Patient Dose Distributions Research Articles

Prediction of dose distributions for non-small cell lung cancer patients using MHA-ResUNet.

The current level of automation in the production of radiotherapy plans for lung cancer patients is relatively low. With the development of artificial intelligence, it has become a reality to use neural networks to predict dose distributions and provide assistance for radiation therapy planning. However, due to the significant individual variability in the distribution of non-small cell lung cancer (NSCLC) planning target volume (PTV) and the complex spatial relationships between the PTV and organs at risk (OARs), there is still a lack of a high-precision dose prediction network tailored to the characteristics of NSCLC. To assist in the development of volumetric modulated arc therapy (VMAT) plans for non-small cell lung cancer patients, a deep neural network is proposed to predict high-precision dose distribution. This study has developed a network called MHA-ResUNet, which combines a large-kernel dilated convolution module and multi-head attention (MHA) modules. The network was trained based on 80 VMAT plans of NSCLC patients. CT images, PTV, and OARs were fed into the independent input channel. The dose distribution was taken as the output to train the model. The performance of this network was compared with that of several commonly used networks, and the networks' performance was evaluated based on the voxel-level mean absolute error (MAE) within the PTV and OARs, as well as the error in clinical dose-volume metrics. The MAE between the predicted dose distribution and the manually planned dose distribution within the PTV is 1.43Gy, and the D95 error is less than 1Gy. Compared with the other three commonly used networks, the dose error of the MHA-ResUNet is the smallest in PTV and OARs. The proposed MHA-ResUNet network improves the receptive field and filters the shallow features to learn the relative spatial relation between the PTV and the OARs, enabling accurate prediction of dose distributions in NSCLC patients undergoing VMAT radiotherapy.

Exploring the impact of field shape on predicted dose distribution in breast cancer patients using deep learning in radiation therapy

BackgroundGeometrical information such as field shape is essential for dose calculation in radiation therapy. However, new methods of dose prediction based on deep learning only use CT images and contouring of patients. The aim of this study is to develop a deep learning method for three-dimensional dose prediction for breast cancer radiotherapy using field shape as well as CT and contouring images of patients. MethodsA total of 150 breast cancer patients' data treated with 3D conformal radiation therapy (CRT) technique were used to train and test 3D-U-Res-F-Net model. The model inputs were CT images, patients' contouring and field shape. The network output was the corresponding the patients’ dose distribution. Then, the dosimetric parameters extracted from the dose volume histograms related to the planned and predicted dose distributions were compared and were statistically analysed with the paired-samples t-test. These parameters include Dmean and Dmax for planning target volume and organs at risk, D95%, D50%, V47.5Gy for PTV, V30Gy and V25Gy for heart and V20Gy for left lung. The gamma index with 3%/3 mm criteria was calculated for PTV and OARs. ResultsThe average absolute difference of the Dmean relative to the prescribed dose for the PTV, heart, left lung, right lung and spinal cord were 1.37 ± 0.4%, 2.02 ± 2%, 2.12 ± 1%, 0.37 ± 0.3% and 0.41 ± 0.3%, respectively. The average absolute difference of the Dmax relative to the prescribed dose for the PTV and OARs varied from 0.23 ± 0.2% to 2.04 ± 2%. The 3D gamma pass rate with 3mm/3% criteria for planning target volume, heart, left lung, right lung, spinal cord and body were 89 ± 3%, 91 ± 5%, 93 ± 5%, 99 ± 3%, 99 ± 1% and 96 ± 2% respectively. ConclusionsThe results of the proposed model demonstrate that there is no significant difference between the predicted and planned dose distributions. The deep learning model can directly predict the 3D dose distribution faster than treatment planning systems. Further studies with more patients and on other cancer sites are essential to fully validate the proposed method.

