A 3D transfer learning approach for identifying multiple simultaneous errors during radiotherapy.

Objective. Deep learning models, such as convolutional neural networks (CNNs), can take full dose comparison images as input and have shown promising results for error identification during treatment. Clinically, complex scenarios should be considered, with the risk of multiple anatomical and/or mechanical errors occurring simultaneously during treatment. The purpose of this study was to evaluate the capability of CNN-based error identification in this more complex scenario.Approach. For 40 lung cancer patients, clinically realistic ranges of combinations of various treatment errors within treatment plans and/or computed tomography (CT) images were simulated. Modified CT images and treatment plans were used to predict 2580 3D dose distributions, which were compared to dose distributions without errors using various gamma analysis criteria and relative dose difference as dose comparison methods. A 3D CNN capable of multilabel classification was trained to identify treatment errors at two classification levels, using dose comparison volumes as input: Level 1 (main error type, e.g. anatomical change, mechanical error) and Level 2 (error subtype, e.g. tumor regression, patient rotation). For training the CNNs, a transfer learning approach was employed. An ensemble model was also evaluated, which consisted of three separate CNNs each taking a region of interest of the dose comparison volume as input. Model performance was evaluated by calculating sample F1-scores for training and validation sets.Main results. The model had high F1-scores for Level 1 classification, but performance for Level 2 was lower, and overfitting became more apparent. Using relative dose difference instead of gamma volumes as input improved performance for Level 2 classification, whereas using an ensemble model additionally reduced overfitting. The models obtained F1-scores of 0.86 and 0.62 on an independent test set for Level 1 and Level 2, respectively.Significance. This study shows that it is possible to identify multiple errors occurring simultaneously in 3D dose verification data.

Shoot-through proton FLASH irradiation lowers linear energy transfer in organs at risk for neurological tumors and is robust against density variations.

Objective. The goal of the study was to test the hypothesis that shoot-through FLASH proton beams would lead to lower dose-averaged LET (LETD) values in critical organs, while providing at least equal normal tissue sparing as clinical proton therapy plans.Approach. For five neurological tumor patients, pencil beam scanning (PBS) shoot-through plans were made, using the maximum energy of 227 MeV and assuming a hypothetical FLASH protective factor (FPF) of 1.5. The effect of different FPF ranging from 1.2 to 1.8 on the clinical goals were also considered. LETDwas calculated for the clinical plan and the shoot-through plan, applying a 2 Gy total dose threshold (RayStation 8 A/9B and 9A-IonRPG). Robust evaluation was performed considering density uncertainty (±3% throughout entire volume).Main results.Clinical plans showed large LETDvariations compared to shoot-through plans and the maximum LETDin OAR is 1.2-8 times lower for the latter. Although less conformal, shoot-through plans met the same clinical goals as the clinical plans, for FLASH protection factors above 1.4. The FLASH shoot-through plans were more robust to density uncertainties with a maximum OAR D2%increase of 0.6 Gy versus 5.7 Gy in the clinical plans.Significance.Shoot-through proton FLASH beams avoid uncertainties in LETDdistributions and proton range, provide adequate target coverage, meet planning constraints and are robust to density variations.

A deep-learning assisted bioluminescence tomography method to enable radiation targeting in rat glioblastoma.

Objective. A novel solution is required for accurate 3D bioluminescence tomography (BLT) based glioblastoma (GBM) targeting. The provided solution should be computationally efficient to support real-time treatment planning, thus reducing the x-ray imaging dose imposed by high-resolution micro cone-beam CT.Approach. A novel deep-learning approach is developed to enable BLT-based tumor targeting and treatment planning for orthotopic rat GBM models. The proposed framework is trained and validated on a set of realistic Monte Carlo simulations. Finally, the trained deep learning model is tested on a limited set of BLI measurements of real rat GBM models.Significance. Bioluminescence imaging (BLI) is a 2D non-invasive optical imaging modality geared toward preclinical cancer research. It can be used to monitor tumor growth in small animal tumor models effectively and without radiation burden. However, the current state-of-the-art does not allow accurate radiation treatment planning using BLI, hence limiting BLI's value in preclinical radiobiology research.Results. The proposed solution can achieve sub-millimeter targeting accuracy on the simulated dataset, with a median dice similarity coefficient (DSC) of 61%. The provided BLT-based planning volume achieves a median encapsulation of more than 97% of the tumor while keeping the median geometrical brain coverage below 4.2%. For the real BLI measurements, the proposed solution provided median geometrical tumor coverage of 95% and a median DSC of 42%. Dose planning using a dedicated small animal treatment planning system indicated good BLT-based treatment planning accuracy compared to ground-truth CT-based planning, where dose-volume metrics for the tumor fall within the limit of agreement for more than 95% of cases.Conclusion. The combination of flexibility, accuracy, and speed of the deep learning solutions make them a viable option for the BLT reconstruction problem and can provide BLT-based tumor targeting for the rat GBM models.

