Digitally Reconstructed Radiographs Research Articles

Physeal Location on the Distal Femoral Cortical Surface Is Consistently Proximal to Apparent Location on the Perfect Lateral Radiograph: A 3-Dimensional Digitally Reconstructed Radiograph Study.

The perfect knee lateral radiograph visualizes anatomic landmarks on the distal femur for clinical and scientific purposes. However, radiographic imaging is a two-dimensional (2D) representation of a three-dimensional (3D) physis. The aim of this study was to characterize the perceived radiographic projection of the femoral physis using perfect lateral digitally reconstructed radiographs (DRRs) and to evaluate discrepancies from this projection to the physis at the lateral and medial cortices. Pediatric patients from a cohort of CT scans were analyzed. Inclusion criteria were an open physis; exclusion criteria were any implant or pathology affecting the physis. CT scans were imported into 3D imaging software and transformed into lateral DRRs and 3D renderings of the femur. The physis was divided into four equal segments, with fiducial markers placed at the "anterior," "midpoint," and "posterior" points. Lines extended from these points in the lateral and medial direction. The vertical distance from these lines, representing the radiographic projection of the physis, was measured relative to the physis at the lateral and medial cortex of the femur on coronal CT slices. Thirty-one patients were included. On the perfect lateral radiograph DRR, the physis on the medial cortex was located proximal to the visualized physis by 6.64 ± 1.74 mm, 11.95 ± 1.67 mm, and 14.30 ± 1.75 mm at the anterior (25%), midpoint (50%), and posterior (75%) locations, respectively. On the lateral side, the physis on the lateral cortex was proximal to the visualized physis by 2.19 ± 1.13 mm, 3.71 ± 1.19 mm, and 6.74 ± 1.25 mm at the anterior, midpoint, and posterior locations, respectively. In this cohort of pediatric patients, the location of the cortical physis was, in all areas measured, proximal to the projection of the visualized physis as seen on the perfect knee lateral DRR. The distance from radiographic physis to cortical physis was greater at the medial cortex compared with the lateral cortex. Descriptive laboratory study. III, observational radiographic anatomic study.

The accuracy of cup anteversion measurement on postoperative pelvic radiographs: A comparative retrospective cohort study between DDH and non-DDH patients

Rationale and objectivesPostoperative pelvic radiographs remain a vital tool for assessing cup orientation after total hip arthroplasty (THA), with the accuracy influenced by various factors. The objective of this study is to investigate the accuracy of cup anteversion measurement in developmental dysplasia of the hip (DDH) patients and others based on postoperative pelvic radiographs conducted under the current heavy workload conditions. Materials and methodsPatients who underwent THA at our hospital with both postoperative X-ray and CT images from January 2020 to December 2022 were included in this retrospective cohort study. Virtual X-ray films were generated using digitally reconstructed radiographs (DRR) technology from CT images, with pelvic position perfectly controlled. Radiographic anteversion (RA) was measured on 3D-CT, virtual X-rays, and actual postoperative X-rays, abbreviated as RA_3D, RA_DRR, and RA_Xray, respectively. A repeated-measures analysis of covariance (ANCOVA) was utilized to evaluate the variations in RA within and between different groups across three methods. The Bland-Altman plot analysis showed the variations among methods in DDH and non-DDH patients, setting a clinically acceptable limits of agreement (LOA) at ±5°. ResultsThis study included 154 hip cases, with 63 DDH and 91 other diseases. Repeated-measures ANCOVA revealed a descending trend in RA across three methods, with differences of 2.64° (DDH) vs. 2.74° (others) from 3D to DRR, and 4.89° (DDH) vs. 1.07° (others) from DRR to X-ray. The group by methods interaction effect were significant (p = 0.002). Significant statistical differences in RA_Xray (P = 0.035) were observed between DDH and non-DDH patients, but not in RA_3D and RA_DRR. Bland-Altman plots showed 71.4 % of DDH patients exceeded the clinically acceptable LOA, compared to 36.3 % of other patients. ConclusionOur study indicated that under the current intense workload, the reliability of assessing cup anteversion using postoperative pelvic radiographs is challenged, especially in patients with DDH.

Isocentric fixed angle irradiation-based DRR: a novel approach to enhance x-ray and CT image registration.

Digitally Reconstructed Radiography (DRR) plays an important role in the registration of intraoperative X-ray and preoperative CT images. However, existing DRR algorithms often neglect the critical Isocentric Fixed Angle Irradiation (IFAI) principle in C-arm imaging, resulting in inaccurate simulation of X-ray images. This limitation degrades registration algorithms relying on DRR image libraries or employing DRR images (DRRs) to train neural network models. To address this issue, we propose a novel IFAI-based DRR method that accurately captures the true projection transformation during X-ray imaging of the human body.
Approach. By strictly adhering to the IFAI principle and utilizing known parameters from intraoperative X-ray images paired with CT scans, our method successfully simulates the real projection transformation and generates DRRs that closely resemble actual X-ray images.
Main result. Experimental results validate the effectiveness of our IFAI-based DRR method by successfully registering intraoperative X-ray images with preoperative CT images from multiple patients who underwent thoracic endovascular aortic procedures.
Significance. The proposed IFAI-based DRR method enhances the quality of DRR images, significantly accelerates the construction of DRR image libraries, and thereby improves the performance of X-ray and CT image registration. Additionally, the method has the generality of registering CT and X-ray images generated by large C-arm devices.&#xD.

