Impact of proton PBS machine operating parameters on the effectiveness of layer rescanning for interplay effect mitigation in lung SBRT treatment.

BACKGROUND: Rescanning is a common technique used in proton pencil beam scanning to mitigate the interplay effect. Advances in machine operating parameters across different generations of particle therapy systems have led to improvements in beam delivery time (BDT). However, the potential impact of these improvements on the effectiveness of rescanning remains an underexplored area in the existing research. METHODS: We systematically investigated the impact of proton machine operating parameters on the effectiveness of layer rescanning in mitigating interplay effect during lung SBRT treatment, using the CIRS phantom. Focused on the Hitachi synchrotron particle therapy system, we explored machine operating parameters from our institution's current (2015) and upcoming systems (2025A and 2025B). Accumulated dynamic 4D dose were reconstructed to assess the interplay effect and layer rescanning effectiveness. RESULTS: Achieving target coverage and dose homogeneity within 2% deviation required 6, 6, and 20 times layer rescanning for the 2015, 2025A, and 2025B machine parameters, respectively. Beyond this point, further increasing the number of layer rescanning did not further improve the dose distribution. BDTs without rescanning were 50.4, 24.4, and 11.4 s for 2015, 2025A, and 2025B, respectively. However, after incorporating proper number of layer rescanning (six for 2015 and 2025A, 20 for 2025B), BDTs increased to 67.0, 39.6, and 42.3 s for 2015, 2025A, and 2025B machine parameters. Our data also demonstrated the potential problem of false negative and false positive if the randomness of the respiratory phase at which the beam is initiated is not considered in the evaluation of interplay effect. CONCLUSION: The effectiveness of layer rescanning for mitigating interplay effect is affected by machine operating parameters. Therefore, past clinical experiences may not be applicable to modern machines.

Direct tumor visual feedback during free breathing in 0.35T MRgRT.

To present a tumor motion control system during free breathing using direct tumor visual feedback to patients in 0.35 T magnetic resonance-guided radiotherapy (MRgRT). We present direct tumor visualization to patients by projecting real-time cine MR images on an MR-compatible display system inside a 0.35 T MRgRT bore. The direct tumor visualization included anatomical images with a target contour and an auto-segmented gating contour. In addition, a beam-status sign was added for patient guidance. The feasibility was investigated with a six-patient clinical evaluation of the system in terms of tumor motion range and beam-on time. Seven patients without visual guidance were used for comparison. Positions of the tumor and the auto-segmented gating contour from the cine MR images were used in probability analysis to evaluate tumor motion control. In addition, beam-on time was recorded to assess the efficacy of the visual feedback system. The direct tumor visualization system was developed and implemented in our clinic. The target contour extended 3 mm outside of the gating contour for 33.6 ± 24.9% of the time without visual guidance, and 37.2 ± 26.4% of the time with visual guidance. The average maximum motion outside of the gating contour was 14.4 ± 11.1 mm without and 13.0 ± 7.9 mm with visual guidance. Beam-on time as a percentage was 43.9 ± 15.3% without visual guidance, and 48.0 ± 21.2% with visual guidance, but was not significantly different (P = 0.34). We demonstrated the clinical feasibility and potential benefits of presenting direct tumor visual feedback to patients in MRgRT. The visual feedback allows patients to visualize and attempt to minimize tumor motion in free breathing. The proposed system and associated clinical workflow can be easily adapted for any type of MRgRT.

Image-domain shading correction for cone-beam CT without prior patient information.

In the era of high-precision radiotherapy, cone-beam CT (CBCT) is frequently utilized for on-board treatment guidance. However, CBCT images usually contain severe shading artifacts due to strong photon scatter from illumination of a large volume and non-optimized patient-specific data measurements, limiting the full clinical applications of CBCT. Many algorithms have been proposed to alleviate this problem by data correction on projections. Sophisticated methods have also been designed when prior patient information is available. Nevertheless, a standard, efficient, and effective approach with large applicability remains elusive for current clinical practice. In this work, we develop a novel algorithm for shading correction directly on CBCT images. Distinct from other image-domain correction methods, our approach does not rely on prior patient information or prior assumption of patient data. In CBCT, projection errors (mostly from scatter and non-ideal usage of bowtie filter) result in dominant low-frequency shading artifacts in image domain. In circular scan geometry, these artifacts often show global or local radial patterns. Hence, the raw CBCT images are first preprocessed into the polar coordinate system. Median filtering and polynomial fitting are applied on the transformed image to estimate the low-frequency shading artifacts (referred to as the bias field) angle-by-angle and slice-by-slice. The low-pass filtering process is done firstly along the angular direction and then the radial direction to preserve image contrast. The estimated bias field is then converted back to the Cartesian coordinate system, followed by 3D low-pass filtering to eliminate possible high-frequency components. The shading-corrected image is finally obtained as the uncorrected volume divided by the bias field. The proposed algorithm was evaluated on CBCT images of a pelvis patient and a head patient. Mean CT number values and spatial non-uniformity on the reconstructed images were used as image quality metrics. Within selected regions of interest, the average CT number error was reduced from around 300 HU to 42 and 38 HU, and the spatial nonuniformity error was reduced from above 17.5% to 2.1% and 1.7% for the pelvis and the head patients, respectively. As our method suppresses only low-frequency shading artifacts, patient anatomy and contrast were retained in the corrected images for both cases. Our shading correction algorithm on CBCT images offers several advantages. It has a high efficiency, since it is deterministic and directly operates on the reconstructed images. It requires no prior information or assumptions, which not only achieves the merits of CBCT-based treatment monitoring by retaining the patient anatomy, but also facilitates its clinical use as an efficient image-correction solution.

