Validation of a CT-based motion model with in-situ fluoroscopy for lung surface deformation estimation

Objective. 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.

We have investigated a registration method between pre-operative 3D CT angiography (3D-CTA) and intra-operative fluoroscopic image sequence (with/without contrast injection) during intervention. Most registration methods are developed for assisting neurosurgery or orthopedic surgery, but our method is developed for interventional procedure such as endovascular stent grafting. In our method, DRR (Digitally Reconstructed Radiograph) are generated by voxel projection of 3D-CTA after extracting an aorta region. By increasing/decreasing CT value in the aorta region of CTA, DRR with/without contrast media injection are obtained. Subsequently we calculate distance (or similarity) measures between DRR and fluoroscopic image iteratively by changing imaging parameters. The most similar DRR to fluoroscopy is selected. We validated our algorithm by using simulated/clinical fluoroscopic images and DRR (with/without contrast media injection) of thorax and abdomen. Several distance (or similarity) measures were investigated in this experiment. Validation results show that M-estimator of residual is good as matching measure, and registration is well performed for almost all cases. However, accuracy is not enough for non-contrasted thoracic images, and calculation time should be reduced for all cases.© (2002) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Purpose: 4D cone beam computed tomography (4DCBCT) suffers from poor image quality, which limits its use as a reference image for image‐guided procedures that use radiography and fluoroscopy. This abstract demonstrates improved image quality in Digitally Reconstructed Radiographs (DRR) derived from thoracic 4DCBCT. Methods: In our method, projection images are used to create both a phase‐rebinned 4D volume and streak‐free 3D volume. These volumes are combined into a new 4DCBCT that includes stationary voxels from 3D volume and moving voxels from 4D volume. Stationary voxels include the chest wall, which has only small motion during breathing, and moving voxels include the soft tissues inside chest wall, such as lung, heart, and abdomen. The chest wall was segmented from 3D volume, and the segmented Result was used to replace corresponding voxels in 4D volume. To generate the DRR, only high intensity voxels, including chest wall, heart, abdomen, main vessels and bronchi, and tumor, were used. This method reduced streak artifacts of 4DCBCT volume, but preserved the motion information present within the lungs and diaphragm. Results: DRR images computed from 4DCBCT are visually poor because of the streak artifacts, while DRR images computed from 3DCBCT suffer from motion blur. The DRR computed from the combined 3D and 4D volumes are visually more similar to the original projection images at corresponding projection angles. Conclusion: DRR image quality declined because of the streak artifacts in 4DCBCT. By combining 4D volume and streak‐free 3D volume and extracting the high intensity voxels to generate DRR, the streak artifacts can be effectively reduced, which results in improvement of the DRR image quality. This work was supported by the National Basic Research Program of China (973) under Grant 2011CB707701 and research foundation of Tsinghua University.

Aims and Objectives: In this paper, a novel approach to the distinction between borrowing and code-switching is proposed, called the Simple View of borrowing and code-switching. Under this view, listedness is seen as the key condition for classifying words or multiword units (MWUs) as borrowings. For MWUs, listedness is operationalized with mutual information (MI) scores: the higher the MI score of a given set of words, the higher the likelihood it is listed in the lexicon. Under the Simple View, the distinction between borrowing and code-switching is seen as a specific instantiation of the distinction between what belongs in the lexicon (fixed, arbitrary patterns) and what is computed online (productive rules), and should therefore be considered as part of the grammar. Methodology: Assumptions from the Simple View were tested on a corpus of switches of single words and MWUs from a Turkish-German code-switching corpus (87,000 words), which was transcribed in CHAT format. Data and analysis: The frequency of switches in either direction, and their morphosyntactic integration patterns were analysed with CLAN. The formulaicity of the MWUs was analysed with MI scores through Sketchengine. Findings: The MI scores of the donor language MWUs were found to be above 3, which is the cut-off point for formulaicity in Corpus Linguistics. Thus, the MWUs were found to be borrowings. In addition, MWUs were found to be more likely to be borrowed than single words. Originality: Insights about formulaic language from Corpus Linguistics and Second Language Acquisition were used to inform analyses of language contact phenomena, and new ways to test the model are proposed. Significance: The Simple View offers a unified approach to borrowing of lexical items and function words and opens a new avenue for research using neuroscientific methods to test whether items are listed in speakers’ mental lexicons.

