a group of experts. Methods Eleven experts from two institutions delineated nine OARs in 10 cases of adjuvant radiotherapy after breast-conserving surgery. Autocontours were then provided to the experts for correction. Overall, 110 manual contours, 110 corrected autocontours, and 10 autocontours of each type of OAR were analyzed. The Dice similarity coefficient (DSC) and Hausdorff distance (HD) were used to compare the degree of agreement between the best manual contour (chosen by an independent expert committee) and each autocontour, corrected autocontour, and manual contour. Higher DSCs and lower HDs indicated a better geometric overlap. The amount of time reduction using the autocontouring system was examined. User satisfaction was evaluated using a survey. Results Manual contours, corrected autocontours, and autocontours had a similar accuracy in the average DSC value (0.88 vs. 0.90 vs. 0.90). The accuracy of autocontours ranked the second place, based on DSCs, and the first place, based on HDs among the manual contours. Interphysician variations among the experts were reduced in corrected autocontours, compared to variations in manual contours (DSC: 0.89–0.90 vs. 0.87–0.90; HD: 4.3–5.8 mm vs. 5.3–7.6 mm). Among the manual delineations, the breast contours had the largest variations, which improved most significantly with the autocontouring system. The total mean times for nine OARs were 37 min for manual contours and 6 min for corrected autocontours. The results of the survey revealed good user satisfaction. Conclusions The autocontouring system had a similar performance in OARs as that of the experts’ manual contouring. This system can be valuable in improving the quality of breast radiotherapy and reducing interphysician variability in clinical practice.

Abstract Background The COVID-19 pandemic has stripped many medical students worldwide of their right to quality education. In response, we developed hybrid courses involving aspects of both online and in-person teaching for radiation oncology medical student clerkship. Methods We entitled students to customize their own rotation schedule using Google Forms and developed a flipped learning online class, which consisted of at least one video clip on basic knowledge of radiation oncology per day (yonsei-radonc.com). Students were instructed to watch online videos before the next day’s discussion session. Required components of the medical education program (e.g., target drawing, site visits to treatment facilities) were also prepared and conducted in accordance with the appropriate level of social distancing measures. Finally, we conducted questionnaire surveys after the completion of the week-long course and clerkship. Results From March to June 2020, 110 fourth-year medical students undertook a clinical module in our 1-week radiation oncology program course. Each day, students completed the flipped learning prior to meeting with the educator and then participated in the online discussion session and conference. All activities were well performed as scheduled. Students’ motivation was high, as was their overall satisfaction with the course. The students were satisfied with the online contents, flipped learning strategy, and instructors. Conclusions We successfully integrated open and virtual educational platforms to improve access to and satisfaction with student clerkship. In the future “new normal,” minimized face-to-face learning interactions, such as flipped learning, should be actively utilized for medical and other students’ education.

Purpose: We aimed to investigate the dose calculation accuracy of Mobius3D for small-field flattening-filter-free x-rays, mainly utilized for stereotactic body radiation therapy (SBRT). The accuracy of beam modeling and multileaf collimator (MLC) modeling in Mobius3D, significantly affecting the dose calculation is investigated. Methods: The commissioning procedures of Mobius3D were performed for unflattened 6 MV and 10 MV x-ray beams of the linear accelerator, including beam model adjustment and dosimetric leaf gap (DLG) optimization. An experimental study with artificial plans was conducted to evaluate the accuracy of small-field modeling. The dose calculation accuracy of Mobius3D was also evaluated for retrospective SBRT plans with multiple targets. Results: Both studies evaluated the dose calculation accuracy through comparisons with the measured data. Relatively large differences were observed for off-axis distances over 5 cm and for small fields less than 1 cm field size. For the study with artificial plans, the maximum absolute error of 9.96% for unflattened 6 MV and 9.07% for unflattened 10 MV was observed when the field size was 1 cm. For the study with patient plans, the mean gamma passing rate with 3%/3 mm gamma criterion was 63.6% for unflattened 6 MV and 82.6% for unflattened 10 MV. The maximum of the average dose difference was −19.9% for unflattened 6MV and −10.1% for unflattened 10MV. Conclusions: The dose calculation accuracy uncertainties of Mobius3D for small-field flattening-filter-free photon beams were observed. The study results indicated that the beam and MLC modeling of Mobius3D must be improved for use in SBRT pretreatment QA in clinical practice.

