3-D Reconstruction of the Spine From Biplanar Radiographs Based on Contour Matching Using the Hough Transform

Strategies for generating synthetic computed tomography-like imaging from radiographs: A scoping review

BACKGROUND

Advancements in tomographic medical imaging have revolutionized diagnostics and treatment monitoring by offering detailed 3D visualization of internal structures. Despite the significant value of computed tomography (CT), challenges such as high radiation dosage and cost barriers limit its accessibility, especially in low- and middle-income countries. Recognizing the potential of radiographic imaging in reconstructing CT images, this scoping review aims to explore the emerging field of synthesizing 3D CT-like images from 2D radiographs by examining the current methodologies.

METHODS

A scoping review was carried out following PRISMA-SR guidelines. Eligibility criteria for the articles included full-text articles published up to September 9, 2024, studying methodologies for the synthesis of 3D CT images from 2D biplanar or four-projection x-ray images. Eligible articles were sourced from PubMed MEDLINE, Embase, and arXiv.

RESULTS

76 studies were included. The majority (50.8 %, n = 30) were published between 2010 and 2020 (38.2 %, n = 29) and from 2020 onwards (36.8 %, n = 28), with European (40.8 %, n = 31), North American (26.3 %, n = 20), and Asian (32.9 %, n = 25) institutions being primary contributors. Anatomical regions varied, with 17.1 % (n = 13) of studies not using clinical data. Further, studies focused on the chest (25 %, n = 19), spine and vertebrae (17.1 %, n = 13), coronary arteries (10.5 %, n = 8), and cranial structures (10.5 %, n = 8), among other anatomical regions. Convolutional neural networks (CNN) (19.7 %, n = 15), generative adversarial networks (21.1 %, n = 16) and statistical shape models (15.8 %, n = 12) emerged as the most applied methodologies. A limited number of studies included explored the use of conditional diffusion models, iterative reconstruction algorithms, statistical shape models, and digital tomosynthesis.

CONCLUSION

This scoping review summarizes current strategies and challenges in synthetic imaging generation. The development of 3D CT-like imaging from 2D radiographs could reduce radiation risk while simultaneously addressing financial and logistical obstacles that impede global access to CT imaging. Despite initial promising results, the field encounters challenges with varied methodologies and frequent lack of proper validation, requiring further research to define synthetic imaging's clinical role.

Anatomical prior-based vertebral landmark detection for spinal disorder diagnosis

As one of fundamental ways to interpret spine images, detection of vertebral landmarks is an informative prerequisite for further diagnosis and management of spine disorders such as scoliosis and fractures. Most existing machine learning-based methods for automatic vertebral landmark detection suffer from overlapping landmarks or abnormally long distances between nearby landmarks against anatomical priors, and thus lack sufficient reliability and interpretability. To tackle the problem, this paper systematically utilizes anatomical prior knowledge in vertebral landmark detection. We explicitly formulate anatomical priors of the spine, related to distances among vertebrae and spatial order within the spine, and integrate these geometrical constraints within training loss, inference procedure, and evaluation metrics. First, we introduce an anatomy-constraint loss to regularize the training process with the aforementioned contextual priors explicitly. Second, we propose a simple-yet-effective anatomy-aided inference procedure by employing sequential prediction rather than a parallel counterpart. Third, we provide novel anatomy-related metrics to quantitatively evaluate to which extent landmark predictions follow the anatomical priors, as is not reflected within the widely-used landmark localization error metric. We employ the localization framework on 1410 anterior-posterior radiographic images. Compared with competitive baseline models, we achieve superior landmark localization accuracy and comparable Cobb angle estimation for scoliosis assessment. Ablation studies demonstrate the effectiveness of designed components on the decrease of localization error and improvement of anatomical plausibility. Additionally, we exhibit effective generalization performance by transferring our detection method onto sagittal 2-D slices of CT scans and boost the performance of downstream compression fracture classification at vertebra-level.