Automatic quality assurance of radiotherapy treatment plans using Bayesian networks: A multi-institutional study.

Purpose: Artificial intelligence applications in radiation oncology have been the focus of study in the last decade. The introduction of automated and intelligent solutions for routine clinical tasks, such as treatment planning and quality assurance, has the potential to increase safety and efficiency of radiotherapy. In this work, we present a multi-institutional study across three different institutions internationally on a Bayesian network (BN)-based initial plan review assistive tool that alerts radiotherapy professionals for potential erroneous or suboptimal treatment plans. Methods: Clinical data were collected from the oncology information systems in three institutes in Europe (Maastro clinic - 8753 patients treated between 2012 and 2020) and the United States of America (University of Vermont Medical Center [UVMMC] - 2733 patients, University of Washington [UW] - 6180 patients, treated between 2018 and 2021). We trained the BN model to detect potential errors in radiotherapy treatment plans using different combinations of institutional data and performed single-site and cross-site validation with simulated plans with embedded errors. The simulated errors consisted of three different categories: i) patient setup, ii) treatment planning and iii) prescription. We also compared the strategy of using only diagnostic parameters or all variables as evidence for the BN. We evaluated the model performance utilizing the area under the receiver-operating characteristic curve (AUC). Results: The best network performance was observed when the BN model is trained and validated using the dataset in the same center. In particular, the testing and validation using UVMMC data has achieved an AUC of 0.92 with all parameters used as evidence. In cross-validation studies, we observed that the BN model performed better when it was trained and validated in institutes with similar technology and treatment protocols (for instance, when testing on UVMMC data, the model trained on UW data achieved an AUC of 0.84, compared with an AUC of 0.64 for the model trained on Maastro data). Also, combining training data from larger clinics (UW and Maastro clinic) and using it on smaller clinics (UVMMC) leads to satisfactory performance with an AUC of 0.85. Lastly, we found that in general the BN model performed better when all variables are considered as evidence. Conclusion: We have developed and validated a Bayesian network model to assist initial treatment plan review using multi-institutional data with different technology and clinical practices. The model has shown good performance even when trained on data from clinics with divergent profiles, suggesting that the model is able to adapt to different data distributions.

What is the optimal input information for deep learning-based pre-treatment error identification in radiotherapy?

Background and purpose: Deep learning (DL) provides high sensitivity for detecting and identifying errors in pre-treatment radiotherapy quality assurance (QA). This work's objective was to systematically evaluate the impact of different dose comparison and image preprocessing methods on DL model performance for error identification in pre-treatment QA. Materials and methods: For 53 volumetric modulated arc therapy (VMAT) and 69 stereotactic body radiotherapy (SBRT) treatment plans of lung cancer patients, mechanical errors were simulated (MLC leaf positions, monitor unit scaling, collimator rotation). Two classification levels were assessed: error type (Level 1) and error magnitude (Level 2). Portal dose images with and without errors were compared using standard (gamma analysis), simple (absolute/relative dose difference, ratio) and alternative (distance-to-agreement, structural similarity index, gradient) dose comparison methods. For preprocessing, different normalization methods (min/max and mean/standard deviation) and image resolutions (32 × 32, 64 × 64 and 128 × 128) were evaluated. All possible combinations of classification level, dose comparison, normalization method and image size resulted in 144 input datasets for DL networks for error identification. Results: Average accuracy was highest for simple dose comparison methods (Level 1: 97.7%, Level 2: 78.1%) while alternative methods scored lowest (Level 1: 91.6%, Level 2: 71.2%). Mean/stdev normalization particularly improved Level 2 classification. Higher image resolution improved error identification, although for SBRT lower image resolution was also sufficient. Conclusions: The choice of dose comparison method has the largest impact on error identification for pre-treatment QA using DL, compared to image preprocessing. Model performance can improve by using simple dose comparison methods, mean/stdev normalization and high image resolution.