Over the past decade, extensive research has explored collocations with high mutual information (MI) scores. However, second-language learners often struggle to incorporate these collocations into their writing due to a lack of awareness and limited practice opportunities. To address this challenge, the present article presents a 10-week writing course that integrates various activities using COCA, an online corpus tool, to enhance students’ awareness and use of collocations with high MI scores. By analyzing student essays and their reflections on class activities, the article examines the benefits and challenges of using COCA for the instruction of collocations with high MI scores in the English language classroom. Pedagogical recommendations are provided to refine instructional practices.

Purpose: To develop a workstation for clinical application of digital tomosynthesis (DTS) for real‐time target localization. Specifically, to accelerate the reconstruction both digitally‐reconstructed radiograph (DRR) and DTS using ultra‐fast interpolation capability of graphics hardware. Method and Materials: The developed workstation for real‐time target localization using DTS technology includes image acquisition, image reconstruction, and image comparison. A critical step to implement this technology is its reconstruction speed, especially real‐time generation of reference DTSs which require DRRs. The new generation graphics hardware provides a fast implementation of image interpolation and allows the bi‐linear/tri‐linear interpolation and the other image operations to be accomplished within second. We implemented classical ray‐casting algorithm for 2D DRR and filtered back‐projection (FBP) algorithm for DTS reconstruction on graphics hardware based on OpenGL programming interface. Specifically, planning CT images and on‐board projection images were loaded into the memory of hardware and saved as textures which were projected onto frame‐buffer according to the given geometry of imaging system. The final images were generated by blending multiple textures projected onto the frame‐buffer. Difference images and correlation coefficients between hardware and software DRRs and DTSs were calculated to assess their similarity. Results: The mean errors and their standard deviations of the differences DRR and DTS images are 1.0% ± 1.5% and 1.2% ± 2.1%. The correlation coefficients are 99.9% for DRRs and 99.5% for DTS. The reconstructed efficiency was improved by a factor of 67 for DRR and 13 for DTS. Conclusions: The hardware method provides fast reconstructions of DRR and DTS with visibility of anatomical structures comparable to those generated using software methods.This allows real‐time target localization for image registration between on‐board and reference DTSs in the application platform.

Gating Verification via Correlation of Dynamic DRRs and Respiration-Correlated MV Fluoroscopic Images

BackgroundBy integrating Digitally Reconstructed Radiograph (DRR) images of pulmonary tumors with Electronic Portal Imaging Device (EPID) images to assist in target segmentation, and subsequently comparing morphological changes in segmented targets across different radiotherapy stages, this approach enables precise quantification of dynamic variations in target volume and shape. This methodological integration provides objective evidence for treatment response evaluation and dynamic optimization of treatment plans, thereby significantly enhancing the precision of radiotherapy delivery.MethodsThe proposed multimodal segmentation framework, named EPIDSeg-Net, comprises an encoder, a multi-scale feature layer, and a decoder. The encoder utilizes a dual-branch architecture: a CNN branch for extracting local texture features and a Swin-Transformer branch for capturing global semantic features. The model first calibrates multimodal input features through a Dual Attention Mechanism (DAM) to adaptively adjust modality-specific weights, thereby enhancing tolerance to missing image information in multi-sequence segmentation. Subsequently, two key modules are implemented within the multi-scale feature layer: a Large-Kernel Grouped Attention Gating (LKG-Gate) module to strengthen local contextual awareness, and a Multi-Path Feature Extraction (MPFE) module to improve feature robustness via a parallel structure. These designs enable the model to effectively focus on lung tumor target regions, optimize segmentation accuracy, and achieve high-performance reconstruction.ResultsThe framework effectively integrates multimodal features, enabling high-precision localization and sharp boundary delineation while preserving anatomical details. Quantitative evaluations demonstrate superior performance: DICE = 93.2 (92.4∼93.9), CE = 0.352, HD95 = 9.42 (6.03∼12.8), IOU = 86.0 (84.1∼87.9), and SENCE = 0.828. Overall, the model excels at preserving gradient information, regional integrity, and fine details; effectively suppresses feature loss; and reduces missed segmentation rates, leading to improvements in both subjective and objective performance metrics.ConclusionThe proposed segmentation method effectively integrates information from EPID and DRR images, enabling more precise localization and segmentation of lesion regions within EPID images while enhancing segmentation accuracy.