CBCT-DRRs superior to CT-DRRs for target-tracking applications for pancreatic SBRT

Objective. In current radiograph-based intra-fraction markerless target-tracking, digitally reconstructed radiographs (DRRs) from planning CTs (CT-DRRs) are often used to train deep learning models that extract information from the intra-fraction radiographs acquired during treatment. Traditional DRR algorithms were designed for patient alignment (i.e. bone matching) and may not replicate the radiographic image quality of intra-fraction radiographs at treatment. Hypothetically, generating DRRs from pre-treatment Cone-Beam CTs (CBCT-DRRs) with DRR algorithms incorporating physical modelling of on-board-imagers (OBIs) could improve the similarity between intra-fraction radiographs and DRRs by eliminating inter-fraction variation and reducing image-quality mismatches between radiographs and DRRs. In this study, we test the two hypotheses that intra-fraction radiographs are more similar to CBCT-DRRs than CT-DRRs, and that intra-fraction radiographs are more similar to DRRs from algorithms incorporating physical models of OBI components than DRRs from algorithms omitting these models. Approach. DRRs were generated from CBCT and CT image sets collected from 20 patients undergoing pancreas stereotactic body radiotherapy. CBCT-DRRs and CT-DRRs were generated replicating the treatment position of patients and the OBI geometry during intra-fraction radiograph acquisition. To investigate whether the modelling of physical OBI components influenced radiograph-DRR similarity, four DRR algorithms were applied for the generation of CBCT-DRRs and CT-DRRs, incorporating and omitting different combinations of OBI component models. The four DRR algorithms were: a traditional DRR algorithm, a DRR algorithm with source-spectrum modelling, a DRR algorithm with source-spectrum and detector modelling, and a DRR algorithm with source-spectrum, detector and patient material modelling. Similarity between radiographs and matched DRRs was quantified using Pearson’s correlation and Czekanowski’s index, calculated on a per-image basis. Distributions of correlations and indexes were compared to test each of the hypotheses. Distribution differences were determined to be statistically significant when Wilcoxon’s signed rank test and the Kolmogorov-Smirnov two sample test returned p ≤ 0.05 for both tests. Main results. Intra-fraction radiographs were more similar to CBCT-DRRs than CT-DRRs for both metrics across all algorithms, with all p ≤ 0.007. Source-spectrum modelling improved radiograph-DRR similarity for both metrics, with all p < 10−6. OBI detector modelling and patient material modelling did not influence radiograph-DRR similarity for either metric. Significance. Generating DRRs from pre-treatment CBCT-DRRs is feasible, and incorporating CBCT-DRRs into markerless target-tracking methods may promote improved target-tracking accuracies. Incorporating source-spectrum modelling into a treatment planning system’s DRR algorithms may reinforce the safe treatment of cancer patients by aiding in patient alignment.

Enhancing fracture diagnosis in pelvic X-rays by deep convolutional neural network with synthesized images from 3D-CT

Pelvic fractures pose significant challenges in medical diagnosis due to the complex structure of the pelvic bones. Timely diagnosis of pelvic fractures is critical to reduce complications and mortality rates. While computed tomography (CT) is highly accurate in detecting pelvic fractures, the initial diagnostic procedure usually involves pelvic X-rays (PXR). In recent years, many deep learning-based methods have been developed utilizing ImageNet-based transfer learning for diagnosing hip and pelvic fractures. However, the ImageNet dataset contains natural RGB images which are different than PXR. In this study, we proposed a two-step transfer learning approach that improved the diagnosis of pelvic fractures in PXR images. The first step involved training a deep convolutional neural network (DCNN) using synthesized PXR images derived from 3D-CT by digitally reconstructed radiographs (DRR). In the second step, the classification layers of the DCNN were fine-tuned using acquired PXR images. The performance of the proposed method was compared with the conventional ImageNet-based transfer learning method. Experimental results demonstrated that the proposed DRR-based method, using 20 synthesized PXR images for each CT, achieved superior performance with the area under the receiver operating characteristic curves (AUROCs) of 0.9327 and 0.8014 for visible and invisible fractures, respectively. The ImageNet-based method yields AUROCs of 0.8908 and 0.7308 for visible and invisible fractures, respectively.