PurposeThe purpose of this study is to propose a technique for delivering accurate doses in an image-guided system by developing an experimental setup optimized for total lymphoid irradiation (TLI) in rat lung transplantation.Materials and MethodsIn this study, a position-controlled shielding system was developed, and the dose was quantitatively evaluated using a 3D rat phantom and Gafchromic EBT3 film. In addition, we made our own image-guided system that allows the position of the rat and the shielding system to be confirmed during TLI.ResultsAs a result of using the position-controlled shielding system, it was found that the doses to the head and lungs were reduced by 93.1 and 87.4%, respectively, of the prescribed doses. In addition, it was shown that the position of the shielding system can be easily confirmed by using the image guidance system.ConclusionA shielding apparatus that can control dose delivery according to the size of the rat can optimize the dose for TLI in rat lung transplantation.

Abstract Background In breast cancer patients receiving radiotherapy (RT), accurate target delineation and reduction of radiation doses to the nearby normal organs is important. However, manual clinical target volume (CTV) and organs-at-risk (OARs) segmentation for treatment planning increases physicians’ workload and inter-physician variability considerably. In this study, we evaluated the potential benefits of deep learning-based auto-segmented contours by comparing them to manually delineated contours for breast cancer patients. Methods CTVs for bilateral breasts, regional lymph nodes, and OARs (including the heart, lungs, esophagus, spinal cord, and thyroid) were manually delineated on planning computed tomography scans of 111 breast cancer patients who received breast-conserving surgery. Subsequently, a two-stage convolutional neural network algorithm was used. Quantitative metrics, including the Dice similarity coefficient (DSC) and 95% Hausdorff distance, and qualitative scoring by two panels from 10 institutions were used for analysis. Inter-observer variability and delineation time were assessed; furthermore, dose-volume histograms and dosimetric parameters were also analyzed using another set of patient data. Results The correlation between the auto-segmented and manual contours was acceptable for OARs, with a mean DSC higher than 0.80 for all OARs. In addition, the CTVs showed favorable results, with mean DSCs higher than 0.70 for all breast and regional lymph node CTVs. Furthermore, qualitative subjective scoring showed that the results were acceptable for all CTVs and OARs, with a median score of at least 8 (possible range: 0–10) for (1) the differences between manual and auto-segmented contours and (2) the extent to which auto-segmentation would assist physicians in clinical practice. The differences in dosimetric parameters between the auto-segmented and manual contours were minimal. Conclusions The feasibility of deep learning-based auto-segmentation in breast RT planning was demonstrated. Although deep learning-based auto-segmentation cannot be a substitute for radiation oncologists, it is a useful tool with excellent potential in assisting radiation oncologists in the future. Trial registration Retrospectively registered.

This paper proposes a classification algorithm utilizing an open set recognition concept to conservatively detect lane change intention of surrounding vehicles. Conservatively predicting the lane change intention of the surrounding vehicles is needed to improve adaptive cruise control (ACC) performance and avoid possible accidents. However, existing machine learning can make incorrect decisions due to information not included in the training data set or confused data even with probability. To cope with this problem, we present a classification algorithm using a multi-class support vector machine applying an open set recognition concept to detect the surrounding vehicles’ lane change intentions. Feature vectors are constructed from lateral information obtained by a Kalman filter using only radar and in-vehicle sensors. The open set recognition concept is adapted using Meta-Recognition based on binary classifiers scores. Furthermore, we analyze lateral information where an object vehicle changes lanes. From experimental results, we observe that the proposed system conservatively deals with wrong decisions and detects and cancels detecting the closest in-path vehicle (CIPV) earlier with average times of 1.4 sec and 0.4 sec compared with a commercial radar system, respectively.