A novel multichannel deep learning model for fast denoising of Monte Carlo dose calculations: preclinical applications.

Objective.In preclinical radiotherapy with kilovolt (kV) x-ray beams, accurate treatment planning is needed to improve the translation potential to clinical trials. Monte Carlo based radiation transport simulations are the gold standard to calculate the absorbed dose distribution in external beam radiotherapy. However, these simulations are notorious for their long computation time, causing a bottleneck in the workflow. Previous studies have used deep learning models to speed up these simulations for clinical megavolt (MV) beams. For kV beams, dose distributions are more affected by tissue type than for MV beams, leading to steep dose gradients. This study aims to speed up preclinical kV dose simulations by proposing a novel deep learning pipeline.Approach.A deep learning model is proposed that denoises low precision (∼106simulated particles) dose distributions to produce high precision (109simulated particles) dose distributions. To effectively denoise the steep dose gradients in preclinical kV dose distributions, the model uses the novel approach to use the low precision Monte Carlo dose calculation as well as the Monte Carlo uncertainty (MCU) map and the mass density map as additional input channels. The model was trained on a large synthetic dataset and tested on a real dataset with a different data distribution. To keep model inference time to a minimum, a novel method for inference optimization was developed as well.Main results.The proposed model provides dose distributions which achieve a median gamma pass rate (3%/0.3 mm) of 98% with a lower bound of 95% when compared to the high precision Monte Carlo dose distributions from the test set, which represents a different dataset distribution than the training set. Using the proposed model together with the novel inference optimization method, the total computation time was reduced from approximately 45 min to less than six seconds on average.Significance.This study presents the first model that can denoise preclinical kV instead of clinical MV Monte Carlo dose distributions. This was achieved by using the MCU and mass density maps as additional model inputs. Additionally, this study shows that training such a model on a synthetic dataset is not only a viable option, but even increases the generalization of the model compared to training on real data due to the sheer size and variety of the synthetic dataset. The application of this model will enable speeding up treatment plan optimization in the preclinical workflow.

A deep learning and Monte Carlo based framework for bioluminescence imaging center of mass-guided glioblastoma targeting.

Objective.Bioluminescence imaging (BLI) is a valuable tool for non-invasive monitoring of glioblastoma multiforme (GBM) tumor-bearing small animals without incurring x-ray radiation burden. However, the use of this imaging modality is limited due to photon scattering and lack of spatial information. Attempts at reconstructing bioluminescence tomography (BLT) using mathematical models of light propagation show limited progress.Approach.This paper employed a different approach by using a deep convolutional neural network (CNN) to predict the tumor's center of mass (CoM). Transfer-learning with a sizeable artificial database is employed to facilitate the training process for, the much smaller, target database including Monte Carlo (MC) simulations of real orthotopic glioblastoma models. Predicted CoM was then used to estimate a BLI-based planning target volume (bPTV), by using the CoM as the center of a sphere, encompassing the tumor. The volume of the encompassing target sphere was estimated based on the total number of photons reaching the skin surface.Main results.Results show sub-millimeter accuracy for CoM prediction with a median error of 0.59 mm. The proposed method also provides promising performance for BLI-based tumor targeting with on average 94% of the tumor inside the bPTV while keeping the average healthy tissue coverage below 10%.Significance.This work introduced a framework for developing and using a CNN for targeted radiation studies for GBM based on BLI. The framework will enable biologists to use BLI as their main image-guidance tool to target GBM tumors in rat models, avoiding delivery of high x-ray imaging dose to the animals.

Inter-observer variability of organ contouring for preclinical studies with cone beam Computed Tomography imaging.