4D-Precise: Learning-based 3D motion estimation and high temporal resolution 4DCT reconstruction from treatment 2D+t X-ray projections

Background and ObjectiveIn radiotherapy treatment planning, respiration-induced motion introduces uncertainty that, if not appropriately considered, could result in dose delivery problems. 4D cone-beam computed tomography (4D-CBCT) has been developed to provide imaging guidance by reconstructing a pseudo-motion sequence of CBCT volumes through binning projection data into breathing phases. However, it suffers from artefacts and erroneously characterizes the averaged breathing motion. Furthermore, conventional 4D-CBCT can only be generated post-hoc using the full sequence of kV projections after the treatment is complete, limiting its utility. Hence, our purpose is to develop a deep-learning motion model for estimating 3D+t CT images from treatment kV projection series. MethodsWe propose an end-to-end learning-based 3D motion modelling and 4DCT reconstruction model named 4D-Precise, abbreviated from Probabilistic reconstruction of image sequences from CBCT kV projections. The model estimates voxel-wise motion fields and simultaneously reconstructs a 3DCT volume at any arbitrary time point of the input projections by transforming a reference CT volume. Developing a Torch-DRR module, it enables end-to-end training by computing Digitally Reconstructed Radiographs (DRRs) in PyTorch. During training, DRRs with matching projection angles to the input kVs are automatically extracted from reconstructed volumes and their structural dissimilarity to inputs is penalised. We introduced a novel loss function to regulate spatio-temporal motion field variations across the CT scan, leveraging planning 4DCT for prior motion distribution estimation. ResultsThe model is trained patient-specifically using three kV scan series, each including over 1200 angular/temporal projections, and tested on three other scan series. Imaging data from five patients are analysed here. Also, the model is validated on a simulated paired 4DCT-DRR dataset created using the Surrogate Parametrised Respiratory Motion Modelling (SuPReMo). The results demonstrate that the reconstructed volumes by 4D-Precise closely resemble the ground-truth volumes in terms of Dice, volume similarity, mean contour distance, and Hausdorff distance, whereas 4D-Precise achieves smoother deformations and fewer negative Jacobian determinants compared to SuPReMo. ConclusionsUnlike conventional 4DCT reconstruction techniques that ignore breath inter-cycle motion variations, the proposed model computes both intra-cycle and inter-cycle motions. It represents motion over an extended timeframe, covering several minutes of kV scan series.

230 Machine Learning for Automated and Real-time 2D-3D Registration of the Spine Using Only a Single Radiograph

INTRODUCTION: Authors present a 2-dimensional to 3-dimensional (2D-3D) registration algorithm that identifies the position of a C-arm relative to a 3D volume. METHODS: This work utilizes digitally reconstructed radiographs (DRR) which are synthetic radiographic images generated by simulating the x-ray projections as they would pass through a CT volume. To evaluate our algorithm, we used cone-beam CT data from 127 patients obtained from a de-identified registry of cervical, thoracic, and lumbar scans. We systematically evaluated and tuned a self-supervised algorithm, quantified the limitations and convergence rate of our model by simulating C-arm registrations with 80 randomly simulated DRRs for each CT volume. The endpoints for this study were the time to convergence, accuracy of convergence for each of the C-arm's degrees of freedom (DoF), and overall registration accuracy based on a voxel-by-voxel measurement. RESULTS: 10,160 unique fluoroscopic images were simulated from 127 CT scans. Our algorithm successfully converged to the correct solution 81% of the time with an average of 1.97 seconds of computation. The fluoroscopic images for which the algorithm converged to the solution achieved 99.9% registration accuracy despite only utilizing single-precision computation for speed. Our algorithm was found to be optimized for convergence when the search space was limited to a ± 45° offset in the RAO/LAO, cranial/caudal and receiver rotation angles with the radiographic isocenter contained within 8000cm3 of the volumetric center of the CT volume. CONCLUSIONS: The machine learning algorithm we present has the potential to aid surgeons with many aspects of spine surgery through an automated 2D-3D registration process. Future work will focus on algorithmic optimizations adn deep-learning to improve the convergence rate and speed profile.

Deep learning-based target decomposition for markerless lung tumor tracking in radiotherapy.

In radiotherapy, real-time tumor tracking can verify tumor position during beam delivery, guide the radiation beam to target the tumor, and reduce the chance of a geometric miss. Markerless kV x-ray image-based tumor tracking is challenging due to the low tumor visibility caused by tumor-obscuring structures. Developing a new method to enhance tumor visibility for real-time tumor tracking is essential. To introduce a novel method for markerless kV image-based tracking of lung tumors via deep learning-based target decomposition. We utilized a conditional Generative Adversarial Network (cGAN), known as Pix2Pix, to build a patient-specific model and generate the synthetic decomposed target image (sDTI) to enhance tumor visibility on the real-time kV projection images acquired by the onboard kV imager equipped on modern linear accelerators. We used 4DCT simulation images to generate the digitally reconstructed radiograph (DRR) and DTI image pairs for model training. We augmented the training dataset by randomly shifting the 4DCT in the superior-inferior, anterior-posterior, and left-right directions during the DRR and DTI generation process. We performed real-time 2D tumor tracking via template matching between the DTI generated from the CT simulation and the sDTI generated from the real-time kV projection images. We validated the proposed method using nine patients' datasets with implanted beacons near the tumor. The sDTI can effectively improve the image contrast around the lung tumors on the kV projection images for the nine patients. With the beacon motion as ground truth, the tracking errors were on average 0.8±0.7mm in the superior-inferior (SI) direction and 0.9±0.8mm in the in-plane left-right (IPLR) direction. The percentage of successful tracking, defined as a tracking error less than 2mm in the SI direction, is 92.2% on the 4312 tested images. The patient-specific model took approximately 12h to train. During testing, it took approximately 35ms to generate one sDTI, and 13ms to perform the tumor tracking using template matching. Our method offers the potential solution for nearly real-time markerless lung tumor tracking. It achieved a high level of accuracy and an impressive tracking rate. Further development of 3D lung tumor tracking is warranted.