A novel 3D-printed brachytherapy applicator and Monte Carlo model for the treatment of conjunctival tumors

PurposeTo develop a custom low dose rate brachytherapy applicator for the treatment of conjunctival malignancies which leverages 3D-printing technology to provide enhanced design flexibility and availability. MethodsAn elliptical shell applicator inspired by the ocular surgery post-operation conformer shells was developed for the placement of the applicator around the cornea of the eye, with a central hole provide patient comfort. The applicator featured two concentric circles of slots for iodine-125 seeds, providing customization of the dose distribution depending on the location of the target. The applicator was modeled using computer-aided design software. The resultant model STL file was used for 3D printing of the applicator and the development of a Monte Carlo model of the applicator and its dose distribution. ResultsThe applicator was successfully 3D printed using biocompatible resin, which could be sterilized for treatment after manual source loading. A Geant4 model of the applicator was created directly from the STL model and was applied to a phantom to estimate the dose distribution delivered by the applicator. The toroidal dose distribution allowed for treatment of the conjunctiva while reducing dose to the cornea compared to traditional eye plaque designs. ConclusionsA custom 3D-printed applicator was successfully developed and modeled for the treatment of conjunctival malignancies. This novel applicator design potentially provides higher quality, more customizable dose distributions for patients and the simplicity of the design makes it accessible for any clinic with 3D-printing technology.

Anatomical changes and dosimetric analysis of the neck region based on FBCT for nasopharyngeal carcinoma patients during radiotherapy.

The study aimed to investigate anatomical changes in the neck region and evaluate their impact on dose distribution in patients with nasopharyngeal carcinoma (NPC) undergoing intensity modulated radiation therapy (IMRT). Additionally, the study sought to determine the optimal time for replanning during the course of treatment. Twenty patients diagnosed with NPC underwent IMRT, with weekly pretreatment kV fan beam computed tomography (FBCT) scans in the treatment room. Metastasized lymph nodes in the neck region and organs at risk (OARs) were redelineation using the images from the FBCT scans. Subsequently, the original treatment plan (PLAN0) was replicated to each FBCT scan to generate new plans labeled as PLAN 1-6. The dose-volume histograms (DVH) of the new plans and the original plan were compared. One-way repeated measure ANOVA was utilized to establish threshold(s) at various time points. The presence of such threshold(s) would signify significant change(s), suggesting the need for replanning. Progressive volume reductions were observed over time in the neck region, the gross target volume for metastatic lymph nodes (GTVnd), as well as the submandibular glands and parotids. Compared to PLAN0, the mean dose (Dmean) of GTVnd-L significantly increased in PLAN5, while the minimum dose covering 95% of the volume (D95%) of PGTVnd-L showed a significant decrease from PLAN3 to PLAN6. Similarly, the Dmean of GTVnd-R significantly increased from PLAN4 to PLAN6, whereas the D95% of PGTVnd-R exhibited a significant decrease during the same period. Furthermore, the dose of bilateral parotid glands, bilateral submandibular glands, brainstem and spinal cord was gradually increased in the middle and late period of treatment. Significant anatomical and dosimetric changes were noted in both the target volumes and OARs. Considering the thresholds identified, it is imperative to undertake replanning at approximately 20 fractions. This measure ensures the delivery of adequate doses to target volumes while mitigating the risk of overdosing on OARs.

Standardizing and improving dose predictions for head and neck cancers using complete sets of OAR contours.

Radiotherapy dose predictions have been trained with data from previously treated patients of similar sites and prescriptions. However, clinical datasets are often inconsistent and do not contain the same number of organ at risk (OAR) structures. The effects of missing contour data in deep learning-based dose prediction models have not been studied. The purpose of this study was to investigate the impacts of incomplete contour sets in the context of deep learning-based radiotherapy dose prediction models trained with clinical datasets and to introduce a novel data substitution method that utilizes automated contours for undefined structures. We trained Standard U-Nets and Cascade U-Nets to predict the volumetric dose distributions of patients with head and neck cancers (HNC) using three input variations to evaluate the effects of missing contours, as well as a novel data substitution method. Each architecture was trained with the original contour (OC) inputs, which included missing information, hybrid contour (HC) inputs, where automated OAR contours generated in software were substituted for missing contour data, and automated contour (AC) inputs containing only automated OAR contours. 120 HNC treatments were used for model training, 30 were used for validation and tuning, and 44 were used for evaluation and testing. Model performance and accuracy were evaluated with global whole body dose agreement, PTV coverage accuracy, and OAR dose agreement. The differences in these values between dataset variations were used to determine the effects of missing data and automated contour substitutions. Automated contours used as substitutions for missing data were found to improve dose prediction accuracy in the Standard U-Net and Cascade U-Net, with a statistically significant difference in some global metrics and/or OAR metrics. For both models, PTV coverage between input variations was unaffected by the substitution technique. Automated contours in HC and AC datasets improved mean dose accuracy for some OAR contours, including the mandible and brainstem, with a greater improvement seen with HC datasets. Global dose metrics, including mean absolute error, mean error, and percent error were different for the Standard U-Net but not for the Cascade U-Net. Automated contours used as a substitution for contour data improved prediction accuracy for some but not all dose prediction metrics. Compared to the Standard U-Net models, the Cascade U-Net achieved greater precision.