BACKGROUND AND PURPOSE: In preclinical radiation studies, there is great interest in quantifying the radiation response of healthy tissues. Manual contouring has significant impact on the treatment-planning because of variation introduced by human interpretation. This results in inconsistencies when assessing normal tissue volumes. Evaluation of these discrepancies can provide a better understanding on the limitations of the current preclinical radiation workflow. In the present work, interobserver variability (IOV) in manual contouring of rodent normal tissues on cone-beam Computed Tomography, in head and thorax regions was evaluated. MATERIALS AND METHODS: Two animal technicians performed manually (assisted) contouring of normal tissues located within the thorax and head regions of rodents, 20 cases per body site. Mean surface distance (MSD), displacement of center of mass (ΔCoM), DICE similarity coefficient (DSC) and the 95th percentile Hausdorff distance (HD95) were calculated between the contours of the two observers to evaluate the IOV. RESULTS: For the thorax organs, right lung had the lowest IOV (ΔCoM: 0.08 ±â€¯0.04 mm, DSC: 0.96 ±â€¯0.01, MSD:0.07 ±â€¯0.01 mm, HD95:0.20 ±â€¯0.03 mm) while spinal cord, the highest IOV (ΔCoM:0.5 ±â€¯0.3 mm, DSC:0.81 ±â€¯0.05, MSD:0.14 ±â€¯0.03 mm, HD95:0.8 ±â€¯0.2 mm). Regarding head organs, right eye demonstrated the lowest IOV (ΔCoM:0.12 ±â€¯0.08 mm, DSC: 0.93 ±â€¯0.02, MSD: 0.15 ±â€¯0.04 mm, HD95: 0.29 ±â€¯0.07 mm) while complete brain, the highest IOV (ΔCoM: 0.2 ±â€¯0.1 mm, DSC: 0.94 ±â€¯0.02, MSD: 0.3 ±â€¯0.1 mm, HD95: 0.5 ±â€¯0.1 mm). CONCLUSIONS: Our findings reveal small IOV, within the sub-mm range, for thorax and head normal tissues in rodents. The set of contours can serve as a basis for developing an automated delineation method for e.g., treatment planning.

Automatic contouring of normal tissues with deep learning for preclinical radiation studies.

Objective.Delineation of relevant normal tissues is a bottleneck in image-guided precision radiotherapy workflows for small animals. A deep learning (DL) model for automatic contouring using standardized 3D micro cone-beam CT (µCBCT) volumes as input is proposed, to provide a fully automatic, generalizable method for normal tissue contouring in preclinical studies.Approach.A 3D U-net was trained to contour organs in the head (whole brain, left/right brain hemisphere, left/right eye) and thorax (complete lungs, left/right lung, heart, spinal cord, thorax bone) regions. As an important preprocessing step, Hounsfield units (HUs) were converted to mass density (MD) values, to remove the energy dependency of theµCBCT scanner and improve generalizability of the DL model. Model performance was evaluated quantitatively by Dice similarity coefficient (DSC), mean surface distance (MSD), 95th percentile Hausdorff distance (HD95p), and center of mass displacement (ΔCoM). For qualitative assessment, DL-generated contours (for 40 and 80 kV images) were scored (0: unacceptable, manual re-contouring needed - 5: no adjustments needed). An uncertainty analysis using Monte Carlo dropout uncertainty was performed for delineation of the heart.Main results.The proposed DL model and accompanying preprocessing method provide high quality contours, with in general median DSC > 0.85, MSD < 0.25 mm, HD95p < 1 mm and ΔCoM < 0.5 mm. The qualitative assessment showed very few contours needed manual adaptations (40 kV: 20/155 contours, 80 kV: 3/155 contours). The uncertainty of the DL model is small (within 2%).Significance.A DL-based model dedicated to preclinical studies has been developed for multi-organ segmentation in two body sites. For the first time, a method independent of image acquisition parameters has been quantitatively evaluated, resulting in sub-millimeter performance, while qualitative assessment demonstrated the high quality of the DL-generated contours. The uncertainty analysis additionally showed that inherent model variability is low.

Automatic coronavirus disease 2019 diagnosis based on chest radiography and deep learning - Success story or dataset bias?