Automatic hip osteoarthritis grading with uncertainty estimation from computed tomography using digitally-reconstructed radiographs.

Progression of hip osteoarthritis (hip OA) leads to pain and disability, likely leading to surgical treatment such as hip arthroplasty at the terminal stage. The severity of hip OA is often classified using the Crowe and Kellgren-Lawrence (KL) classifications. However, as the classification is subjective, we aimed to develop an automated approach to classify the disease severity based on the two grades using digitally-reconstructed radiographs from CT images. Automatic grading of the hip OA severity was performed using deep learning-based models. The models were trained to predict the disease grade using two grading schemes, i.e., predicting the Crowe and KL grades separately, and predicting a new ordinal label combining both grades and representing the disease progression of hip OA. The models were trained in classification and regression settings. In addition, the model uncertainty was estimated and validated as a predictor of classification accuracy. The models were trained and validated on a database of 197 hip OA patients, and externally validated on 52 patients. The model accuracy was evaluated using exact class accuracy (ECA), one-neighbor class accuracy (ONCA), and balanced accuracy. The deep learning models produced a comparable accuracy of approximately 0.65 (ECA) and 0.95 (ONCA) in the classification and regression settings. The model uncertainty was significantly larger in cases with large classification errors ( ). In this study, an automatic approach for grading hip OA severity from CT images was developed. The models have shown comparable performance with high ONCA, which facilitates automated grading in large-scale CT databases and indicates the potential for further disease progression analysis. Classification accuracy was correlated with the model uncertainty, which would allow for the prediction of classification errors. The code will be made publicly available at https://github.com/NAIST-ICB/HipOA-Grading .

Automated detection of vertebral body misalignments in orthogonal kV and MV guided radiotherapy: application to a comprehensive retrospective dataset

Objective. In image-guided radiotherapy (IGRT), off-by-one vertebral body misalignments are rare but potentially catastrophic. In this study, a novel detection method for such misalignments in IGRT was investigated using densely-connected convolutional networks (DenseNets) for applications towards real-time error prevention and retrospective error auditing. Approach. A total of 4213 images acquired from 527 radiotherapy patients aligned with planar kV or MV radiographs were used to develop and test error-detection software modules. Digitally reconstructed radiographs (DRRs) and setup images were retrieved and co-registered according to the clinically applied alignment contained in the DICOM REG files. A semi-automated algorithm was developed to simulate patient positioning errors on the anterior-posterior (AP) and lateral (LAT) images shifted by one vertebral body. A DenseNet architecture was designed to classify either AP images individually or AP and LAT image pairs. Receiver-operator characteristic curves (ROC) and areas under the curves (AUC) were computed to evaluate the classifiers on test subsets. Subsequently, the algorithm was applied to the entire dataset in order to retrospectively determine the absolute off-by-one vertebral body error rate for planar radiograph guided RT at our institution from 2011–2021. Main results. The AUCs for the kV models were 0.98 for unpaired AP and 0.99 for paired AP-LAT. The AUC for the MV AP model was 0.92. For a specificity of 95%, the paired kV model achieved a sensitivity of 99%. Application of the model to the entire dataset yielded a per-fraction off-by-one vertebral body error rate of 0.044% [0.0022%, 0.21%] for paired kV IGRT including one previously unreported error. Significance. Our error detection algorithm was successful in classifying vertebral body positioning errors with sufficient accuracy for retrospective quality control and real-time error prevention. The reported positioning error rate for planar radiograph IGRT is unique in being determined independently of an error reporting system.

Estimation of Cochlear Implant Insertion Depth Using 2D-3D Registration of Postoperative X-Ray and Preoperative CT Images.