Abstract Background Despite detailed instruction for full bladder, patients are unable to maintain consistent bladder filling during a 5-week pelvic radiation therapy (RT) course. We investigated the best bladder volume estimation procedure for verifying consistent bladder volume. Methods We reviewed 462 patients who underwent pelvic RT. Biofeedback using a bladder scanner was conducted before simulation and during treatment. Exact bladder volume was calculated by bladder inner wall contour based on CT images (Vctsim). Bladder volume was estimated either by bladder scanner (Vscan) or anatomical features from the presacral promontory to the bladder base and dome in the sagittal plane of CT (Vratio). The feasibility of Vratio was validated using daily megavoltage or kV cone-beam CT before treatment. Results Mean Vctsim was 335.6 ± 147.5 cc. Despite a positive correlation between Vctsim and Vscan (R 2 = 0.278) and between Vctsim and Vratio (R 2 = 0.424), Vratio yielded more consistent results than Vscan, with a mean percentage error of 26.3 (SD 19.6, p

Dyslipidemia and autophagy have been implicated in the pathogenesis of blinding neovascular age-related macular degeneration (NV-AMD). VLDL receptor (VLDLR), expressed in photoreceptors with a high metabolic rate, facilitates the uptake of triglyceride-derived fatty acids. Since fatty acid uptake is reduced in Vldlr–/– tissues, more remain in circulation, and the retina is fuel deficient, driving the formation in mice of neovascular lesions reminiscent of retinal angiomatous proliferation (RAP), a subtype of NV-AMD. Nutrient scarcity and energy failure are classically mitigated by increasing autophagy. We found that excess circulating lipids restrained retinal autophagy, which contributed to pathological angiogenesis in the Vldlr–/– RAP model. Triglyceride-derived fatty acid sensed by free fatty acid receptor 1 (FFAR1) restricted autophagy and oxidative metabolism in photoreceptors. FFAR1 suppressed transcription factor EB (TFEB), a master regulator of autophagy and lipid metabolism. Reduced TFEB, in turn, decreased sirtuin-3 expression and mitochondrial respiration. Metabolomic signatures of mouse RAP-like retinas were consistent with a role in promoting angiogenesis. This signature was also found in human NV-AMD vitreous. Restoring photoreceptor autophagy in Vldlr–/– retinas, either pharmacologically or by deleting Ffar1, enhanced metabolic efficiency and suppressed pathological angiogenesis. Dysregulated autophagy by circulating lipids might therefore contribute to the energy failure of photoreceptors driving neovascular eye diseases, and FFAR1 may be a target for intervention.

Purpose: Various deformable image registration (DIR) methods have been used to evaluate organ deformations in 4-dimensional computed tomography (4D CT) images scanned during the respiratory motions of a patient. This study assesses the performance of 10 DIR algorithms using 4D CT images of 5 patients with fiducial markers (FMs) implanted during the postoperative radiosurgery of multiple lung metastases. Methods: To evaluate DIR algorithms, 4D CT images of 5 patients were used, and ground-truths of FMs and tumors were generated by physicians based on their medical expertise. The positions of FMs and tumors in each 4D CT phase image were determined using 10 DIR algorithms, and the deformed results were compared with ground-truth data. Results: The target registration errors (TREs) between the FM positions estimated by optical flow algorithms and the ground-truth ranged from 1.82 ± 1.05 to 1.98 ± 1.17 mm, which is within the uncertainty of the ground-truth position. Two algorithm groups, namely, optical flow and demons, were used to estimate tumor positions with TREs ranging from 1.29 ± 1.21 to 1.78 ± 1.75 mm. With respect to the deformed position for tumors, for the 2 DIR algorithm groups, the maximum differences of the deformed positions for gross tumor volume tracking were approximately 4.55 to 7.55 times higher than the mean differences. Errors caused by the aforementioned difference in the Hounsfield unit values were also observed. Conclusions: We quantitatively evaluated 10 DIR algorithms using 4D CT images of 5 patients and compared the results with ground-truth data. The optical flow algorithms showed reasonable FM-tracking results in patient 4D CT images. The iterative optical flow method delivered the best performance in this study. With respect to the tumor volume, the optical flow and demons algorithms delivered the best performance.