PURPOSE: Over the last 2 years, the artificial intelligence (AI) community has presented several automatic screening tools for coronavirus disease 2019 (COVID-19) based on chest radiography (CXR), with reported accuracies often well over 90%. However, it has been noted that many of these studies have likely suffered from dataset bias, leading to overly optimistic results. The purpose of this study was to thoroughly investigate to what extent biases have influenced the performance of a range of previously proposed and promising convolutional neural networks (CNNs), and to determine what performance can be expected with current CNNs on a realistic and unbiased dataset. METHODS: Five CNNs for COVID-19 positive/negative classification were implemented for evaluation, namely VGG19, ResNet50, InceptionV3, DenseNet201, and COVID-Net. To perform both internal and cross-dataset evaluations, four datasets were created. The first dataset Valencian Region Medical Image Bank (BIMCV) followed strict reverse transcriptase-polymerase chain reaction (RT-PCR) test criteria and was created from a single reliable open access databank, while the second dataset (COVIDxB8) was created through a combination of six online CXR repositories. The third and fourth datasets were created by combining the opposing classes from the BIMCV and COVIDxB8 datasets. To decrease inter-dataset variability, a pre-processing workflow of resizing, normalization, and histogram equalization were applied to all datasets. Classification performance was evaluated on unseen test sets using precision and recall. A qualitative sanity check was performed by evaluating saliency maps displaying the top 5%, 10%, and 20% most salient segments in the input CXRs, to evaluate whether the CNNs were using relevant information for decision making. In an additional experiment and to further investigate the origin of potential dataset bias, all pixel values outside the lungs were set to zero through automatic lung segmentation before training and testing. RESULTS: When trained and evaluated on the single online source dataset (BIMCV), the performance of all CNNs is relatively low (precision: 0.65-0.72, recall: 0.59-0.71), but remains relatively consistent during external evaluation (precision: 0.58-0.82, recall: 0.57-0.72). On the contrary, when trained and internally evaluated on the combinatory datasets, all CNNs performed well across all metrics (precision: 0.94-1.00, recall: 0.77-1.00). However, when subsequently evaluated cross-dataset, results dropped substantially (precision: 0.10-0.61, recall: 0.04-0.80). For all datasets, saliency maps revealed the CNNs rarely focus on areas inside the lungs for their decision-making. However, even when setting all pixel values outside the lungs to zero, classification performance does not change and dataset bias remains. CONCLUSIONS: Results in this study confirm that when trained on a combinatory dataset, CNNs tend to learn the origin of the CXRs rather than the presence or absence of disease, a behavior known as short-cut learning. The bias is shown to originate from differences in overall pixel values rather than embedded text or symbols, despite consistent image pre-processing. When trained on a reliable, and realistic single-source dataset in which non-lung pixels have been masked, CNNs currently show limited sensitivity (<70%) for COVID-19 infection in CXR, questioning their use as a reliable automatic screening tool.

Considerations for shoot-through FLASH proton therapy.

PURPOSE: To discuss several pertinent issues related to shoot-through FLASH proton therapy based on an illustrative case. METHODS: We argue that with the advent of FLASH proton radiotherapy and due to the issues associated with conventional proton radiotherapy regarding the uncertainties of positioning of the Bragg peaks, the difficulties of in vivo verification of the dose distribution, the use of treatment margins and the uncertainties surrounding linear energy transfer (LET) and relative biological effectiveness (RBE), a special mode of shoot-through FLASH proton radiotherapy should be investigated. In shoot-through FLASH, the proton beams have sufficient energy to reach the distal exit side of the patient. Due to the FLASH sparing effect of normal tissues at both the proximal and distal side of tumors, radiotherapy plans can be developed that meet current planning constraints and issues regarding RBE can be avoided. RESULTS: A preliminary proton plan for a neurological tumor in close proximity to various organs at risk (OAR) with strict dose constraints was studied. A plan with four beams mostly met the constraints for the OAR, using a treatment planning system that was not optimized for this novel treatment modality. When new treatment planning algorithms would be developed for shoot-through FLASH, constraints would be easier to meet. The shoot-through FLASH plan led to a significant effective dose reduction in large parts of the healthy tissue. The plan had no uncertainties associated to Bragg peak positioning, needed in principle no large proximal or distal margins and LET increases near the Bragg peak became irrelevant. CONCLUSION: Shoot-through FLASH proton radiotherapy may be an interesting treatment modality to explore further. It would remove some of the current sources of uncertainty in proton radiotherapy. An additional advantage could be that portal dosimetry may be possible with beams penetrating the patient and impinging on a distally placed imaging detector, potentially leading to a practical treatment verification method. With current proton accelerator technology, trials could be conducted for neurological, head&neck and thoracic cancers. For abdominal and pelvic cancer a higher proton energy would be required.