To improve estimation of cochlear implant (CI) insertion depth in postoperative skull x-rays using synthesized information from preoperative CT scans. Retrospective cohort. Tertiary referral center. Ten adult cochlear implant recipients with preoperative and postoperative temporal bone computed tomography (CT)scans and postoperative skull x-ray imaging. Postoperative x-rays and digitally reconstructed radiographs (DRR) from preoperative CTs were registered using 3D Slicer and MATLAB to enhance localization of the round window and modiolus. Angular insertion depth (AID) was estimated in unmodified and registration-enhanced x-rays and DRRs in the cochlear view. Linear insertion depth (LID) was estimated in registered images by two methods that localized the proximal CI electrode or segmented the cochlea. Ground truth assessments were made in postoperative CTs. Errors of insertion depth estimates were calculated relative to ground truth measurements and compared with paired t t ests. Pearson correlation coefficient was used to assess inter-rater reliability of two reviewer's measurements of AID in unmodified x-rays. In postoperative x-rays, AID estimation errors were similar with and without registration enhancement (-1.3 ± 20.7° and -4.8 ± 24.9°, respectively; mean ± SD; p = 0.6). AID estimation in unmodified x-rays demonstrated strong interrater agreement (ρ = 0.79, p < 0.05) and interrater differences (-15.0 ± 35.3°) comparable to estimate errors. Registering images allowed measurement of AID in the cochlear view with estimation errors of 14.6 ± 30.6° and measurement of LID, with estimate errors that were similar between proximal electrode localization and cochlear segmentation methods (-0.9 ± 2.2 mm and -2.1 ± 2.7 mm, respectively; p = 0.3). 2D-3D image registration allows measurement of AID in the cochlear view and LID using postoperative x-rays and preoperative CT imaging. The use of this technique may reduce the need for postimplantation CT studies to assess these metrics of CI electrode position. Further work is needed to improve the accuracy of AID assessment in the postoperative x-ray view with registered images compared with established methods.

Deep learning in computed tomography super resolution using multi-modality data training.

One of the limitations in leveraging the potential of artificial intelligence in X-ray imaging is the limited availability of annotated training data. As X-ray and CT shares similar imaging physics, one could achieve cross-domain data sharing, so to generate labeled synthetic X-ray images from annotated CT volumes as digitally reconstructed radiographs (DRRs). To account for the lower resolution of CT and the CT-generated DRRs as compared to the real X-ray images, we propose the use of super-resolution (SR) techniques to enhance the CT resolution before DRRgeneration. As spatial resolution can be defined by the modulation transfer function kernel in CT physics, we propose to train a SR network using paired low-resolution (LR) and high-resolution (HR) images by varying the kernel's shape and cutoff frequency. This is different to previous deep learning-based SR techniques on RGB and medical images which focused on refining the sampling grid. Instead of generating LR images by bicubic interpolation, we aim to create realistic multi-detector CT (MDCT) like LR images from HR cone-beam CT (CBCT)scans. We propose and evaluate the use of a SR U-Net for the mapping between LR and HR CBCT image slices. We reconstructed paired LR and HR training volumes from the same CT scans with small in-plane sampling grid size of . We used the residual U-Net architecture to train two models. SRUN : trained with kernel-based LR images, and SRUN : trained with bicubic downsampled data as baseline. Both models are trained on one CBCT dataset (n = 13 391). The performance of both models was then evaluated on unseen kernel-based and interpolation-based LR CBCT images (n = 10 950), and also on MDCT images (n = 1392). Five-fold cross validation and ablation study were performed to find the optimal hyperparameters. Both SRUN and SRUN models show significant improvements (p-value 0.05) in mean absolute error (MAE), peak signal-to-noise ratio (PSNR) and structural similarity index measures (SSIMs) on unseen CBCT images. Also, the improvement percentages in MAE, PSNR, and SSIM by SRUN is larger than SRUN . For SRUN , MAE is reduced by 14%, and PSNR and SSIMs increased by 6 and 8%, respectively. To conclude, SRUN outperforms SRUN , which the former generates sharper images when tested with kernel-based LR CBCT images as well as cross-modality LR MDCTdata. Our proposed method showed better performance than the baseline interpolation approach on unseen LR CBCT. We showed that the frequency behavior of the used data is important for learning the SR features. Additionally, we showed cross-modality resolution improvements to LR MDCT images. Our approach is, therefore, a first and essential step in enabling realistic high spatial resolution CT-generated DRRs for deep learningtraining.

Deep-learning soft-tissue decomposition in chest radiography using fast fuzzy C-means clustering with CT datasets

Chest radiography is the most routinely used X-ray imaging technique for screening and diagnosing lung and chest disease, such as lung cancer and pneumonia. However, the clinical interpretation of the hidden and obscured anatomy in chest X-ray images remains challenging because of the bony structures overlapping the lung area. Thus, multi-perspective imaging with a high radiation dose is often required. In this study, to address this problem, we propose a deep-learning soft-tissue decomposition method using fast fuzzy C-means (FFCM) clustering with computed tomography (CT) datasets. In this method, FFCM clustering is used to decompose a CT image into bone and soft-tissue components, which are synthesized into digitally reconstructed radiographs (DRRs) to obtain large amounts of X-ray decomposition datasets as ground truths for training. In the training stage, chest and soft-tissue DRRs are used as input and label data, respectively, for training the network. During the testing, a chest X-ray image is fed into the trained network to output the corresponding soft-tissue image component. To verify the efficacy of the proposed method, we conducted a feasibility study on clinical CT datasets available from the AAPM Lung CT Challenge. According to our results, the proposed method effectively yielded soft-tissue decomposition from chest X-ray images; this is encouraging for reducing the visual complexity of chest X-ray images. Consequently, the finding of our feasibility study indicate that the proposed method can offer a promising outcome for this purpose.