Quantifying Bone and Skin Movement in the Residual Limb-Socket Interface of Individuals With Transtibial Limb Loss Using Dynamic Stereo X-Ray: Protocol for a Lower Limb Loss Cadaver and Clinical Study

BACKGROUND

Relative motion between the residual limb and socket in individuals with transtibial limb loss can lead to substantial consequences that limit mobility. Although assessments of the relative motion between the residual limb and socket have been performed, there remains a substantial gap in understanding the complex mechanics of the residual limb-socket interface during dynamic activities that limits the ability to improve socket design. However, dynamic stereo x-ray (DSX) is an advanced imaging technology that can quantify 3D bone movement and skin deformation inside a socket during dynamic activities.

OBJECTIVE

This study aims to develop analytical tools using DSX to quantify the dynamic, in vivo kinematics between the residual limb and socket and the mechanism of residual tissue deformation.

METHODS

A lower limb cadaver study will first be performed to optimize the placement of an array of radiopaque beads and markers on the socket, liner, and skin to simultaneously assess dynamic tibial movement and residual tissue and liner deformation. Five cadaver limbs will be used in an iterative process to develop an optimal marker setup. Stance phase gait will be simulated during each session to induce bone movement and skin and liner deformation. The number, shape, size, and placement of each marker will be evaluated after each session to refine the marker set. Once an optimal marker setup is identified, 21 participants with transtibial limb loss will be fitted with a socket capable of being suspended via both elevated vacuum and traditional suction. Participants will undergo a 4-week acclimation period and then be tested in the DSX system to track tibial, skin, and liner motion under both suspension techniques during 3 activities: treadmill walking at a self-selected speed, at a walking speed 10% faster, and during a step-down movement. The performance of the 2 suspension techniques will be evaluated by quantifying the 3D bone movement of the residual tibia with respect to the socket and quantifying liner and skin deformation at the socket-residuum interface.

RESULTS

This study was funded in October 2021. Cadaver testing began in January 2023. Enrollment began in February 2024. Data collection is expected to conclude in December 2025. The initial dissemination of results is expected in November 2026.

CONCLUSIONS

The successful completion of this study will help develop analytical methods for the accurate assessment of residual limb-socket motion. The results will significantly advance the understanding of the complex biomechanical interactions between the residual limb and the socket, which can aid in evidence-based clinical practice and socket prescription guidelines. This critical foundational information can aid in the development of future socket technology that has the potential to reduce secondary comorbidities that result from complications of poor prosthesis load transmission.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

DERR1-10.2196/57329.

TomoRay: Generating Synthetic Computed Tomography of the Spine From Biplanar Radiographs

OBJECTIVE

Computed tomography (CT) imaging is a cornerstone in the assessment of patients with spinal trauma and in the planning of spinal interventions. However, CT studies are associated with logistical problems, acquisition costs, and radiation exposure. In this proof-of-concept study, the feasibility of generating synthetic spinal CT images using biplanar radiographs was explored. This could expand the potential applications of x-ray machines pre-, post-, and even intraoperatively.

METHODS

A cohort of 209 patients who underwent spinal CT imaging from the VerSe2020 dataset was used to train the algorithm. The model was subsequently evaluated using an internal and external validation set containing 55 from the VerSe2020 dataset and a subset of 56 images from the CTSpine1K dataset, respectively. Digitally reconstructed radiographs served as input for training and evaluation of the 2-dimensional (2D)-to-3-dimentional (3D) generative adversarial model. Model performance was assessed using peak signal to noise ratio (PSNR), structural similarity index (SSIM), and cosine similarity (CS).

RESULTS

At external validation, the developed model achieved a PSNR of 21.139 ± 1.018 dB (mean ± standard deviation). The SSIM and CS amounted to 0.947 ± 0.010 and 0.671 ± 0.691, respectively.

CONCLUSION

Generating an artificial 3D output from 2D imaging is challenging, especially for spinal imaging, where x-rays are known to deliver insufficient information frequently. Although the synthetic CT scans derived from our model do not perfectly match their ground truth CT, our proof-of-concept study warrants further exploration of the potential of this technology.