Identification of treatment error types for lung cancer patients using convolutional neural networks and EPID dosimetry.

BACKGROUND/PURPOSE: Electronic portal imaging device (EPID) dosimetry aims to detect treatment errors, potentially leading to treatment adaptation. Clinically used threshold classification methods for detecting errors lead to loss of information (from multi-dimensional EPID data to a few numbers) and cannot be used for identifying causes of errors. Advanced classification methods, such as deep learning, can use all available information. In this study, convolutional neural networks (CNNs) were trained to detect and identify error type and magnitude of simulated treatment errors in lung cancer patients. The purpose of this simulation study is to provide a proof-of-concept of CNNs for error identification using EPID dosimetry in an in vivo scenario. MATERIALS AND METHODS: Clinically realistic ranges of anatomical changes, positioning errors and mechanical errors were simulated for lung cancer patients. Predicted portal dose images (PDIs) containing errors were compared to error-free PDIs using the widely used gamma analysis. CNNs were trained to classify errors using 2D gamma maps. Three classification levels were assessed: Level 1 (main error type, e.g., anatomical change), Level 2 (error subtype, e.g., tumor regression) and Level 3 (error magnitude, e.g., >50% tumor regression). RESULTS: CNNs showed good performance for all classification levels (training/test accuracy 99.5%/96.1%, 92.5%/86.8%, 82.0%/72.9%). For Level 3, overfitting became more apparent. CONCLUSION: This simulation study indicates that deep learning is a promising powerful tool for identifying types and magnitude of treatment errors with EPID dosimetry, providing additional information not currently available from EPID dosimetry. This is a first step towards rapid, automated models for identification of treatment errors using EPID dosimetry.

External validation of a hidden Markov model for gamma-based classification of anatomical changes in lung cancer patients using EPID dosimetry.

PURPOSE: To externally validate a hidden Markov model (HMM) for classifying gamma analysis results of in vivo electronic portal imaging device (EPID) measurements into different categories of anatomical change for lung cancer patients. Additionally, the relationship between HMM classification and deviations in dose-volume histogram (DVH) metrics was evaluated. METHODS: The HMM was developed at CHU de Québec (CHUQ), and trained on features extracted from gamma analysis maps of in vivo EPID measurements from 483 fractions (24 patients, treated with three-dimensional 3D-CRT or intensity modulated radiotherapy), using the EPID measurement of the first treatment fraction as reference. The model inputs were the average gamma value, standard deviation, and average value of the highest 1% of gamma values, all averaged over all beams in a fraction. The HMM classified each fraction into one of three categories: no anatomical change (Category 1), some anatomical change (no clinical action needed, Category 2) and severe anatomical change (clinical action needed, Category 3). The external validation dataset consisted of EPID measurements from 263 fractions of 30 patients treated at Maastro with volumetric modulated arc therapy (VMAT) or hybrid plans (containing both static beams and VMAT arcs). Gamma analysis features were extracted in the same way as in the CHUQ dataset, by using the EPID measurement of the first fraction as reference (γQ), and additionally by using an EPID dose prediction as reference (γM). For Maastro patients, cone beam computed tomography (CBCT) scans and image-guided radiotherapy (IGRT) classification of these images were available for each fraction. Contours were propagated from the planning CT to the CBCTs, and the dose was recalculated using a Monte Carlo dose engine. Dose-volume histogram metrics for targets and organs-at-risk (OARs: lungs, heart, mediastinum, spinal cord, brachial plexus) were extracted for each fraction, and compared to the planned dose. HMM classification of the external validation set was compared to threshold classification based on the average gamma value alone (a surrogate for clinical classification at CHUQ), IGRT classification as performed at Maastro, and differences in DVH metrics extracted from 3D dose recalculations on the CBCTs. RESULTS: The HMM achieved 65.4%/65.0% accuracy for γQ and γM, respectively, compared to average gamma threshold classification. When comparing HMM classification with IGRT classification, the overall accuracy was 29.7% for γQ and 23.2% for γM. Hence, HMM classification and IGRT classification of anatomical changes did not correspond. However, there is a trend towards higher deviations in DVH metrics with classification into higher categories by the HMM for large OARs (lungs, heart, mediastinum), but not for the targets and small OARs (spinal cord, brachial plexus). CONCLUSION: The external validation shows that transferring the HMM for anatomical change classification to a different center is challenging, but can still be valuable. The HMM trained at CHUQ cannot be used directly to classify anatomical changes in the Maastro data. However, it may be possible to use the model in a different capacity, as an indicator for changes in the 3D dose based on two-dimensional EPID measurements.