DRR acceleration using inexpensive GPUs for model-image registration based joint kinematic measurements

Model-image registration methods are commonly used in research to measure three-dimensional joint kinematics from single-plane and bi-plane x-ray images. These methods have the potential to be beneficial if used clinically, but current techniques are too slow or expensive to be clinically practical. One technical element of these methods for measuring natural bone motion is the use of digitally reconstructed radiographs (DRRs). DRRs can be very expensive to compute, or require expensive and fast computer hardware. In this technical development, a numerically efficient Siddon-Jacobs algorithm for computing DRRs was implemented on a consumer-grade graphics card using a programming language for parallel architectures. Compared to traditional voxel projection algorithms with a central-processing-unit-only implementation, the parallel computation implementation on the graphics card provided speedups of 650–1546 times faster rendering, while retaining equivalent performance for joint kinematics measurements. The use of consumer grade graphics hardware may contribute to making model-image registration measurements of joint kinematics practical for clinical use.

Validity of the Center-Center Method for Syndesmotic Fixation Axis Against the Trans-Syndesmotic Axis: A Cross-Sectional Study

Category: Trauma; Ankle Introduction/Purpose: In surgery for ankle fractures with tibiofibular diastasis, it is important to vertically grasp the distal tibiofibular joint. However, it is difficult to do that because the tibiofibular joint surface is not under direct vision. Center-center method (CC) is an intraoperative fluoroscopic technique to determine the position of the syndesmotic clamp. CC has been reported to reduce the risk of syndesmotic malreduction, but It is unclear whether the axis by CC (CC axis) coincides with the axis perpendicular to the distal tibiofibular joint (trans-syndesmosis axis: TS axis). Therefore, the present study aims to 1) measure the difference between the CC axis and TS axis, and 2) compare the measured values when CC is performed at different distances from the ankle joint surface. Methods: The subjects were 150 patients who underwent CT scans for ankle joint disease and lower extremity trauma. CC is a method in which the ankle joint is rotated, and the midpoint of the anteroposterior margin of the tibia and fibula is superimposed by fluoroscopy. In the present study, CC was performed using Digitally reconstructed radiography (DRR) created from a CT image. At the ankle joint surface (0mm), CC axis and TS axis were projected onto the CT axis linked with DRR. The angle between the axes was measured (inter-axes angle). Also, the distance of the intersection of these axes through the medial cortex of the tibia was measured (inter-axes distance). It was also performed at 10mm and 20mm proximal to the ankle joint surface. One-way repeated measures analysis of variance and the post hoc Bonferroni tests were used to compare the measured values' difference according to the articular surface's height. Results: Overall, the TS axis was externally rotated relative to the CC axis, with a mean inter-axes angle of 8.5±6.8degrees. The inter-axes angle of 0mm, 10mm, and 20mm were 1.9±4.0degrees, 10±5.3degrees, and 13±5.1degrees, respectively (p &lt; 0.001). Also, the intersection point of the TS axis with the medial tibial cortex was anterior to the intersection of the CC axis, with an inter-axes distance of 7.7±6.3mm overall. The inter-axes distance of 0mm, 10mm, and 20mm were 2.0±4.2mm, 9.0±4.9mm, and 11.3±5.2mm, respectively (p &lt; 0.001). Both the inter-axes angle and the inter-axes distance were larger when the CC method was performed higher to the tibial plafond (P &lt; .001), with the post hoc tests showing significant differences among all heights. Conclusion: The TS axis was externally rotated to the CC axis and positioned slightly anteriorly. Also, the inter-axes angle and distance were larger when the CC method was performed higher to the tibial plafond. If the surgeon aimed to clamp the syndesmosis along the TS axis, the CC method should be applied closer to the ankle joint surface, and the medial tip of the forceps be placed slightly anterior to the point determined by the CC axis.

Dosimetric Advantages of VMAT TBI Plans: A Direct Comparison to Conventional TBI Plans.

Total body irradiation (TBI) continues to play an integral role in the conditioning of patients undergoing bone marrow transplantation. Historically, conventional TBI (C-TBI) has been delivered based on a clinical plan without CT simulation. However, volumetric modulated arc therapy TBI (V-TBI) is emerging as an alternative method to deliver TBI in a contemporary fashion. We aimed to compare these two methods of TBI delivery. Patients undergoing treatment with V-TBI were identified. C-TBI plans were created using their existing simulation CT images. Patient thickness was measured on the scan and compensators such as lead sheets to attenuate dose in areas with less separation and lung blocks drawn on digitally reconstructed radiographs (DRR) were added when necessary. A 3D dose distribution was then calculated allowing for the direct comparison between C-TBI and V-TBI on the same patient using the same CT image set. Dosimetric data from each plan including target volume coverage, dose homogeneity, absolute max dose, and dose to lungs were recorded. V-TBI and C-TBI plans for a total of four patients were preliminarily analyzed. Two patients were prescribed 200 cGy in a single fraction, while the other two were prescribed 1200 cGy in eight fractions. V-TBI resulted in a more favorable dosimetry for all four patients in most evaluated metrics including dose coverage, dose homogeneity, and lung dose (Table 1). V-TBI did result in an increased absolute maximum dose to all four patients compared to c-TBI, but still met the desired constraint of D0.03cc<125%. V-TBI resulted in more favorable dosimetry for all four patients compared to C-TBI. To our knowledge, this is the first direct dosimetric comparison between the two methods. Analysis of 8 more V-TBI cases is currently underway. In the future, we plan to design a prospective study to evaluate the clinical outcomes of patients undergoing V-TBI vs C-TBI.