Enhanced IDOL segmentation framework using personalized hyperspace learning IDOL.

BACKGROUND: Adaptive radiotherapy (ART) workflows have been increasingly adopted to achieve dose escalation and tissue sparing under shifting anatomic conditions, but the necessity of recontouring and the associated time burden hinders a real-time or online ART workflow. In response to this challenge, approaches to auto-segmentation involving deformable image registration, atlas-based segmentation, and deep learning-based segmentation (DLS) have been developed. Despite the particular promise shown by DLS methods, implementing these approaches in a clinical setting remains a challenge, namely due to the difficulty of curating a data set of sufficient size and quality so as to achieve generalizability in a trained model. PURPOSE: To address this challenge, we have developed an intentional deep overfit learning (IDOL) framework tailored to the auto-segmentation task. However, certain limitations were identified, particularly the insufficiency of the personalized dataset to effectively overfit the model. In this study, we introduce a personalized hyperspace learning (PHL)-IDOL segmentation framework capable of generating datasets that induce the model to overfit specific patient characteristics for medical image segmentation. METHODS: The PHL-IDOL model is trained in two stages. In the first, a conventional, general model is trained with a diverse set of patient data (n = 100 patients) consisting of CT images and clinical contours. Following this, the general model is tuned with a data set consisting of two components: (a) selection of a subset of the patient data (m < n) using the similarity metrics (mean square error (MSE), peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and the universal quality image index (UQI) values); (b) adjust the CT and the clinical contours using a deformed vector generated from the reference patient and the selected patients using (a). After training, the general model, the continual model, the conventional IDOL model, and the proposed PHL-IDOL model were evaluated using the volumetric dice similarity coefficient (VDSC) and the Hausdorff distance 95% (HD95%) computed for 18 structures in 20 test patients. RESULTS: Implementing the PHL-IDOL framework resulted in improved segmentation performance for each patient. The Dice scores increased from 0.81 ± $ \pm $ 0.05 with the general model, 0.83 ± 0.04 $ \pm 0.04$ for the continual model, 0.83 ± 0.04 $ \pm 0.04$ for the conventional IDOL model to an average of 0.87 ± 0.03 $ \pm 0.03$ with the PHL-IDOL model. Similarly, the Hausdorff distance decreased from 3.06 ± 0.99 $ \pm 0.99$ with the general model, 2.84 ± 0.69 $ \pm 0.69$ for the continual model, 2.79 ± 0.79 $ \pm 0.79$ for the conventional IDOL model and 2.36 ± 0.52 $ \pm 0.52$ for the PHL-IDOL model. All the standard deviations were decreased by nearly half of the values comparing the general model and the PHL-IDOL model. CONCLUSION: The PHL-IDOL framework applied to the auto-segmentation task achieves improved performance compared to the general DLS approach, demonstrating the promise of leveraging patient-specific prior information in a task central to online ART workflows.

AAPM Task Group 334: A guidance document to using radiotherapy immobilization devices and accessories in an MR environment.

Use of magnetic resonance (MR) imaging in radiation therapy has increased substantially in recent years as more radiotherapy centers are having MR simulators installed, requesting more time on clinical diagnostic MR systems, or even treating with combination MR linear accelerator (MR-linac) systems. With this increased use, to ensure the most accurate integration of images into radiotherapy (RT), RT immobilization devices and accessories must be able to be used safely in the MR environment and produce minimal perturbations. The determination of the safety profile and considerations often falls to the medical physicist or other support staff members who at a minimum should be a Level 2 personnel as per the ACR. The purpose of this guidance document will be to help guide the user in making determinations on MR Safety labeling (i.e., MR Safe, Conditional, or Unsafe) including standard testing, and verification of image quality, when using RT immobilization devices and accessories in an MR environment.

Ensemble learning and personalized training for the improvement of unsupervised deep learning-based synthetic CT reconstruction.

BACKGROUND: The growing adoption of magnetic resonance imaging (MRI)-guided radiation therapy (RT) platforms and a focus on MRI-only RT workflows have brought the technical challenge of synthetic computed tomography (sCT) reconstruction to the forefront. Unpaired-data deep learning-based approaches to the problem offer the attractive characteristic of not requiring paired training data, but the gap between paired- and unpaired-data results can be limiting. PURPOSE: We present two distinct approaches aimed at improving unpaired-data sCT reconstruction results: a cascade ensemble that combines multiple models and a personalized training strategy originally designed for the paired-data setting. METHODS: Comparisons are made between the following models: (1) the paired-data fully convolutional DenseNet (FCDN), (2) the FCDN with the Intentional Deep Overfit Learning (IDOL) personalized training strategy, (3) the unpaired-data CycleGAN, (4) the CycleGAN with the IDOL training strategy, and (5) the CycleGAN as an intermediate model in a cascade ensemble approach. Evaluation of the various models over 25 total patients is carried out using a five-fold cross-validation scheme, with the patient-specific IDOL models being trained for the five patients of fold 3, chosen at random. RESULTS: In both the paired- and unpaired-data settings, adopting the IDOL training strategy led to improvements in the mean absolute error (MAE) between true CT images and sCT outputs within the body contour (mean improvement, paired- and unpaired-data approaches, respectively: 38%, 9%) and in regions of bone (52%, 5%), the peak signal-to-noise ratio (PSNR; 15%, 7%), and the structural similarity index (SSIM; 6%, <1%). The ensemble approach offered additional benefits over the IDOL approach in all three metrics (mean improvement over unpaired-data approach in fold 3; MAE: 20%; bone MAE: 16%; PSNR: 10%; SSIM: 2%), and differences in body MAE between the ensemble approach and the paired-data approach are statistically insignificant. CONCLUSIONS: We have demonstrated that both a cascade ensemble approach and a personalized training strategy designed initially for the paired-data setting offer significant improvements in image quality metrics for the unpaired-data sCT reconstruction task. Closing the gap between paired- and unpaired-data approaches is a step toward fully enabling these powerful and attractive unpaired-data frameworks.