Treatment plan quality assessment for radiotherapy of rectal cancer patients using prediction of organ-at-risk dose metrics.

BACKGROUND AND PURPOSE: Radiotherapy centers frequently lack simple tools for periodic treatment plan verification and feedback on current plan quality. It is difficult to measure treatment quality over different years or during the planning process. Here, we implemented plan quality assurance (QA) by developing a database of dose-volume histogram (DVH) metrics and a prediction model. These tools were used to assess automatically optimized treatment plans for rectal cancer patients, based on cohort analysis. MATERIAL AND METHODS: A treatment plan QA framework was established and an overlap volume histogram based model was used to predict DVH parameters for cohorts of patients treated in 2018 and 2019 and grouped according to planning technique. A training cohort of 22 re-optimized treatment plans was used to make the prediction model. The prediction model was validated on 95 automatically generated treatment plans (automatically optimized cohort) and 93 manually optimized plans (manually optimized cohort). RESULTS: For the manually optimized cohort, on average the prediction deviated less than 0.3 ± 1.4 Gy and -4.3 ± 5.5 Gy, for the mean doses to the bowel bag and bladder, respectively; for the automatically optimized cohort a smaller deviation was observed: -0.1 ± 1.1 Gy and -0.2 ± 2.5 Gy, respectively. The interquartile range of DVH parameters was on average smaller for the automatically optimized cohort, indicating less variation within each parameter compared to manual planning. CONCLUSION: An automated framework to monitor treatment quality with a DVH prediction model was successfully implemented clinically and revealed less variation in DVH parameters for automated in comparison to manually optimized plans. The framework also allowed for individual feedback and DVH estimation.

Should dose from small fields be limited for dose verification procedures?: uncertainty versus small field dose in VMAT treatments.

Over the years, radiotherapy treatments have become more complex and conformal, leading to an increased use of small field segments in volumetric modulated arc therapy (VMAT) arcs. The impact of small field dose inaccuracy on dose verification methods has not been studied yet. The aim of this work is therefore to quantify the relationship between the uncertainty of a 2D pre-treatment dose prediction model and the proportion of dose coming from small fields in VMAT arcs for a range of clinical plans. The model evaluated in this work predicts 2D portal dose images (PDIs) without a patient or phantom in the beam. The uncertainty of the model was calculated through simulation of model parameter deviations. The proportion of dose from small fields in a VMAT arc was determined by comparing a PDI with only dose from small fields with the original PDI. The uncertainty and proportion of dose from small fields were calculated for 109 VMAT arcs (41 head and neck, 33 lung, 35 prostate). The correlation was assessed with a linear regression. There is a statistically significant positive correlation between the uncertainty of the model and the proportion of dose from small fields in a VMAT arc, for each treatment site individually, as well as for all tumor sites together. The strongest relationship is found for the prostate cases. As there is a positive relationship between the uncertainty of the 2D pre-treatment dose prediction model, it may be wise to limit the dose from small fields in VMAT arcs, to avoid additional uncertainty in the dose verification process.

Validation and uncertainty analysis of a pre-treatment 2D dose prediction model.

Independent verification of complex treatment delivery with megavolt photon beam radiotherapy (RT) has been effectively used to detect and prevent errors. This work presents the validation and uncertainty analysis of a model that predicts 2D portal dose images (PDIs) without a patient or phantom in the beam. The prediction model is based on an exponential point dose model with separable primary and secondary photon fluence components. The model includes a scatter kernel, off-axis ratio map, transmission values and penumbra kernels for beam-delimiting components. These parameters were derived through a model fitting procedure supplied with point dose and dose profile measurements of radiation fields. The model was validated against a treatment planning system (TPS; Eclipse) and radiochromic film measurements for complex clinical scenarios, including volumetric modulated arc therapy (VMAT). Confidence limits on fitted model parameters were calculated based on simulated measurements. A sensitivity analysis was performed to evaluate the effect of the parameter uncertainties on the model output. For the maximum uncertainty, the maximum deviating measurement sets were propagated through the fitting procedure and the model. The overall uncertainty was assessed using all simulated measurements. The validation of the prediction model against the TPS and the film showed a good agreement, with on average 90.8% and 90.5% of pixels passing a (2%,2 mm) global gamma analysis respectively, with a low dose threshold of 10%. The maximum and overall uncertainty of the model is dependent on the type of clinical plan used as input. The results can be used to study the robustness of the model. A model for predicting accurate 2D pre-treatment PDIs in complex RT scenarios can be used clinically and its uncertainties can be taken into account.