Patient-Specific 3D CT Images Reconstruction from 2D KV Images Via Vision Transformer-Based Deep-Learning

In some proton therapy facilities, patient alignment relies on two 2D orthogonal kV images, taken at fixed, oblique angles, as no 3D on-the-bed-imaging is available. The visibility of the tumor in kV images is limited since the patient's 3D anatomy is projected onto a 2D plane, especially when the tumor is behind a high-density structure such as bone. This can lead to a large patient setup error. A solution to this problem is to reconstruct the 3D CT image from the kV images obtained in the treatment position. An asymmetric autoencoder-like network built with vision-transformer blocks was developed. The data was collected from a head and neck patient: 2 orthogonal kV images (1024X1024 voxels), 1 3D CT with padding (512X512X512) acquired from the in-room CT-on-rails before kVs were taken and 2 digitally-reconstructed-radiograph (DRR) images (512X512) based on the CT. We resampled kV images every 8 voxels and DRR and CT every 4 voxels, thus formed a dataset consisting of 262,144 samples, in which the images had a dimension of 128 for each direction. The value of each voxel in CT was normalized to range 0-1 with a uniform shift of 1000 and a denominator of 4000. For kV and DRR, we ranked all voxels value in an ascending order and normalized the values of the first 80% voxels to range 0-0.8 and the rest to range 0.8-1, thus yielding a quasi-Gaussian distribution, which was favorable by the deep neural networks. We further cropped kV and DRR images with a self-supervised bitmap based on the voxels' gradients. In training, both kV and DRR were utilized, and the encoder was encouraged to learn the same feature maps for kV images and its corresponding DRR images with mean-absolute-error (MAE) as the similarity loss. Then the decoder would reconstruct the 3D CT image from the feature maps of the kV images with the CT-on-rails as ground-truth (gCT) and MAE as the reconstruction loss. In testing, only independent kV images were used. The full-size synthetic CT (sCT) was achieved by concatenating the sCTs generated by the model according to their spatial information. The image quality of the sCT was evaluated using MAE and per-voxel-absolute-CT-number-difference volume histogram (CDVH). The proposed network was implemented with PyTorch deep learning library and both distributed data parallel (DDP) and automatic mixed precision (AMP) were applied to saving memory and accelerating the training speed. We used the AdamW optimizer with β1 = 0.9 and β2 = 0.999 and a cosine annealing learning rate scheduler with an initial learning of 1e-7 and 20 warm-up epochs. The model achieved a MAE of <40HU and the CDVH showed that <5% of the voxels had a per-voxel-absolute-CT-number-difference larger than 185HU. The profile of a typical gCT slice and its corresponding sCT slice exhibited a high agreement, indicating the high similarity between the gCT and sCT. A patient-specific vision-transformer-based network was developed and shown to be accurate and efficient to reconstruct 3D CT images from kV images.

Deep learning for automated measurement of CSA related acromion morphological parameters on anteroposterior radiographs

BackgroundThe Critical Shoulder Angle Related Acromion Morphological Parameter (CSA- RAMP) is a valuable tool in the analyzing the etiology of the rotator cuff tears (RCTs). However, its clinical application has been limited by the time-consuming and prone to inter- and intra-user variability of the measurement process. ObjectivesTo develop and validate a deep learning algorithm for fully automated assessment of shoulder anteroposterior radiographs associated with RCTs and calculation of CSA-RAMP. MethodsRetrospective analysis was conducted on radiographs obtained from computed tomography (CT) scans and X-rays performed between 2018 and 2020 at our institution. The development of the system involved the utilization of digitally reconstructed radiographs (DRRs) generated from each CT scan. The system's performance was evaluated by comparing it with manual and semiautomated measurements on two separate test datasets: dataset I (DRRs) and dataset II (X-rays). Standard metrics, including mean average precision (AP), were utilized to assess the segmentation performance. Additionally, the consistency among fully automated, semiautomated, and manual measurements was comprehensively evaluated using the Pearson correlation coefficient and Bland-Altman analysis. ResultsA total of 1080 DRRs generated from 120 consecutive CT scans and 159 X-ray films were included in the study. The algorithm demonstrated excellent segmentation performance, with a mean AP of 57.67 and an AP50 of 94.31. Strong inter-group correlations were observed for all CSA-RAMP measurements in both test datasets I (automated versus manual, automated versus semiautomated, and semiautomated versus manual; r = [0.990–––0.997], P < 0.001) and dataset II (r = [0.984–––0.995], P < 0.001). Bland-Altman analysis revealed low bias for all CSA-RAMP measurements in both test datasets I and II, except for CD (with a maximum bias of 2.49%). ConclusionsWe have successfully developed a fully automated algorithm capable of rapidly and accurately measuring CSA-RAMP on shoulder anteroposterior radiographs. A consistent automated CSA- RAMP measurement system may accelerate powerful and precise studies of disease biology in future large cohorts of RCTs patients.