Deep neural networks have shown great improvements in low-dose computed tomography (CT) denoising. Early algorithms were primarily optimized to obtain an accurate image with low distortion between the denoised image and reference full-dose image at the cost of yielding an overly smoothed unrealistic CT image. Recent research has sought to preserve the fine details of denoised images with high perceptual quality, which has been accompanied by a decrease in objective quality due to a trade-off between perceptual quality and distortion. We pursue a network that can generate accurate and realistic CT images with high objective and perceptual quality within one network, achieving a better perception-distortion trade-off. To achieve this goal, we propose a stationary wavelet transform-assisted network employing the characteristics of high- and low-frequency domains of the wavelet transform and frequency subband-specific losses defined in the wavelet domain. We first introduce a stationary wavelet transform for the network training procedure. Then, we train the network using objective loss functions defined for high- and low-frequency domains to enhance the objective quality of the denoised CT image. With this network design, we train the network again after replacing the objective loss functions with perceptual loss functions in high- and low-frequency domains. As a result, we acquired denoised CT images with high perceptual quality using this strategy while minimizing the objective quality loss. We evaluated our algorithms on the phantom and clinical images, and the quantitative and qualitative results indicate that ours outperform the existing state-of-the-art algorithms in terms of objective and perceptual quality.

Abstract Background Rotational beam delivery enables concurrent acquisition of cone-beam CT (CBCT), thereby facilitating further geometric verification of patient setup during radiation treatment. However, it is challenging to acquire CBCT during stereotactic body radiation therapy (SBRT) using flattening-filter free X-ray beams, in which a high radiation dose is delivered. This study presents quantitative evaluation results of the image quality in four-dimensional (4D) in-treatment CBCT acquired during SBRT delivery. Methods The impact of megavoltage (MV) scatter and acquisition parameters on the image quality was evaluated using Catphan 503 and XSight lung tracking phantoms. The in-treatment CBCT images of the phantoms were acquired while delivering 16 SBRT plans. The uniformity, contrast, and contrast-to-noise ratio (CNR) of the in-treatment CBCT images were calculated and compared to those of CBCT images acquired without SBRT delivery. Furthermore, the localizing accuracy of the moving target in the XSight lung phantom was evaluated for 10 respiratory phases. Results The CNR of the 3D-reconstucted Catphan CBCT images was reduced from 6.3 to 2.6 due to the effect of MV treatment scatter. Both for the Catphan and XSight phantoms, the CBCT image quality was affected by the tube current and monitor units (MUs) of the treatment plan. The lung target in the XSight tracking phantom was most visible for extreme phases; the mean CNRs of the lung target in the in-treatment CBCT images (with 40 mA tube current) across the SBRT plans were 3.2 for the end-of-exhalation phase and 3.0 for the end-of-inhalation phase. The lung target was localized with sub-millimeter accuracy for the extreme respiratory phases. Conclusions Full-arc acquisition with an increased tube current (e.g. 40 mA) is recommended to compensate for degradation in the CBCT image quality due to unflattened MV beam scatter. Acquiring in-treatment CBCT with a high-MU treatment beam is also suggested to improve the resulting CBCT image quality.