Detection of anatomical changes in lung cancer patients with 2D time-integrated, 2D time-resolved and 3D time-integrated portal dosimetry: a simulation study.

The aim of this work is to assess the performance of 2D time-integrated (2D-TI), 2D time-resolved (2D-TR) and 3D time-integrated (3D-TI) portal dosimetry in detecting dose discrepancies between the planned and (simulated) delivered dose caused by simulated changes in the anatomy of lung cancer patients. For six lung cancer patients, tumor shift, tumor regression and pleural effusion are simulated by modifying their CT images. Based on the modified CT images, time-integrated (TI) and time-resolved (TR) portal dose images (PDIs) are simulated and 3D-TI doses are calculated. The modified and original PDIs and 3D doses are compared by a gamma analysis with various gamma criteria. Furthermore, the difference in the D 95% (ΔD 95%) of the GTV is calculated and used as a gold standard. The correlation between the gamma fail rate and the ΔD 95% is investigated, as well the sensitivity and specificity of all combinations of portal dosimetry method, gamma criteria and gamma fail rate threshold. On the individual patient level, there is a correlation between the gamma fail rate and the ΔD 95%, which cannot be found at the group level. The sensitivity and specificity analysis showed that there is not one combination of portal dosimetry method, gamma criteria and gamma fail rate threshold that can detect all simulated anatomical changes. This work shows that it will be more beneficial to relate portal dosimetry and DVH analysis on the patient level, rather than trying to quantify a relationship for a group of patients. With regards to optimizing sensitivity and specificity, different combinations of portal dosimetry method, gamma criteria and gamma fail rate should be used to optimally detect certain types of anatomical changes.

The bilirubin albumin ratio in the management of hyperbilirubinemia in preterm infants to improve neurodevelopmental outcome: a randomized controlled trial--BARTrial.

BACKGROUND AND OBJECTIVE: High bilirubin/albumin (B/A) ratios increase the risk of bilirubin neurotoxicity. The B/A ratio may be a valuable measure, in addition to the total serum bilirubin (TSB), in the management of hyperbilirubinemia. We aimed to assess whether the additional use of B/A ratios in the management of hyperbilirubinemia in preterm infants improved neurodevelopmental outcome. METHODS: In a prospective, randomized controlled trial, 615 preterm infants of 32 weeks' gestation or less were randomly assigned to treatment based on either B/A ratio and TSB thresholds (consensus-based), whichever threshold was crossed first, or on the TSB thresholds only. The primary outcome was neurodevelopment at 18 to 24 months' corrected age as assessed with the Bayley Scales of Infant Development III by investigators unaware of treatment allocation. Secondary outcomes included complications of preterm birth and death. RESULTS: Composite motor (100 ± 13 vs. 101 ± 12) and cognitive (101 ± 12 vs. 101 ± 11) scores did not differ between the B/A ratio and TSB groups. Demographic characteristics, maximal TSB levels, B/A ratios, and other secondary outcomes were similar. The rates of death and/or severe neurodevelopmental impairment for the B/A ratio versus TSB groups were 15.4% versus 15.5% (P = 1.0) and 2.8% versus 1.4% (P = 0.62) for birth weights ≤ 1000 g and 1.8% versus 5.8% (P = 0.03) and 4.1% versus 2.0% (P = 0.26) for birth weights of >1000 g. CONCLUSIONS: The additional use of B/A ratio in the management of hyperbilirubinemia in preterm infants did not improve their neurodevelopmental outcome. TRIAL REGISTRATION: Controlled-Trials.com ISRCTN74465643.