A deep learning-based dose prediction method for evaluation of radiotherapy treatment planning

[Purpose]To address the issue of low accuracy in dose distribution prediction in radiotherapy, we propose a deep learning-based model for predicting three-dimensional dose distribution in tumor radiation therapy. The model utilizes quantitative evaluation methods to assess the treatment plans. [Methods]We selected a dataset of 130 cervical cancer patients, including CT images and target region files. A deep learning U-Net model based on convolutional neural networks and residual blocks was employed to automatically extract multi-scale and multi-level feature maps of CT images, target regions, and anatomical structures of critical organs for intensity-modulated radiation therapy (IMRT) treatment plans and perform three-dimensional dose distribution prediction. Quantitative analysis methods, including error measures such as maximum dose (Dmax), mean dose (Dmean), V20, and D95, were used. [Results]For cervical cancer cases, DVH (dose-volume histogram) graphs were generated based on the evaluation results, directly reflecting the differences between the actual and predicted doses. The actual errors met the basic requirements, and a quantitative evaluation approach was used to optimize the dosimetric parameters. The specific quantification results are: DSC: 86.52 ± 9.31, 95% HD: 3.74 ± 1.49, JD: 0.112 ± 0.026, MSD: 0.067 ± 0.031. [Conclusion]Through training the deep learning model, we have successfully captured the complex nonlinear relationship between IMRT plan feature map parameters and three-dimensional dose distribution. In practical clinical applications, this trained model can accurately predict personalized three-dimensional dose distribution for new patients and effectively assess treatment plans in a quantitative manner. The source code is available at: https://github.com/xiebw9509/Radiotherapy_dose_prediction.

Development of an automatic optimization program for the weight of multi-field irradiation in BNCT based on a genetic algorithm

ABSTRACT Boron Neutron Capture Therapy (BNCT) is a promising radiotherapy technique for cancer treatment. Optimizing dose distribution in tumors is a significant challenge for BNCT, especially for tumors that are deep-seated. The method of optimizing dose distribution through multi-field irradiation holds great potential for clinical applications in BNCT. This study presents an optimization program that couples genetic algorithm and Monte Carlo simulation, which can automatically optimize irradiation field weights for multi-field irradiation, ensuring optimal dose distribution for cancer patients. Using this program, a glioblastoma case was used to analyze the optimization results for different numbers of irradiation fields. The results demonstrate that the program possesses an efficient and accurate automatic optimization feature for multi-field irradiation weights. The minimum dose, mean dose, and D80% of tumor dose improve with the increase in irradiation fields, while the homogeneity index (HI) of tumor dose decreases, without exceeding dose constraints for organs at risk. The developed optimization program provides a method for determining irradiation field weights in BNCT treatment planning systems, which has the potential to improve treatment outcomes.

A Photograph-Based Visualization and Prediction Framework for Radiation-Induced Dermatitis.

This study aimed to suggest a photograph-based prediction system for acute radiation-induced dermatitis (RID), which can be applied to notify patients about the risk of the development of skin discomfort during radiotherapy. The proposed system compared the spatial dose distribution with the RID region using the following methods. Skin photographs of patients were taken using an RGB-depth camera to acquire the shape information of RID. The skin surface data measured from the camera was registered with the shape of the external body contour using an iterative closest point algorithm. Spatial dose distribution of skin was extracted from the external body contour to a depth of 2 mm and projected onto the plane of the skin photograph using a transformation matrix for skin depth data. To compare the spatial distribution of skin dose with the shape of RID, the region of RID in patients' skin was delineated on photographs into three toxicity symptoms referring to the CTCAE criteria grade 1 (skin redness), grade 2 (dry desquamation), and grade 3 (moist desquamation). The degree of overlap between the shape of each RID and skin dose distribution was evaluated using the dice similarity coefficient (DSC). Threshold doses for predicting RID occurrence were estimated by skin isodose lines with the highest DSC. The developed system was validated using data from 19 patients who received volumetric modulated arc therapy for head-neck cancer at a single institution. Threshold doses for RID grades 1, 2, and 3 were estimated using 18, 18, and 2 individual RID labels delineated on skin photographs, respectively. Isodose lines with the highest DSC for RID grades 1, 2, and 3 were calculated as 26.0 Gy, 36.5 Gy, and 54.0 Gy, respectively. A strong overlap (average DSC > 0.6) was observed between isodose skin lines and the shape of RID labels in all RID grades. Assessing the spatial information of skin dose can be helpful in predicting acute RID. The region of RID shows a strong similarity with the skin dose distribution in head-neck patients. Visualization of skin dose on the patient photograph is potent to patient education for preparing the cosmetic discomfort during radiotherapy, which may lead to the improvement of the patient satisfaction in treatment.