Technical note: Progressive deep learning: An accelerated training strategy for medical image segmentation.

BACKGROUND: Recent advancements in Deep Learning (DL) methodologies have led to state-of-the-art performance in a wide range of applications especially in object recognition, classification, and segmentation of medical images. However, training modern DL models requires a large amount of computation and long training times due to the complex nature of network structures and the large number of training datasets involved. Moreover, it is an intensive, repetitive manual process to select the optimized configuration of hyperparameters for a given DL network. PURPOSE: In this study, we present a novel approach to accelerate the training time of DL models via the progressive feeding of training datasets based on similarity measures for medical image segmentation. We term this approach Progressive Deep Learning (PDL). METHODS: The two-stage PDL approach was tested on the auto-segmentation task for two imaging modalities: CT and MRI. The training datasets were ranked according to similarity measures between each sample based on Mean Square Error (MSE), Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), and the Universal Quality Image Index (UQI) values. At the start of the training process, a relatively coarse sampling of training datasets with higher ranks was used to optimize the hyperparameters of the DL network. Following this, the samples with higher ranks were used in step 1 to yield accelerated loss minimization in early training epochs and the total dataset was added in step 2 for the remainder of training. RESULTS: Our results demonstrate that the PDL approach can reduce the training time by nearly half (∼49%) and can predict segmentations (CT U-net/DenseNet dice coefficient: 0.9506/0.9508, MR U-net/DenseNet dice coefficient: 0.9508/0.9510) without major statistical difference (Wilcoxon signed-rank test) compared to the conventional DL approach. The total training times with a fixed cutoff at 0.95 DSC for the CT dataset using DenseNet and U-Net architectures, respectively, were 17 h, 20 min and 4 h, 45 min in the conventional case compared to 8 h, 45 min and 2 h, 20 min with PDL. For the MRI dataset, the total training times using the same architectures were 2 h, 54 min and 52 min in the conventional case and 1 h, 14 min and 25 min with PDL. CONCLUSION: The proposed PDL training approach offers the ability to substantially reduce the training time for medical image segmentation while maintaining the performance achieved in the conventional case.

A high-resolution cone beam computed tomography (HRCBCT) reconstruction framework for CBCT-guided online adaptive therapy.

BACKGROUND: Kilo-voltage cone-beam computed tomography (CBCT) is a prevalent modality used for adaptive radiotherapy (ART) due to its compatibility with linear accelerators and ability to provide online imaging. However, the widely-used Feldkamp-Davis-Kress (FDK) reconstruction algorithm has several limitations, including potential streak aliasing artifacts and elevated noise levels. Iterative reconstruction (IR) techniques, such as total variation (TV) minimization, dictionary-based methods, and prior information-based methods, have emerged as viable solutions to address these limitations and improve the quality and applicability of CBCT in ART. PURPOSE: One of the primary challenges in IR-based techniques is finding the right balance between minimizing image noise and preserving image resolution. To overcome this challenge, we have developed a new reconstruction technique called high-resolution CBCT (HRCBCT) that specifically focuses on improving image resolution while reducing noise levels. METHODS: The HRCBCT reconstruction technique builds upon the conventional IR approach, incorporating three components: the data fidelity term, the resolution preservation term, and the regularization term. The data fidelity term ensures alignment between reconstructed values and measured projection data, while the resolution preservation term exploits the high resolution of the initial Feldkamp-Davis-Kress (FDK) algorithm. The regularization term mitigates noise during the IR process. To enhance convergence and resolution at each iterative stage, we applied Iterative Filtered Backprojection (IFBP) to the data fidelity minimization process. RESULTS: We evaluated the performance of the proposed HRCBCT algorithm using data from two physical phantoms and one head and neck patient. The HRCBCT algorithm outperformed all four different algorithms; FDK, Iterative Filtered Back Projection (IFBP), Compressed Sensing based Iterative Reconstruction (CSIR), and Prior Image Constrained Compressed Sensing (PICCS) methods in terms of resolution and noise reduction for all data sets. Line profiles across three line pairs of resolution revealed that the HRCBCT algorithm delivered the highest distinguishable line pairs compared to the other algorithms. Similarly, the Modulation Transfer Function (MTF) measurements, obtained from the tungsten wire insert on the CatPhan 600 physical phantom, showed a significant improvement with HRCBCT over traditional algorithms. CONCLUSION: The proposed HRCBCT algorithm offers a promising solution for enhancing CBCT image quality in adaptive radiotherapy settings. By addressing the challenges inherent in traditional IR methods, the algorithm delivers high-definition CBCT images with improved resolution and reduced noise throughout each iterative step. Implementing the HR CBCT algorithm could significantly impact the accuracy of treatment planning during online adaptive therapy.