Simultaneous object detection and segmentation for patient-specific markerless lung tumor tracking in simulated radiographs with deep learning.

Real-time tumor tracking is one motion management method to address motion-induced uncertainty. To date, fiducial markers are often required to reliably track lung tumors with X-ray imaging, which carries risks of complications and leads to prolonged treatment time. A markerless tracking approach is thus desirable. Deep learning-based approaches have shown promise for markerless tracking, but systematic evaluation and procedures to investigate applicability in individual cases are missing. Moreover, few efforts have been made to provide bounding box prediction and mask segmentation simultaneously, which could allow either rigid or deformable multi-leaf collimatortracking. The purpose of this study was to implement a deep learning-based markerless lung tumor tracking model exploiting patient-specific training which outputs both a bounding box and a mask segmentation simultaneously. We also aimed to compare the two kinds of predictions and to implement a specific procedure to understand the feasibility of markerless tracking on individualcases. We first trained a Retina U-Net baseline model on digitally reconstructed radiographs (DRRs) generated from a public dataset containing 875 CT scans and corresponding lung nodule annotations. Afterwards, we used an independent cohort of 97 lung patients to develop a patient-specific refinement procedure. In order to determine the optimal hyperparameters for automatic patient-specific training, we selected 13 patients for validation where the baseline model predicted a bounding box on planning CT (PCT)-DRR with intersection over union (IoU) with the ground-truth higher than 0.7. The final test set contained the remaining 84 patients with varying PCT-DRR IoU. For each testing patient, the baseline model was refined on the PCT-DRR to generate a patient-specific model, which was then tested on a separate 10-phase 4DCT-DRR to mimic the intrafraction motion during treatment. A template matching algorithm served as benchmark model. The testing results were evaluated by four metrics: the center of mass (COM) error and the Dice similarity coefficient (DSC) for segmentation masks, and the center of box (COB) error and the DSC for bounding box detections. Performance was compared to the benchmark model including statistical testing forsignificance. A PCT-DRR IoU value of 0.2 was shown to be the threshold dividing inconsistent (68%) and consistent (100%) success (defined as mean bounding box DSC > 0.6) of PS models on 4DCT-DRRs. Thirty-seven out of the eighty-four testing cases had a PCT-DRR IoU above 0.2. For these 37 cases, the mean COM error was 2.6 mm, the mean segmentation DSC was 0.78, the mean COB error was 2.7 mm, and the mean box DSC was 0.83. Including the validation cases, the model was applicable to 50 out of 97 patients when using the PCT-DRR IoU threshold of 0.2. The inference time per frame was 170 ms. The model outperformed the benchmark model on all metrics, and the comparison was significant (p < 0.001) over the 37 PCT-DRR IoU > 0.2 cases, but not over the undifferentiated 84 testingcases. The implemented patient-specific refinement approach based on a pre-trained baseline model was shown to be applicable to markerless tumor tracking in simulated radiographs for lungcases.

Validity and application of Doiguchi's pelvic tilt measurement method

BackgroundThe validity of Doiguchi's pelvic tilt measurement method has not been proven. The objective in our study was to validate the method. MethodsOur investigation included 73 total hip arthroplasties (THAs) performed using our cup placement procedure from July 2020 to November 2021. Pelvic tilt formed by the pubic symphysis and sacral promontory (PTPS) in supine and lateral positions was calculated by two methods (the Doiguchi method and the digital reconstructed radiograph (DRR) method using a 3D computer templating system) based on the transverse and longitudinal diameters of the pelvic ring measured immediately before THA. ResultsThere was a strong/moderate correlation in the values of PTPS between the Doiguchi and DRR methods. However, the value of PTPS calculated by the Doiguchi method was significantly lower than that calculated by DRR, and there was a partially direct match. On the other hand, there was no significant difference in the value of PT change from supine to lateral position between the Doiguchi and DRR methods. The PT changes based on both methods were strongly correlated, and the PT change calculated by the Doiguchi method was almost identical to that calculated by the DRR method. ConclusionsDoiguchi's pelvic tilt measurement method was validated for the first time. These results demonstrated that the ratio of the transverse and longitudinal diameters of the pelvic ring was an important factor defining the change in pelvic tilt. The slope in the linear function of the Doiguchi method was found to be almost the correct value, although the intercept of the linear function exhibited individual differences.