Can a 10 MV-FFF Photon Beam be Used in Stereotactic Thalamotomy to Reduce Beam Delivery Times: A Dosimetric Evaluation

Stereotactic radiosurgery (SRS) relies on small fields to ablate lesions. Treating functional diseases, such as essential tremor, requires a significantly higher dose (150 Gy) to a small target in a single fraction. Currently, Linac based treatment is delivered via pinhole-shaped fields formed by a 4 mm circular cone using a 6 MV photon beam. Utilizing a higher energy, high dose rate beam such as a 10 MV flattening-filter-free (FFF) beam with dose rates up to 2400 cGy/min could reduce long radiation delivery times by a factor of 4 comparing to the standard 6 MV beam with dose rate up to 600 cGy/min. This study aims to assess the benefit and feasibility of using a 10 MV-FFF beam by evaluating the dosimetric effects between the proposed 10 MV-FFF and commonly used 6 MV beam and comparing the dose to organ-at-risk (OARs) based on a typical real patient tremor treatment plan. The planning tool codes BEAMnrc/DOSXYZnrc were used to generate beams of 6 MV and 10 MV-FFF for a medical linear accelerator. Calculated patient dose distributions were compared between two energy beams based on a realistic treatment plan for thalamotomy. Additionally, dose to OARs were evaluated by using the dose-volume histograms (DVH) for the same target dose coverage. The mean energies of photons within the 4 mm cone projected area were 1.73 MeV and 2.09 MeV and the values of percentage depth-dose curve (%dd) at d = 5 cm and SSD = 95 cm were 75.5% and 78.3%, for the 6 MV and 10 MV-FFF beams, respectively. A patient delivered treatment plan with a prescribed dose of 15,000 cGy utilizing the 6MV-SRS beam was developed which used a 4 mm cone with a 21-arc delivery. Using that plan as the template, doses to OARs were evaluated comparing the 6MV to the 10MV-FFF energies. The maximum doses to the nearest OAR, brainstem, was 733 cGy and 822 cGy for the 6 MV and 10 MV-FFF beams, respectively. The details of maximum and mean doses to OARs are presented in Table 1 between the two energies. Compared to the 6 MV delivery, there is only a 12% increase in the maximum dose to the brain stem when using the 10 MV-FFF delivery based on the patient treatment plan. If the mild dose increase to OAR is acceptable for using 10 MV-FFF, the treatment time with 10 MV-FFF beam can be reduced to less than 1 hour from 3 hours with the 6 MV beam. The shorter treatment time can increase the accuracy of treatment target positioning as well patient comfort. Table 1 Calculated doses to OARs based on the real patient plan to deliver the same 15,000 cGy target dose between using 6 MV and 10 MV-FFF beams.

Simulation of consequences of using nonideal detectors during beam data commissioning measurements.