GateCT™ surface tracking system for respiratory signal reconstruction in 4DCT imaging.

PURPOSE: To assess the temporal and spatial accuracy of the GateCT™ system (VisionRT, London, UK), a recently released respiratory tracking system for 4DCT, under both ideal and nonideal respiratory conditions. METHODS: Three experiments were performed by benchmarking and comparing its results with the ground-truth input data and those generated by the widely used Varian RPM™ system (Real-time Position Management, Varian, Palo Alto, CA). The first experiment used 10 sinusoidal breathing patterns (constant amplitude and frequency using sin(6)ωt), 10 "consistent" patient breathing patterns, and 10 "sporadic" patient breathing patterns. Motion was simulated with the quasar™ Programmable Respiratory Motion Platform (MODUS, London, Canada) as the surrogate. The GateCT™ and RPM™ systems were used to track the breathing patterns. The data from both systems were then analyzed in the Fourier domain, to evaluate temporal/phase accuracy, using the Pearson's correlation coefficient (PCC). The analysis correlated the ground-truth input data against the GateCT™ and RPM™ tracking results, respectively. The second experiment used 10 ideal sinusoidal breathing patterns, five of period 2.0 s, and five of period 5.0 s, with varying abdominal amplitudes found in clinical cases (peak-to-peak range: 1.67-10 mm) to test the sensitivity of the system to reconstruct various range of motion. And, the third experiment used 12 consecutive clinical patients to track the abdominal motion simultaneously by the GateCT™ and RPM™ systems. The baseline of the tracking results from both the two systems was analyzed via the mean-position-estimate (MPE) calculations. All experiments were tracked for at least 120 s. RESULTS: In the first experiment, the average PCC values (±SD) of all thirty breathing patterns were 0.9995 ± 0.00035 and 0.9994 ± 0.00041 for the GateCT™ and the RPM™ system, respectively. These nearly identical results demonstrated similar temporal/phase tracking accuracy for the two systems. The results in the second experiment, however, revealed a pattern for the GateCT™ system in which the uncertainty of its mean-position tracking increased as the amplitude of the breathing pattern decreased. For example, a non-negligible baseline drift of up to 29.3% with respect to the peak-to-peak amplitude of 1.67-mm was observed. On the contrary, the RPM™ system displayed a more consistent recording of amplitudes over time with the greatest drift being <7.7%. The third experiment confirmed these findings in the clinical setting. Consistent decrease in PCC values due to the increase in artificial amplitude drifts, as the breathing amplitude decreased, was found. The lowest PCC value was 0.7239 for a patient with 1.57-mm peak-to-peak amplitude. CONCLUSIONS: The GateCT™ system revealed its consistency in temporal/phase tracking but had limitations in accurately tracking the absolute abdominal positions, thus suggesting its appropriateness for phase-sorting of 4DCT rather than amplitude-sorting. In contrast, the RPM™ system demonstrated stable respiratory signal tracking in all ranges and accurately both in phase and amplitude, and is a robust system to use for both phase-sorting and amplitude-sorting techniques. The impact of the observed mean-position drift in the GateCT™ system on the resulting 4DCT image quality, in amplitude-sorting, needs further investigation.

Fast compressed sensing-based CBCT reconstruction using Barzilai-Borwein formulation for application to on-line IGRT.

PURPOSE: Compressed sensing theory has enabled an accurate, low-dose cone-beam computed tomography (CBCT) reconstruction using a minimal number of noisy projections. However, the reconstruction time remains a significant challenge for practical implementation in the clinic. In this work, we propose a novel gradient projection algorithm, based on the Gradient-Projection-Barzilai-Borwein formulation (GP-BB), that handles the total variation (TV)-norm regularization-based least squares problem for the CBCT reconstruction in a highly efficient manner, with speed acceptable for routine use in the clinic. METHODS: CBCT is reconstructed by minimizing an energy function consisting of a data fidelity term and a TV-norm regularization term. Both terms are simultaneously minimized by calculating the gradient projection of the energy function with the step size determined using an approximate Hessian calculation at each iteration, based on the Barzilai-Borwein formulation. To speed up the process, a multiresolution optimization is used. In addition, the entire algorithm was designed to run with a single graphics processing unit (GPU) card. To evaluate the performance, the Shepp-Logan numerical phantom, the CatPhan 600 physical phantom, and a clinically-treated head-and-neck patient were acquired from the TrueBeam™ system (Varian Medical Systems, Palo Alto, CA). For each scan, in total, 364 projections were acquired in a 200° rotation. The imager has 1024 × 768 pixels with 0.388 × 0.388-mm resolution. This was down-sampled to 512 × 384 pixels with 0.776 × 0.776-mm resolution for reconstruction. Evenly spaced angles were subsampled and used for varying the number of projections for the image reconstruction. To assess the performance of our GP-BB algorithm, we have implemented and compared with three compressed sensing-type algorithms, the two of which are popular and published (forward-backward splitting techniques), and the other one with a basic line-search technique. In addition, the conventional Feldkamp-Davis-Kress (FDK) reconstruction of the clinical patient data is compared as well. RESULTS: In comparison with the other compressed sensing-type algorithms, our algorithm showed convergence in ≤30 iterations whereas other published algorithms need at least 50 iterations in order to reconstruct the Shepp-Logan phantom image. With the CatPhan phantom, the GP-BB algorithm achieved a clinically-reasonable image with 40 projections in 12 iterations, in less than 12.6 s. This is at least an order of magnitude faster in reconstruction time compared with the most recent reports utilizing GPU technology given the same input projections. For the head-and-neck clinical scan, clinically-reasonable images were obtained from 120 projections in 34-78 s converging in 12-30 iterations. In this reconstruction range (i.e., 120 projections) the image quality is visually similar to or better than the conventional FDK reconstructed images using 364 projections. This represents a dose reduction of nearly 67% (120∕364 projections) while maintaining a reasonable speed in clinical implementation. CONCLUSIONS: In this paper, we proposed a novel, fast, low-dose CBCT reconstruction algorithm using the Barzilai-Borwein step-size calculation. A clinically viable head-and-neck image can be obtained within ∼34-78 s while simultaneously cutting the dose by approximately 67%. This makes our GP-BB algorithm potentially useful in an on-line image-guided radiation therapy (IGRT).