For pancreatic cancer patients, image guided radiation therapy and real-time tumor tracking (RTTT) techniques can deliver radiation to the target accurately. Currently, for the radiation therapy machine with kV X-ray imaging systems, internal markers must be implemented to facilitate tumor tracking. The purpose of this study was to develop a markerless deep learning-based pancreatic tumor positioning procedure for real-time tumor tracking with a kV X-ray imaging system. Fourteen pancreatic cancer patients treated with intensity-modulated radiation therapy from six fixed gantry angles with a gimbal-head radiotherapy system were included in this study. For a gimbal-head radiotherapy system, the three-dimensional (3D) intrafraction target position can be determined using an orthogonal kV X-ray imaging system. All patients underwent four-dimensional computed tomography (4DCT) simulations for treatment planning, which were divided into 10 respiratory phases. After a patient's 4DCT was acquired, for each X-ray tube angle, 10 digitally reconstructed radiograph (DRR) images were obtained. Then, a data augmentation procedure was conducted. The data augmentation procedure first rotated the CT volume around the superior-inferior and anterior-posterior directions from -3° to 3° in 1.5° intervals. Then, the Super-SloMo model was adapted to interpolate 10 frames between respiratory phases. In total, the data augmentation procedure expanded the data scale 250-fold. In this study, for each patient, 12 datasets containing the DRR images from each specific X-ray tube angle based on the radiation therapy plan were obtained. The augmented dataset was randomly divided into training and testing datasets. The training dataset contained 2000 DRR images with clinical target volume (CTV) contours labeled for fine-tuning the pre-trained target contour prediction model. After the fine-tuning, the patient and X-ray tube angle-specific CTV contour prediction model was acquired. The testing dataset contained the remaining 500 images to evaluate the performance of the CTV contour prediction model. The dice similarity coefficient (DSC) between the area enclosed by the CTV contour and predicted contour was calculated to evaluate the model's contour prediction performance. The 3D position of the CTV was calculated based on the centroid of the contour in the orthogonal DRR images, and the 3D error of the prediction position was calculated to evaluate the CTV positioning performance. For each patient, the DSC results from 12 X-ray tube angles and 3D error from 6 gantry angles were calculated, representing the novelty of this study. The mean and standard deviation (SD) of all patients' DSCs were 0.98 and 0.015, respectively. The mean and SD of the 3D error were 0.29mm and 0.14mm, respectively. The global maximum 3D error was 1.66mm, and the global minimum DSC was 0.81. The mean calculation time for CTV contour prediction was 55ms per image. This fulfills the requirement of RTTT. Regarding the positioning accuracy and calculation efficiency, the presented procedure can provide a solution for markerless real-time tumor tracking for pancreatic cancer patients.

Transfer learning from an artificial radiograph-landmark dataset for registration of the anatomic skull model to dual fluoroscopic X-ray images

Radiation therapy of prostate cancer requires a high geometric accuracy of dose delivery. Unfortunately, this accuracy can be hardly obtained because of imperfect patients repositioning throughout the preparation and execution of the treatment. In order to correct for such errors we developed methods for online registration of portal images (PI) and digitally reconstructed radiographs (DRR). To improve the preconditions for registration of the Pis and DRRs image preprocessing, such as contrast-limited adaptive histogram equalization (CLAHE) of the Pis, noise and edge filtering of the Pis and the DRRs, has been performed. Then, similarity measures as mean square error, cross correlation and mutual information were investigated. A fully automatic bony-based registration method has been developed. As a reference method a semi-automatic stereometric fiducial-based registration has also been developed. This method uses gold seeds implanted into the prostate as internal fiducial markers for assessment of the similarity measures. Small lead spheres as well were adhered as external fiducial markers on the Alderson phantom. This method was also used for the verification of the automatic bony-based registration. The registration of repositioning was then performed fully automatically in less than 30 s with the Alderson phantom and with the patient data. At the present stage, there remains a registration error of less than 2 mm.

Purpose: To facilitate image‐guided stereotactic body radiotherapy (SBRT) with high precision and fully automated target localization and patient position verification. Method and Materials: A mutual information based image registration algorithm was developed to register the patient's daily 3D computed tomography (CT) images with those taken for treatment planning to assess and correct patient position shifts in all three translational and three rotational directions. A digitally reconstructed radiograph (DRR) algorithm generates and displays the corresponding DRRs in one frame in a split screen format along with their intensity differences to enhance verification. In the same way, the portal image verification algorithm processes the portal image and displays it with the corresponding planning DRR in the same frame as final verification of the patient's position. The algorithms were programmed using MatLab and were assembled seamlessly in one package and the operation was fully automated. Results: Target localization and patient position verification process in our image‐guided SBRT is significantly simplified and the time needed for this process is reduced from 30 minutes or more to 3–4 minutes. The image registration algorithm was found to be accurate to within 0.1 mm with a head and neck phantom. The DRR generation algorithm generates high quality DRRs with a spatial resolution of 0.289×0.289 mm2/pixel. It removes unrelated anatomical structures for a cleaner background in the DRR to help identify the target isocenter more easily and more accurately. The DRR comparison and portal image verification algorithms readily detect misalignments of less than 0.2 mm. Conclusion: Application of the algorithms in determining patient position shift in actual SBRT cases demonstrated that they were accurate, fast, and reliable. They serve as a useful tool for image‐guided radiotherapy.