Beam data commissioning is a core task of radiotherapy physicists. Despite multiple detectors available, a feasible measurement program compromises between detector properties and time constraints. Therefore, it is important to understand how nonideal measurement data propagates into patient dose calculation. We simulated the effects of realistic errors, due to beam commissioning with presumably nonoptimal detectors, on the resulting patient dose distributions. Additionally, the detectability of such beam commissioning errors during patient plan quality assurance (QA) was evaluated. A clinically used beam model was re-commissioned introducing changes to depth dose curves, output factors, profiles or combinations of those. Seventeen altered beam models with incremental changes of the modelling parameters were created to analyze dose changes on simplified anatomical phantoms. Additionally, fourteen altered models incorporate changes in the order of signal differences reported for typically used detectors. Eighteen treatment plans of different types were recalculated on patient CT data sets using the altered beam models. For the majority of clinical plans, dose distributions in the target volume recalculated on the patient computed tomography data were similar between the original and the modified beam models, yielding global 2%/2mm gamma pass rates above 98.9%. Larger changes were observed for certain combinations of beam modelling errors and anatomical sites, most extreme for output factor changes in a small target volume plan with a pass rate of 80.6%. Modelling an enlarged penumbra as if measured with a 0.125cm3 ion chamber had the largest effect on the dose distribution (average pass rate of 96.5%, lowest 85.4%). On different QA phantom geometries, dose distributions between calculations with modified and unmodified models typically changed too little to be detected in actual measurements. While the simulated errors during beam modelling had little effect on most plans, in some cases changes were considerable. High-quality penumbra and small field output factor should be a main focus of commissioning measurements. Detecting modelling issues using standard patient QA phantoms is unlikely. Verification of a beam model should be performed especially for plans with high modulation and in different depths or geometries representing the variety of situations expected clinically.

Improvement of deep learning prediction model in patient-specific QA for VMAT with MLC leaf position map and patient's dose distribution.

Deep learning-based virtual patient-specific quality assurance (QA) is a novel technique that enables patient QA without measurement. However, this method could be improved by further evaluating the optimal data to be used as input. Therefore, a deep learning-based model that uses multileaf collimator (MLC) information per control point and dose distribution in patient's CT as inputs was developed. Overall, 96 volumetric-modulated arc therapy plans generated for prostate cancer treatment were used. We developed a model (Model 1) that can predict measurement-based gamma passing rate (GPR) for a treatment plan using data stored as a map reflecting the MLC leaf position at each control point (MLPM) and data of the dose distribution in patient's CT as inputs. The evaluation of the model was based on the mean absolute error (MAE) and Pearson's correlation coefficient (r) between the measured and predicted GPR. For comparison, we also analyzed models trained with the dose distribution in patient's CT alone (Model 2) and with dose distributions recalculated on a virtual phantom CT (Model 3). At the 2%/2mm criterion, MAE[%] and r for Model 1, Model 2, and Model 3 were 2.32%±0.43% and 0.54±0.03, 2.70%±0.26%, and 0.32±0.08, and 2.96%±0.23% and 0.24±0.22, respectively; at the 3%/3mm criterion, these values were 1.25%±0.05% and 0.36±0.18, 1.57%±0.35% and 0.19±0.20, and 1.39%±0.32% and 0.17±0.22, respectively. This result showed that Model 1 exhibited the lowest MAE and highest r at both criteria of 2%/2mm and 3%3mm. These findings showed that a model that combines the MLPM and dose distribution in patient's CT exhibited a better GPR prediction performance compared with the other two studied models.

The effects of different photon beam energies in stereotactic radiosurgery with cones.

Stereotactic radiosurgery (SRS) relies on small fields to ablate lesions. Currently, linac based treatment is delivered via circular cones using a 6 MV beam. There is interest in both lower energy photon beams, which can offer steeper dose fall off as well as higher energy photon beams, which have higher dose rates, thus reducing radiation delivery times. Of interest in this study is the 2.5 MV beam developed for imaging applications and both the 6 and 10 MV flattening-filter-free (FFF) beams, which can achieve dose rates up to 2400 cGy/min. This study aims to assess the benefit and feasibility among different energy beams ranging from 2.5 to 10 MV beams by evaluating the dosimetric effects of each beam and comparing the dose to organs-at-risk (OARs) for two separate patient plans. One based on a typical real patient tremor utilizing a 4mm cone and the other a typical brain metastasis delivered with a 10mm cone. The Monte Carlo codes BEAMnrc/DOSXYZnrc were used to generate beams of 2.5 MV, 6 MV-FFF, 6 MV-SRS, 6 MV, 10 MV-FFF, and 10 MV from a Varian TrueBeam except 6 MV-SRS, which is taken from a Varian TX model linear accelerator. Each beam's energy spectrum, mean energy, %dd curve, and dose profile were obtained by analyzing the simulated beams. Calculated patient dose distributions were compared among six different energy beam configurations based on a realistic treatment plan for thalamotomy and a conventional brain metastasis plan. Dose to OARs were evaluated using dose-volume histograms for the same target dose coverage. The mean energies of photons within the primary beam projected area were insensitive to cone sizes and the values of percentage depth-dose curves (%dd) at d=5cm and SSD=95cm for a 4mm (10mm) cone ranges from 62.6 (64.4) to 82.2 (85.7) for beam energy ranging from 2.5 to 10 MV beams, respectively. Doses to OARs were evaluated among these beams based on real treatment plans delivering 15000 and 2200cGy to the target with a 4 and 10mm cone, respectively. The maximum doses to the brainstem, which is 10mm away from the isocenter, was found to be 434 (300), 632 (352), 691 (362), 733 (375), 822 (403), and 975 (441)cGy for 2.5 MV, 6 MV-FFF, 6 MV-SRS, 6 MV, 10 MV-FFF, and 10 MV beams delivering 15000 (2200) cGy target dose, respectively. Using the 6 MV-SRS as reference, changes of the maximum dose (691 cGy) to the brain stem are -37%, -9%, +6%, +19%, and 41% for 2.5 MV, 6 MV-FFF, 6 MV, 10 MV-FFF, and 10 MV beams, respectively, based on the thalamotomy plan, where the "-" or "+" signs indicate the percentage decrease or increase. Changes of the maximum dose (362 cGy) to brain stem, based on the brain metastasis plan are much less for respective beam energies. The sum of 21 arcs beam-on time was 39 min on our 6 MV-SRS beam with 1000 cGy/min for thalamotomy. The beam-on time can be reduced to 16 min with 10 MV-FFF.