Liver motion during cone beam computed tomography guided stereotactic body radiation therapy.

PURPOSE: Understanding motion characteristics of liver such as, interfractional and intrafractional motion variability, difference in motion within different locations in the organ, and their complex relationship with the breathing cycles are particularly important for image-guided liver SBRT. The purpose of this study was to investigate such motion characteristics based on fiducial markers tracked with the x-ray projections of the CBCT scans, taken immediately prior to the treatments. METHODS: Twenty liver SBRT patients were analyzed. Each patient had three fiducial markers (2 × 5-mm gold) percutaneously implanted around the gross tumor. The prescription ranged from 2 to 8 fractions per patient. The CBCT projections data for each fraction (∼650 projections∕scan), for each patient, were analyzed and the 2D positions of the markers were extracted using an in-house algorithm. In total, >55 000 x-ray projections were analyzed from 85 CBCT scans. From the 2D extracted positions, a 3D motion trajectory of the markers was constructed, from each CBCT scans, resulting in left-right (LR), anterior-posterior (AP), and cranio-caudal (CC) location information of the markers with >55 000 data points. The authors then analyzed the interfraction and intrafraction liver motion variability, within different locations in the organ, and as a function of the breathing cycle. The authors also compared the motion characteristics against the planning 4DCT and the RPM™ (Varian Medical Systems, Palo Alto, CA) breathing traces. Variations in the appropriate gating window (defined as the percent of the maximum range at which 50% of the marker positions are contained), between fractions were calculated as well. RESULTS: The range of motion for the 20 patients were 3.0 ± 2.0 mm, 5.1 ± 3.1 mm, and 17.9 ± 5.1 mm in the planning 4DCT, and 2.8 ± 1.6 mm, 5.3 ± 3.1 mm, and 16.5 ± 5.7 mm in the treatment CBCT, for LR, AP, and CC directions, respectively. The range of respiratory period was 3.9 ± 0.7 and 4.2 ± 0.8 s during the 4DCT simulation and the CBCT scans, respectively. The authors found that breathing-induced AP and CC motions are highly correlated. That is, all markers moved cranially also moved posteriorly and vice versa, irrespective of the location. The LR motion had a more variable relationship with the AP∕CC motions, and appeared random with respect to the location. That is, when the markers moved toward cranial-posterior direction, 58% of the markers moved to the patient-right, 22% of the markers moved to the patient-left, and 20% of the markers had minimal∕none motion. The absolute difference in the motion magnitude between the markers, in different locations within the liver, had a positive correlation with the absolute distance between the markers (R(2) = 0.69, linear-fit). The interfractional gating window varied significantly for some patients, with the largest having 29.4%-56.4% range between fractions. CONCLUSIONS: This study analyzed the liver motion characteristics of 20 patients undergoing SBRT. A large variation in motion was observed, interfractionally and intrafractionally, and that as the distance between the markers increased, the difference in the absolute range of motion also increased. This suggests that marker(s) in closest proximity to the target be used.

Real-time liver tumor localization via a single x-ray projection using deep graph neural network-assisted biomechanical modeling.

Objective.Real-time imaging is highly desirable in image-guided radiotherapy, as it provides instantaneous knowledge of patients' anatomy and motion during treatments and enables online treatment adaptation to achieve the highest tumor targeting accuracy. Due to extremely limited acquisition time, only one or few x-ray projections can be acquired for real-time imaging, which poses a substantial challenge to localize the tumor from the scarce projections. For liver radiotherapy, such a challenge is further exacerbated by the diminished contrast between the tumor and the surrounding normal liver tissues. Here, we propose a framework combining graph neural network-based deep learning and biomechanical modeling to track liver tumor in real-time from a single onboard x-ray projection.Approach.Liver tumor tracking is achieved in two steps. First, a deep learning network is developed to predict the liver surface deformation using image features learned from the x-ray projection. Second, the intra-liver deformation is estimated through biomechanical modeling, using the liver surface deformation as the boundary condition to solve tumor motion by finite element analysis. The accuracy of the proposed framework was evaluated using a dataset of 10 patients with liver cancer.Main results.The results show accurate liver surface registration from the graph neural network-based deep learning model, which translates into accurate, fiducial-less liver tumor localization after biomechanical modeling (<1.2 (±1.2) mm average localization error).Significance.The method demonstrates its potentiality towards intra-treatment and real-time 3D liver tumor monitoring and localization. It could be applied to facilitate 4D dose accumulation, multi-leaf collimator tracking and real-time plan adaptation. The method can be adapted to other anatomical sites as well.