We present a spectral-based method for the 2D/3D rigid registration of X-ray images to a CT scan. The method uses a Fourier-based representation to decompose the six rigid transformation parameters problem into a twoparameter out-of-plane rotation and a four-parameter in-plane transformation problems. Preoperatively, a set of Digitally Reconstructed Radiographs (DRRs) are generated offline from the CT in the expected in-plane location ranges of the fluoroscopic X-ray imaging devices. Each DRR is transformed into a imaging device in-plane invariant features space. Intraoperatively, a few 2D projections of the patient anatomy are acquired with an X-ray imaging device. Each projection is transformed into its in-plane invariant representation. The out-of-plane parameters are first computed by maximization of the Normalized Cross-Correlation between the invariant representations of the DRRs and the X-ray images. Then, the in-plane parameters are computed with the phase correlation method based on the Fourier-Mellin transform. Experimental results on publicly available data sets show that our method can robustly estimate the out-of-plane parameters with accuracy of 1.5° in less than 1sec for out-of-plane rotations of 10° or more, and perform the entire registration in less than 10secs.

BackgroundLower limb rotation at the time of imaging may affect the measurement of joint line angles using plain radiographs, potentially compromising measurement accuracy. Accurate joint line angle assessment is important for orthopaedic surgical planning and limb alignment. This study aimed to investigate changes in distal femoral and proximal tibial joint line angles in response to limb rotation and to evaluate the correlation between these changes using digitally reconstructed radiographs (DRRs) generated from computed tomography (CT) images.MethodsPreoperative CT data from 50 knees scheduled for total knee arthroplasty (TKA) or unicompartmental knee arthroplasty (UKA) at our institution were analysed using TKA planning software. The femur and tibia were aligned perpendicular to their mechanical axes in the coronal and sagittal planes. The surgical epicondylar axis (SEA) and Akagi’s line were used as reference axes for femoral and tibial rotation, respectively, with 0° defined as neutral rotation. Each bone was rotated from 20° external rotation (ER) to 20° internal rotation (IR) in 5° increments using the software, and DRR images were generated at each position. The lateral distal femoral angle (LDFA), medial proximal tibial angle (MPTA), and posterior tibial slope (PTS) were measured at each rotational angle. The absolute values, their variations with rotation, and the correlations between each angle and its respective changes were analysed.ResultsAt 0°, the mean ± standard deviation values were 87.5 ± 2.4°, 83.9 ± 2.6°, and 8.4 ± 3.2° for LDFA, MPTA, and PTS, respectively. The mean changes across the range of 20° ER to 20° IR were 0.3° (1.8°–0.6°), 1.1° (1.0°–1.5°), and 6.2° (1.9°–2.4°) for LDFA, MPTA, and PTS, respectively. MPTA was negatively correlated with ΔPTS (r = -0.7). Based on MPTA values, patients were categorised into three groups: <82.5°, 82.5–84.5°, and > 84.5°. ΔPTS was significantly different between the < 82.5° (8.6 ± 2.3°) and 82.5–84.5° (6.9 ± 1.6°) groups, and between the 82.5–84.5° and > 84.5° (4.0 ± 1.9°) groups (P < 0.05 and P < 0.001, respectively).ConclusionsLimb rotation significantly affects PTS, particularly in cases with greater medial inclination of the tibial plateau. Caution should be exercised when using plain radiographs for preoperative planning or postoperative evaluation.

Cluster labelling is a technique which provides useful information about the cluster to the end users. In this paper, we propose a novel approach which is the follow-up of our previous work. Our earlier approach generates clusters of web documents by using a modified apriori approach which is more efficient and faster than the traditional apriori approach. To label the clusters, the proposed approach used an effective feature selection technique which selects the top features of a cluster. Rather than labelling the cluster with 'bag of words', a concept driven mechanism has been developed which uses the Wikipedia that takes the top features of a cluster as input to generate the possible candidate labels. Mutual information (MI) score technique has been used for ranking the candidate labels and then the topmost candidates are considered as potential labels of a cluster. Experimental results on two benchmark datasets demonstrate the efficiency of our approach.