Effect of internal port on dose distribution in post-mastectomy radiotherapy for breast cancer patients after expander breast reconstruction.

In patients with expander-based reconstruction a few dosimetric analyses detected radiation therapy dose perturbation due to the internal port of an expander, potentially leading to toxicity or loss of local control. This study aimed at adding data on this field. A dosimetric analysis was conducted in 30 chest wall treatment planning without and with correction for port artifact. In plans with artifact correction density was overwritten as 1 g/cm3. Medium, minimum and maximum chest wall doses were compared in the two plans. Both plans, with and without correction, were compared on an anthropomorphic phantom with a tissue expander on the chest covered by a bolus simulating the skin. Ex vivo dosimetry was carried out on the phantom and in vivo dosimetry in three patients by using film strips during one treatment fraction. Estimated doses and measured film doses were compared. No significant differences emerged in the minimum, medium and maximum doses in the two plans, without and with correction for port artifacts. Ex vivo and in vivo analyses showed a good correspondence between detected and calculated doses without and with correction. The port did not significantly affect dose distribution in patients who will receive post-mastectomy radiation therapy.

Radiotherapy delivery error detection with electronic portal imaging device (EPID)-based in vivo dosimetry.

Radiotherapy delivery error detection with electronic portal imaging device (EPID)-based in vivo dosimetry.

Experiment of proof-of-principle on prompt gamma-positron emission tomography (PG-PET) system for in-vivo dose distribution verification in proton therapy

In our previous study, we proposed an integrated PG-PET-based imaging method to increase the prediction accuracy for patient dose distributions. The purpose of the present study is to experimentally validate the feasibility of the PG-PET system. Based on the detector geometry optimized in the previous study, we constructed a dual-head PG-PET system consisting of a 16 × 16 GAGG scintillator and KETEK SiPM arrays, BaSO4 reflectors, and an 8 × 8 parallel-hole tungsten collimator. The performance of this system as equipped with a proof of principle, we measured the PG and positron emission (PE) distributions from a 3 × 6 × 10 cm3 PMMA phantom for a 45 MeV proton beam. The measured depth was about 17 mm and the expected depth was 16 mm in the computation simulation under the same conditions as the measurements. In the comparison result, we can find a 1 mm difference between computation simulation and measurement. In this study, our results show the feasibility of the PG-PET system for in-vivo range verification. However, further study should be followed with the consideration of the typical measurement conditions in the clinic application.