Intentional deep overfit learning (IDOL): A novel deep learning strategy for adaptive radiation therapy.

PURPOSE: Applications of deep learning (DL) are essential to realizing an effective adaptive radiotherapy (ART) workflow. Despite the promise demonstrated by DL approaches in several critical ART tasks, there remain unsolved challenges to achieve satisfactory generalizability of a trained model in a clinical setting. Foremost among these is the difficulty of collecting a task-specific training dataset with high-quality, consistent annotations for supervised learning applications. In this study, we propose a tailored DL framework for patient-specific performance that leverages the behavior of a model intentionally overfitted to a patient-specific training dataset augmented from the prior information available in an ART workflow-an approach we term Intentional Deep Overfit Learning (IDOL). METHODS: Implementing the IDOL framework in any task in radiotherapy consists of two training stages: (1) training a generalized model with a diverse training dataset of N patients, just as in the conventional DL approach, and (2) intentionally overfitting this general model to a small training dataset-specific the patient of interest ( N + 1 ) generated through perturbations and augmentations of the available task- and patient-specific prior information to establish a personalized IDOL model. The IDOL framework itself is task-agnostic and is, thus, widely applicable to many components of the ART workflow, three of which we use as a proof of concept here: the autocontouring task on replanning CTs for traditional ART, the MRI super-resolution (SR) task for MRI-guided ART, and the synthetic CT (sCT) reconstruction task for MRI-only ART. RESULTS: In the replanning CT autocontouring task, the accuracy measured by the Dice similarity coefficient improves from 0.847 with the general model to 0.935 by adopting the IDOL model. In the case of MRI SR, the mean absolute error (MAE) is improved by 40% using the IDOL framework over the conventional model. Finally, in the sCT reconstruction task, the MAE is reduced from 68 to 22 HU by utilizing the IDOL framework. CONCLUSIONS: In this study, we propose a novel IDOL framework for ART and demonstrate its feasibility using three ART tasks. We expect the IDOL framework to be especially useful in creating personally tailored models in situations with limited availability of training data but existing prior information, which is usually true in the medical setting in general and is especially true in ART.

Four-dimensional cone-beam computed tomography and digital tomosynthesis reconstructions using respiratory signals extracted from transcutaneously inserted metal markers for liver SBRT.

PURPOSE: Respiration-induced intrafraction target motion is a concern in liver cancer radiotherapy, especially in stereotactic body radiotherapy (SBRT), and therefore, verification of its motion is necessary. An effective means to localize the liver cancer is to insert metal fiducial markers to or near the tumor with simultaneous imaging using cone-beam computed tomography (CBCT). Utilizing the fiducial markers, the authors have demonstrated a method to generate breath-induced motion signal of liver for reconstructing 4D digital tomosynthesis (4DDTS) and 4DCBCT images based on phasewise and/or amplitudewise sorting of projection data. METHODS: The marker extraction algorithm is based on template matching of a prior known marker image and has been coded to optimally extract marker positions in CBCT projections from the On-Board Imager (Varian Medical Systems, Palo Alto, CA). To validate the algorithm, multiple projection images of moving thorax phantom and five patient cases were examined. Upon extraction of the motion signals from the markers, 4D image sorting and image reconstructions were subsequently performed. In the case of incomplete signals due to projections with missing markers, the authors have implemented signal profiling to replace the missing portion. RESULTS: The proposed marker extraction algorithm was shown to be very robust and accurate in the phantom and patient cases examined. The maximum discrepancy of the algorithm predicted marker location versus operator selected location was < 1.2 mm, with the overall average of 0.51 +/- 0.15 mm, for 500 projections. The resulting 4DDTS and 4DCBCT images showed clear reduction in motion-induced blur of the markers and the anatomy for an effective image guidance. The signal profiling method was useful in replacing missing signals. CONCLUSIONS: The authors have successfully demonstrated that motion tracking of fiducial markers and the subsequent 4D reconstruction of CBCT and DTS are possible. Due to the significant reduction in motion-induced image blur, it is anticipated that such technology will be useful in image-guided liver SBRT treatments.

Evaluation of super-resolution on 50 pancreatic cancer patients with real-time cine MRI from 0.35T MRgRT.

MR-guided radiotherapy (MRgRT) systems provide excellent soft tissue imaging immediately prior to and in real time during radiation delivery for cancer treatment. However, 2D cine MRI often has limited spatial resolution due to high temporal resolution. This work applies a super resolution machine learning framework to 3.5 mm pixel edge length, low resolution (LR), sagittal 2D cine MRI images acquired on a MRgRT system to generate 0.9 mm pixel edge length, super resolution (SR), images originally acquired at 4 frames per second (FPS). LR images were collected from 50 pancreatic cancer patients treated on a ViewRay MR-LINAC. SR images were evaluated using three methods. 1) The first method utilized intrinsic image quality metrics for evaluation. 2) The second used relative metrics including edge detection and structural similarity index (SSIM). 3) Finally, automatically generated tumor contours were created on both low resolution and super resolution images to evaluate target delineation and compared with DICE and SSIM. Intrinsic image quality metrics all had statistically significant improvements for SR images versus LR images, with mean (±1 SD) BRISQUE scores of 29.65 ± 2.98 and 42.48 ± 0.98 for SR and LR, respectively. SR images showed good agreement with LR images in SSIM evaluation, indicating there was not significant distortion of the images. Comparison of LR and SR images with paired high resolution (HR) 3D images showed that SR images had a mean (±1 SD) SSIM value of 0.633 ± 0.063 and LR a value of 0.587 ± 0.067 (p ≪ 0.05). Contours generated on SR images were also more robust to noise addition than those generated on LR images. This study shows that super resolution with a machine learning framework can generate high spatial resolution images from 4fps low spatial resolution cine MRI acquired on the ViewRay MR-LINAC while maintaining tumor contour quality and without significant acquisition or post processing delay.