Abstract Background Since intensity-modulated radiation therapy (IMRT) has become popular for the treatment of gynecologic cancers, the contouring process has become more critical. This study evaluated the feasibility of atlas-based auto-segmentation (ABAS) for contouring in patients with endometrial and cervical cancers. Methods A total of 75 sets of planning CT images from 75 patients were collected. Contours for the pelvic nodal clinical target volume (CTV), femur, and bladder were carefully generated by two skilled radiation oncologists. Of 75 patients, 60 were randomly registered in three different atlas libraries for ABAS in groups of 20, 40, or 60. ABAS was conducted in 15 patients, followed by manual correction (ABASc). The time required to generate all contours was recorded, and the accuracy of segmentation was assessed using Dice’s coefficient (DC) and the Hausdorff distance (HD) and compared to those of manually delineated contours. Results For ABAS-CTV, the best results were achieved with groups of 60 patients (DC, 0.79; HD, 19.7 mm) and the worst results with groups of 20 patients (DC, 0.75; p = 0.012; HD, 21.3 mm; p = 0.002). ABASc-CTV performed better than ABAS-CTV in terms of both HD and DC (ABASc [n = 60]; DC, 0.84; HD, 15.6 mm; all p 0.9 and HD ≤10.0 mm), with significant time reduction compared to that needed for manual delineation (p 40 mm). Furthermore, ABASc-Bladder required a longer processing time than manual contouring to achieve the same accuracy. Conclusions ABAS could help physicians to delineate the CTV and organs-at-risk (e.g., femurs) in IMRT planning considering its consistency, efficacy, and accuracy.

Abstract Background Tomotherapy-based total body irradiation (TBI) is performed using the head-first position (HFP) and feet-first position (FFP) due to treatment length exceeding the 135 cm limit. To reduce the dosimetric variation at the match lines, we propose and verify a volumetric gradient matching technique (VGMT) by combining TomoHelical (TH) and TomoDirect (TD) modes. Methods Two planning CT image sets were acquired with HFP and FFP using 15 × 55 × 18 cm3 of solid water phantom. Planning target volume (PTV) was divided into upper, lower, and gradient volumes. The junction comprised 2-cm thick five and seven gradient volumes (5-GVs and 7-GVs) to create a dose distribution with a gentle slope. TH-IMRT and TD-IMRT plans were generated with 5-GVs and 7-GVs. The setup error in the calculated dose was assessed by shifting dose distribution of the FFP plan by 5, 10, 15, and 20 mm in the longitudinal direction and comparing it with the original. Doses for 95% (D95) and 5% of the PTV (D5) were calculated for all simulated setup error plans. Absolute dose measurements were performed using an ionization chamber in the junction. Results The TH&TD plan produced a linear gradient in junction volume, comparable to that of the TH&TH plan. D5 of the PTV was 110% of the prescribed dose when the FFP plan was shifted 0.7 cm and 1.2 cm in the superior direction for 5-GVs and 7-GVs. D95 of the PTV decreased to

PurposeTo develop an internal target volume (ITV) margin determination framework (or decision-supporting framework) for treating multiple lung metastases using CyberKnife Synchrony with intraoperatively implanted fiducial markers (IIFMs). The feasibility of using non-ideally implanted fiducial markers (a limited number and/or far from a target) for tracking-based lung stereotactic ablative radiotherapy (SABR) was investigated.MethodsIn the developed margin determination framework, an optimal set of IIFMs was determined to minimize a tracking uncertainty-specific ITV (ITVtracking) margin (margin required to cover target-to-marker motion discrepancy), i.e., minimize the motion discrepancies between gross tumor volume (GTV) and the selected set of fiducial markers (FMs). The developed margin determination framework was evaluated in 17 patients with lung metastases. To automatically calculate the respiratory motions of the FMs, a template matching-based FM tracking algorithm was developed, and GTV motion was manually measured. Furthermore, during-treatment motions of the selected FMs were analyzed using log files and compared with those calculated using 4D CTs.ResultsFor 41 of 42 lesions in 17 patients (97.6%), an optimal set of the IIFMs was successfully determined, requiring an ITVtracking margin less than 5 mm. The template matching-based FM tracking algorithm calculated the FM motions with a sub-millimeter accuracy compared with the manual measurements. The patient respiratory motions during treatment were, on average, significantly smaller than those measured at simulation for the patient cohort considered.ConclusionUse of the developed margin determination framework employing CyberKnife Synchrony with a limited number of IIFMs is feasible for lung SABR.