Characterization of patient dose distributions and probabilistic risk analysis in cardiovascular catheterization laboratories

Purpose: Patient radiation doses in cardiovascular and interventional radiology are highly variable for similar procedures. This random nature may be better described by a distribution function, compared to a linear regression. This study develops a distribution function to characterize patient dose distributions and estimate probabilistic risk. Methods: Reference air kerma from 8647 patients over six years were retrospectively collected from an EP lab and two Cath labs. Data was first sorted into low dose (<100mGy) and high dose cases (>100mGy), and histograms of the data created. Dagum and Inverse gamma distributions were chosen to initially fit to both low and high dose cases. Fits between model and the data was optimized, and a linear regression analysis performed to obtain R squared values and standard errors for the correlation between model and data. Risk probabilities were estimated according to the modeled distribution function. BMI and time distributions were analyzed to understand their influence on the inverse gamma distribution error found in the data. 75th percentiles from both descriptive statistics and model were calculated. Results: The inverse gamma distribution can be used to characterize radiation dose distributions. Model predicted cases for radiation dose 3000mGy< x <5000mGy and >5000mGy are approximately 42 and 0 for 3651 cases for lab#1, and 14 and 1 for 3197 cases for lab#2, respectively, while the actual cases are 10 and 0, and 16 and 2. Descriptive and model statistics generated different 75th percentile levels for sorted data compared to unsorted data. Time has a greater influence on inverse gamma distribution function than BMI. Conclusion: This study creates a framework to understand the random error present in radiology practices that cause wide variations in patient radiation doses. It also provides an approach to evaluate different IR areas in terms of effectiveness of dose reduction measures.

New technique and application of truncated CBCT processing in adaptive radiotherapy for breast cancer

ObjectiveA generative adversarial network (TCBCTNet) was proposed to generate synthetic computed tomography (sCT) from truncated low-dose cone-beam computed tomography (CBCT) and planning CT (pCT). The sCT was applied to the dose calculation of radiotherapy for patients with breast cancer. MethodsThe low-dose CBCT and pCT images of 80 female thoracic patients were used for training. The CBCT, pCT, and replanning CT (rCT) images of 20 thoracic patients and 20 patients with breast cancer were used for testing. All patients were fixed in the same posture with a vacuum pad. The CBCT images were scanned under the Fast Chest M20 protocol with a 50% reduction in projection frames compared with the standard Chest M20 protocol. Rigid registration was performed between pCT and CBCT, and deformation registration was performed between rCT and CBCT. In the training stage of the TCBCTNet, truncated CBCT images obtained from complete CBCT images by simulation were used. The input of the CBCT→CT generator was truncated CBCT and pCT, and TCBCTNet was applied to patients with breast cancer after training. The accuracy of the sCT was evaluated by anatomy and dosimetry and compared with the generative adversarial network with UNet and ResNet as the generators (named as UnetGAN, ResGAN). ResultsThe three models could improve the image quality of CBCT and reduce the scattering artifacts while preserving the anatomical geometry of CBCT. For the chest test set, TCBCTNet achieved the best mean absolute error (MAE, 21.18±3.76 HU), better than 23.06±3.90 HU in UnetGAN and 22.47±3.57 HU in ResGAN. When applied to patients with breast cancer, TCBCTNet performance decreased, and MAE was 25.34±6.09 HU. Compared with rCT, sCT by TCBCTNet showed consistent dose distribution and subtle absolute dose differences between the target and the organ at risk. The 3D gamma pass rates were 98.98%±0.64% and 99.69%±0.22% at 2 mm/2% and 3 mm/3%, respectively. Ablation experiments confirmed that pCT and content loss played important roles in TCBCTNet. ConclusionsHigh-quality sCT images could be synthesized from truncated low-dose CBCT and pCT by using the proposed TCBCTNet model. In addition, sCT could be used to accurately calculate the dose distribution for patients with breast cancer.

Use of a neural network-based prediction method to calculate the therapeutic dose in boron neutron capture therapy of patients with glioblastoma.