Abdominal synthetic CT reconstruction with intensity projection prior for MRI-only adaptive radiotherapy.

Objective. Owing to the superior soft tissue contrast of MRI, MRI-guided adaptive radiotherapy (ART) is well-suited to managing interfractional changes in anatomy. An MRI-only workflow is desirable, but producing synthetic CT (sCT) data through paired data-driven deep learning (DL) for abdominal dose calculations remains a challenge due to the highly variable presence of intestinal gas. We present the preliminary dosimetric evaluation of our novel approach to sCT reconstruction that is well suited to handling intestinal gas in abdominal MRI-only ART.Approach. We utilize a paired data DL approach enabled by the intensity projection prior, in which well-matching training pairs are created by propagating air from MRI to corresponding CT scans. Evaluations focus on two classes: patients with (1) little involvement of intestinal gas, and (2) notable differences in intestinal gas presence between corresponding scans. Comparisons between sCT-based plans and CT-based clinical plans for both classes are made at the first treatment fraction to highlight the dosimetric impact of the variable presence of intestinal gas.Main results. Class 1 patients (n= 13) demonstrate differences in prescribed dose coverage of the PTV of 1.3 ± 2.1% between clinical plans and sCT-based plans. Mean DVH differences in all structures for Class 1 patients are found to be statistically insignificant. In Class 2 (n= 20), target coverage is 13.3 ± 11.0% higher in the clinical plans and mean DVH differences are found to be statistically significant.Significance. Significant deviations in calculated doses arising from the variable presence of intestinal gas in corresponding CT and MRI scans result in uncertainty in high-dose regions that may limit the effectiveness of adaptive dose escalation efforts. We have proposed a paired data-driven DL approach to sCT reconstruction for accurate dose calculations in abdominal ART enabled by the creation of a clinically unavailable training data set with well-matching representations of intestinal gas.

Development of Dosimetric Verification System for Patient-Specific Quality Assurance of High-Dose-Rate Brachytherapy.

Purpose: The aim of this study was to develop a dosimetric verification system (DVS) using a solid phantom for patient-specific quality assurance (QA) of high-dose-rate brachytherapy (HDR-BT). Methods: The proposed DVS consists of three parts: dose measurement, dose calculation, and analysis. All the dose measurements were performed using EBT3 film and a solid phantom. The solid phantom made of acrylonitrile butadiene styrene (ABS, density = 1.04 g/cm3) was used to measure the dose distribution. To improve the accuracy of dose calculation by using the solid phantom, a conversion factor [CF(r)] according to the radial distance between the water and the solid phantom material was determined by Monte Carlo simulations. In addition, an independent dose calculation program (IDCP) was developed by applying the obtained CF(r). To validate the DVS, dosimetric verification was performed using gamma analysis with 3% dose difference and 3 mm distance-to-agreement criterion for three simulated cases: single dwell position, elliptical dose distribution, and concave elliptical dose distribution. In addition, the possibility of applying the DVS in the high-dose range (up to 15 Gy) was evaluated. Results: The CF(r) between the ABS and water phantom was 0.88 at 0.5 cm. The factor gradually increased with increasing radial distance and converged to 1.08 at 6.0 cm. The point doses 1 cm below the source were 400 cGy in the treatment planning system (TPS), 373.73 cGy in IDCP, and 370.48 cGy in film measurement. The gamma passing rates of dose distributions obtained from TPS and IDCP compared with the dose distribution measured by the film for the simulated cases were 99.41 and 100% for the single dwell position, 96.80 and 100% for the elliptical dose distribution, 88.91 and 99.70% for the concave elliptical dose distribution, respectively. For the high-dose range, the gamma passing rates in the dose distributions between the DVS and measurements were above 98% and higher than those between TPS and measurements. Conclusion: The proposed DVS is applicable for dosimetric verification of HDR-BT, as confirmed through simulated cases for various doses.

Patient-specific motion artifacts in 4DCT.

PURPOSE: Four-dimensional computed tomography (4DCT) has enhanced images of the thorax and upper abdomen during respiration, but intraphase residual motion artifacts will persist in cine-mode scanning. In this study, the source and magnitude of projection artifacts due to intraphase target motion is investigated. METHODS: A theoretical model of geometric uncertainty due to partial projection artifacts in cine-mode 4DCT was derived based on ideal periodic motion. Predicted artifacts were compared to measured errors with a rigid lung phantom attached to a programmable motion platform. Ideal periodic motion and actual patient breathing patterns were used as input for phantom motion. Reconstructed target dimensions were measured along the direction of motion and compared to the actual, known dimensions. RESULTS: Artifacts due to intraphase residual motion in cine-mode 4DCT range from a few mm up to a few cm on a given scanner, and can be predicted based on target motion and CT gantry rotation time. Errors in ITV and GTV dimensions were accurately characterized by the theoretical uncertainty at all phases when sinusoidal motion was considered, and in 96% of 300 measurements when patient breathing patterns were used as motion input. When peak-to-peak motion of 1.5 cm is combined with a breathing period of 4 s and gantry rotation time of 1 s, errors due to partial projection artifacts can be greater than 1 cm near midventilation and are a few mm in the inhale and exhale phases. Incorporation of such uncertainty into margin design should be considered in addition to other uncertainties. CONCLUSIONS: Artifacts due to intraphase residual motion exist in 4DCT, even for ideal breathing motions (e.g., sine waves). It was determined that these motion artifacts depend on patient-specific tumor motion and CT gantry rotation speed. Thus, if the patient-specific motion parameters are known (i.e., amplitude and period), a patient-specific margin can and should be designed to compensate for this uncertainty.