Purpose: This study was conducted to develop a deep learning-based automatic segmentation (DLBAS) model of head and neck organs for radiotherapy (RT) in dogs, and to evaluate the feasibility for delineating the RT planning.Materials and Methods: The segmentation indicated that there were potentially 15 organs at risk (OARs) in the head and neck of dogs. Post-contrast computed tomography (CT) was performed in 90 dogs. The training and validation sets comprised 80 CT data sets, including 20 test sets. The accuracy of the segmentation was assessed using both the Dice similarity coefficient (DSC) and the Hausdorff distance (HD), and by referencing the expert contours as the ground truth. An additional 10 clinical test sets with relatively large displacement or deformation of organs were selected for verification in cancer patients. To evaluate the applicability in cancer patients, and the impact of expert intervention, three methods–HA, DLBAS, and the readjustment of the predicted data obtained via the DLBAS of the clinical test sets (HA_DLBAS)–were compared.Results: The DLBAS model (in the 20 test sets) showed reliable DSC and HD values; it also had a short contouring time of ~3 s. The average (mean ± standard deviation) DSC (0.83 ± 0.04) and HD (2.71 ± 1.01 mm) values were similar to those of previous human studies. The DLBAS was highly accurate and had no large displacement of head and neck organs. However, the DLBAS in the 10 clinical test sets showed lower DSC (0.78 ± 0.11) and higher HD (4.30 ± 3.69 mm) values than those of the test sets. The HA_DLBAS was comparable to both the HA (DSC: 0.85 ± 0.06 and HD: 2.74 ± 1.18 mm) and DLBAS presented better comparison metrics and decreased statistical deviations (DSC: 0.94 ± 0.03 and HD: 2.30 ± 0.41 mm). In addition, the contouring time of HA_DLBAS (30 min) was less than that of HA (80 min).Conclusion: In conclusion, HA_DLBAS method and the proposed DLBAS was highly consistent and robust in its performance. Thus, DLBAS has great potential as a single or supportive tool to the key process in RT planning.

PurposeRespiratory training system that can be used by patients themselves was developed with a micro-electro-mechanical-system (MEMS)-based patch-type magnetic sensor. We conducted a basic function test and clinical usability evaluation to determine the system’s clinical applicability.MethodsThe system is designed with a sensor attached to the patient’s chest and a magnet on the back to monitor the patient’s respiration by measuring changes in magnetic intensity related to respiratory movements of the thoracic surface. The system comprises a MEMS-based patch-type magnetic sensor capable of wireless communication and being applied to measurement magnets and mobile applications. System performance was evaluated by the level of systemic noise, the precision of the sensor in various breathing patterns, how measurement signals change for varying distances, or the presence or absence of material between the sensor and the magnet. Various breathing patterns were created using the QUASAR respiratory motion phantom; the data obtained were analyzed using the fitting and peak value analysis methods.ResultsThe sensor had a noise ratio of