Boron neutron capture therapy (BNCT) is a binary radiotherapy based on the 10 B(n, α)7 Li capture reaction. Nonradioactive isotope 10 B atoms which selectively concentrated in tumor cells will react with low energy neutrons (mainly thermal neutrons) to produce secondary particles with high linear energy transfer, thus depositing dose in tumor cells. In clinical practice, an appropriate treatment plan needs to be set on the basis of the treatment planning system (TPS). Existing BNCT TPSs usually use the Monte Carlo method to determine the three-dimensional (3D) therapeutic dose distribution, which often requires a lot of calculation time due to the complexity of simulating neutron transportation. A neural network-based BNCT dose prediction method is proposed to achieve the rapid and accurate acquisition of BNCT 3D therapeutic dose distribution for patients with glioblastoma to solve the time-consuming problem of BNCT dose calculation in clinic. The clinical data of 122 patients with glioblastoma are collected. Eighteen patients are used as a test set, and the rest are used as a training set. The 3D-UNET is constructed through the design optimization of input and output data sets based on radiation field information and patient CT information to enable the prediction of 3D dose distribution of BNCT. The average mean absolute error of the predicted and simulated equivalent doses of each organ are all less than 1Gy. For the dose to 95% of the GTV volume (D95 ), the relative deviation between predicted and simulated results are all less than 2%. The average 2mm/2% gamma index is 89.67%, and the average 3mm/3% gamma index is 96.78%. The calculation takes about 6h to simulate the 3D therapeutic dose distribution of a patient with glioblastoma by Monte Carlo method using Intel Xeon E5-2699 v4, whereas the time required by the method proposed in this study is almost less than 1 s using a Titan-V graphics card. This study proposes a 3D dose prediction method based on 3D-UNET architecture in BNCT, and the feasibility of this method is demonstrated. Results indicate that the method can remarkably reduce the time required for calculation and ensure the accuracy of the predicted 3D therapeutic dose-effect. This work is expected to promote the clinical development of BNCT in the future.

Direct Dose Prediction With Deep Learning for Postoperative Cervical Cancer Underwent Volumetric Modulated Arc Therapy.

To predict the voxel-based dose distribution for postoperative cervical cancer patients underwent volumetric modulated arc therapy using deep learning models. A total of 254 patients with cervical cancer received volumetric modulated arc therapy in authors' hospital from January 2018 to September 2021 were enrolled in this retrospective study. Two deep learning networks (3D deep residual neural network and 3DUnet) were adapted to train (203 cases) and test (51 cases) the feasibility and effectiveness of the prediction method. The performance of deep learning models was evaluated by comparing the results with those of treatment planning system according to metrics of dose-volume histogram of target volumes and organs at risk. The dose distributions predicted by deep learning models were clinically acceptable. The automatic dose prediction time was around 5 to 10 min, which was about one-eighth to one-tenth of the manual optimization time. The maximum dose difference was observed in D98 of rectum with a | δD| of 5.00 ± 3.40% and 4.88 ± 3.99% for Unet3D and ResUnet3D, respectively. The minimum difference was observed in the D2 of clinical target volume with a |δD| of 0.53 ± 0.45% and 0.83 ± 0.45% for ResUnet3D and Unet3D, respectively. The 2 deep learning models adapted in the study showed the feasibility and reasonable accuracy in the voxel-based dose prediction for postoperative cervical cancer underwent volumetric modulated arc therapy. Automatic dose distribution prediction of volumetric modulated arc therapy with deep learning models is of clinical significance for the postoperative management of patients with cervical cancer.

In Silico Studies of the Impact of Rotational Errors on Translation Shifts and Dose Distribution in Image-Guided Radiotherapy.

Objective: To compare the 6-dimensional errors of different immobilization devices and body regions based on 3-dimensional cone beam computed tomography for image-guided radiotherapy and to further quantitatively evaluate the impact of rotational corrections on translational shifts and dose distribution based on anthropomorphic phantoms. Materials and Methods: Two hundred ninety patients with cone beam computed tomographies from 3835 fractions were retrospectively analyzed for brain, head & neck, chest, abdomen, pelvis, and breast cases. A phantom experiment was conducted to investigate the impact of rotational errors on translational shifts using cone beam computed tomography and the registration system. For the dosimetry study, pitch rotations were simulated by adjusting the breast bracket by ±2.5°. Roll and yaw rotations were simulated by rotating the gantry and couch in the planning system by ±3.0°, respectively. The original plan for the breast region was designed in the computed tomography image space without rotation. With the same planning parameters, the original plan was transplanted into the image space with different rotations for dose recalculation. The effect of these errors on the breast target and organs at risk was assessed by dose-volume histograms. Results: Most of the mean rotational errors in the breast region were >1°. A single uncorrected yaw of 3° caused a change of 2.9 mm in longitudinal translation. A phantom study for the breast region demonstrated that when the pitch rotations were -2.5° and 2.5° and roll and yaw were both 3°, the reductions in the planning target volumes-V50 Gy were 20.07% and 29.58% of the original values, respectively. When the pitch rotation was +2.5°, the left lung V5 Gy and heart Dmean were 7.49% and 165.76 Gy larger, respectively, than the original values. Conclusions: Uncorrected rotations may cause changes in the values and directions of translational shifts. Rotational corrections may improve the patient setup and dose distribution accuracy.