Development of Volumetric Independent Dose Calculation System for Verification of the Treatment Plan in Image-Guided Adaptive Brachytherapy.

Purpose: This study aimed to develop a volumetric independent dose calculation (vIDC) system for verification of the treatment plan in image-guided adaptive brachytherapy (IGABT) and to evaluate the feasibility of the vIDC in clinical practice with simulated cases. Methods: The vIDC is based on the formalism of TG-43. Four simulated cases of cervical cancer were selected to retrospectively evaluate the dose distributions in IGABT. Some reference point doses, such as points A and B and rectal points, were calculated by vIDC using absolute coordinate. The 3D dose volume was also calculated to acquire dose-volume histograms (DVHs) with grid resolutions of 1.0 × 1.0 (G1.0), 2.5 × 2.5 (G2.5), and 0.5 × 0.5 mm2 (G0.5). Dosimetric parameters such as D90% and D2cc doses covering 90% of the high-risk critical target volume (HR-CTV) and 2 cc of the organs at risk (OARs) were obtained from DVHs. D90% also converted to equivalent dose in 2-Gy fractions (EQD2) to produce the same radiobiological effect as external beam radiotherapy. In addition, D90% was obtained in two types with or without the applicator volume to confirm the effect of the applicator itself. Validation of the vIDC was also performed using gamma evaluation by comparison with Monte Carlo simulation. Results: The average percentage difference of point doses was <2.28%. The DVHs for the HR-CTV and OARs showed no significant differences between the vIDC and the treatment planning system (TPS). Without considering the applicator volume, the D90% of the HR-CTV calculated by the vIDC decreases with a decreasing calculated dose-grid size (32.4, 5.65, and -2.20 cGy in G2.5, G1.0, and G0.5, respectively). The overall D90% is higher when considering the applicator volume. The converted D90% by EQD2 ranged from -1.29 to 1.00%. The D2cc of the OARs showed that the averaged dose deviation is <10 cGy regardless of the dose-grid size. Based on gamma analysis, the passing rate was 98.81% for 3%/3-mm criteria. Conclusion: The vIDC was developed as an independent dose verification system for verification of the treatment plan in IGABT. We confirmed that the vIDC is suitable for second-check dose validation of the TPS under various conditions.

Technical Note: Real-time 3D MRI in the presence of motion for MRI-guided radiotherapy: 3D Dynamic keyhole imaging with super-resolution.

PURPOSE: The purpose of this study was to present real-time three-dimensional (3D) magnetic resonance imaging (MRI) in the presence of motion for MRI-guided radiotherapy (MRgRT) using dynamic keyhole imaging for high-temporal acquisition and super-resolution generative (SRG) model for high-spatial reconstruction. METHOD: We propose a unique real-time 3D MRI technique by combining a data sharing technique (3D dynamic keyhole imaging) with a SRG model using cascaded deep learning technique. 3D dynamic keyhole imaging utilizes the data sharing mechanism by combining keyhole central k-space data acquired in real-time with high-spatial, high-temporal resolution prior peripheral k-space data at various motion positions prepared by the SRG model. The efficacy of the 3D dynamic keyhole imaging with super-resolution (SR_dKeyhole) was compared to the ground-truth super-resolution images with the original full k-space data. It was also compared with the zero-filling reconstruction (zero-filling), conventional keyhole reconstruction with low-spatial high-temporal prior data (LR_cKeyhole), and conventional keyhole reconstruction with super-resolution prior data (SR_cKeyhole). RESULTS: High-spatial, high-temporal resolution 3D MRI datasets (1.5 × 1.5 × 6 mm3 ) were generated from low-spatial, high-temporal resolution 3D MRI datasets (6 × 6 × 6 mm3 ) using the cascaded deep learning SRG framework (<100 ms/volume). 3D dynamic keyhole imaging with the SRG model provided high-spatial, high-temporal resolution images (1.5 × 1.5 × 6 mm3 , 455 ms) with the highest similarity to the ground-truth SR images without any noticeable artifacts. Structural similarity indices (SSIM) of the reconstructed 3D MRI to the original SR 3D MRI were 0.65, 0.66, 0.86, and 0.89 for zero-filling, LR_cKeyhole, SR_cKeyhole, and SR_dKeyhole, respectively (1 for identical image). In addition, average value of image relative error (IRE) of the reconstructed 3D MRI to the original SR 3D MRI were 0.169, 0.191, 0.079, and 0.067 for zero-filling, LR_cKeyhole, SR_cKeyhole, and SR_dKeyhole, respectively (0 for identical image). CONCLUSIONS: We demonstrated that high-spatial, high-temporal resolution 3D MRI was feasible by combing 3D dynamic keyhole imaging with a SRG model in terms of image quality and imaging time. The proposed technique can be utilized for real-time 3D MRgRT.