Systems based on remote sensing technology, which use reciprocal acoustic signals to continuously monitor changes in the coastal oceanic environment, are referred to as coastal acoustic tomography (CAT) systems. These systems have been applied in regions in which heavy ship traffics, fishing and marine aquaculture activities make it difficult to establish in situ oceanic sensor moorings. Conventionally, CAT measurements were used to successfully produce horizontal maps of the depth-averaged current velocity and temperature in these coastal regions without attempting to produce a vertical temperature profile. This prompted us to propose a new method for vertical temperature profile estimation (VTPE) from CAT data using the available sea surface temperature (SST), near-bottom temperature (NBT), and water depth. The VTPE method was validated using data-assimilated and tide-included high-resolution ocean model outputs, including tide data, by comparing the estimated and simulated temperatures. Measurements were performed in the southern coastal region of Korea, where two CAT stations were moored to establish a continuous coastal ocean monitoring system. The validation results revealed that the algorithm performed well across all seasons. Sensitivity tests of the VTPE method with reasonable realistic random errors in the SST, NBT, and acoustic travel time measurements demonstrate that the method is applicable to CAT observation data because the monthly mean root-mean-squared difference (RMSD) for the vertical profiles for February, May, August, and November were 0.23, 0.30, 0.50, and 0.24°C, respectively. The VTPE method was applied to the CAT observation datasets acquired in February and August. The transceivers at the CAT stations were at depths 11 and 6 m on average. The RMSD between the estimated and observed temperatures in the middle layer (∼3 m depth) between two stations in February and August were 0.08 and 0.60°C, respectively, the accuracy of which is sufficient in largely time-varying coastal environments. We provide a novel method for continuous coastal subsurface environmental monitoring without interrupting maritime traffic, fishing, and marine aquaculture activities.

Deep convolutional neural network (CNN) helped enhance image quality of cone-beam computed tomography (CBCT) by generating synthetic CT. Most of the previous works, however, trained network by intensity-based loss functions, possibly undermining to promote image feature similarity. The verifications were not sufficient to demonstrate clinical applicability, either. This work investigated the effect of variable loss functions combining feature- and intensity-driven losses in synthetic CT generation, followed by strengthening the verification of generated images in both image similarity and dosimetry accuracy. The proposed strategy highlighted the feature-driven quantification in (1) training the network by perceptual loss, besides L1 and structural similarity (SSIM) losses regarding anatomical similarity, and (2) evaluating image similarity by feature mapping ratio (FMR), besides conventional metrics. In addition, the synthetic CT images were assessed in terms of dose calculating accuracy by a commercial Monte-Carlo algorithm. The network was trained with 50 paired CBCT-CT scans acquired at the same CT simulator and treatment unit to constrain environmental factors any other than loss functions. For 10 independent cases, incorporating perceptual loss into L1 and SSIM losses outperformed the other combinations, which enhanced FMR of image similarity by 10%, and the dose calculating accuracy by 1–2% of gamma passing rate in 1%/1mm criterion.

Purpose: The aim of this study was to investigate the relationship between the anatomical location of thrombi in the lower extremities and the development of pulmonary embolism (PE). Materials and Methods : We collected and analyzed the data of patients diagnosed with deep vein thrombosis (DVT) of the lower extremities between 2006 and 2015, and included those whose computed tomography (CT) data were available for PE identification. We evaluated the relationship between the laterality and the proximal/distal location of the thrombi in lower extremites and the location of PE. Results : CT images were available for 388/452 patients with DVT. After excluding 32 cases with bilateral involvement, 356 cases were included for analysis in this study. The ratio of DVT in the left:right leg was 232:124. PEs developed in 121 (52.2%) patients with left-sided DVT and in 78 (62.9%) with right-sided DVT (P=0.052). PEs in the main pulmonary arteries developed in 36 (15.5%) patients with left leg DVT and in 30 (24.2%) with right leg DVT (P=0.045). The most frequent site of thrombosis associated with the development of PE was the left iliac vein (59/199, 29.6%). According to the anatomical segment of the leg affected by DVT, patients with DVT in the right femoral vein (50/71, 70.4%; P=0.016) had the highest rate of occurrence of PE. Conclusion : PE develops more frequently in patients with right-sided DVT than in those with left-sided DVT. Therefore, careful observation for the possible development of PE is recommended in cases with right-sided DVT of the lower extremity.