Three-dimensional spine reconstruction from biplane radiographs using convolutional neural networks

PURPOSE

The purpose of this study was to develop and evaluate a deep learning network for three-dimensional reconstruction of the spine from biplanar radiographs.

METHODS

The proposed approach focused on extracting similar features and multiscale features of bone tissue in biplanar radiographs. Bone tissue features were reconstructed for feature representation across dimensions to generate three-dimensional volumes. The number of feature mappings was gradually reduced in the reconstruction to transform the high-dimensional features into the three-dimensional image domain. We produced and made eight public datasets to train and test the proposed network. Two evaluation metrics were proposed and combined with four classical evaluation metrics to measure the performance of the method.

RESULTS

In comparative experiments, the reconstruction results of this method achieved a Hausdorff distance of 1.85 mm, a surface overlap of 0.2 mm, a volume overlap of 0.9664, and an offset distance of only 0.21 mm from the vertebral body centroid. The results of this study indicate that the proposed method is reliable.

Comparison of 3D Bone Position Estimation Using QR Code and Metal Bead Markers

To improve the accuracy of a 3D bone position estimation system that displays 3D images in response to changes in the position of fluoroscopic images, modified markers using quick response (QR) codes were developed. The aims of this study were to assess the accuracy of the estimated bone position on 3D images with reference to QR code markers on fluoroscopic images and to compare its accuracy with metal bead markers. Bone positions were estimated from reference points on a fluoroscopic image compared with those on a 3D image. The positional relationships of QR code and metal bead markers on the fluoroscopic image were compared with those on the 3D image in order to establish whether a 3D image may be drawn by tracking positional changes in radius models. Differences were investigated by comparing the distance between markers on the fluoroscopic image and that on the 3D image, which was projected on the monitor. The error ratio, which was defined as the difference in the measurement between the fluoroscopic and 3D images divided by the fluoroscopic measurement, was compared between QR code and metal bead markers. Error ratios for the QR code markers were 5.0 ± 2.0%, 6.4 ± 7.6%, and 1.0 ± 0.8% in the anterior–posterior view, ulnar side lateral view, and posterior–anterior view, respectively. Error ratios for the metal bead markers were 1.3 ± 1.7%, 13.8 ± 14.5%, and 4.7 ± 5.7% in the anterior–posterior view, ulnar side lateral view, and posterior–anterior view, respectively. The error ratio for the metal bead markers was smaller in the initial position (p < 0.01). However, the error ratios for the QR code markers were smaller in the lateral position and the posterior–anterior position (p < 0.05). In QR code marker tracking, tracking was successful even with discontinuous images. The accuracy of a 3D bone position estimation was increased by using the QR code marker system. QR code marker tracking facilitates real-time comparisons of dynamic changes in preoperative 3D and intraoperative fluoroscopic images.

BX2S-Net: Learning to reconstruct 3D spinal structures from bi-planar X-ray images

Grasping good understanding of the weight-bearing spatial structure of the spine of a human subject in a standing position is critical for the treatment of spinal disorders. Such disorders are commonly diagnosed via 2D X-ray imaging of the human subject in a standing position. However, 3D reconstruction techniques based on bi-planar X-ray imaging can enable better exploration and analysis of the spinal structure. In particular, compared to earlier deformable modeling approaches, the recently-developed deep-learning-based 3D reconstruction methods exhibit higher efficiency and generalizability. But these methods usually employ simple architectures with 2D encoders and 3D decoders. Thus, these methods have several drawbacks, namely, the existence of a semantic gap between dimensionally-inconsistent feature maps, the difficulty of jointly handling multi-view inputs, and the information source limitations for the decoding process. In order to better assist clinicians and tackle these problems, we propose a novel convolutional neural network framework, which we call BX2S-Net, to effectively achieve 3D spine reconstruction based on bi-planar X-ray images. In particular, a dimensionally-consistent encoder-decoder architecture is designed in conjunction with a dimensionality enhancement method in order to reduce the semantic gap between feature maps and achieve information fusion for multi-view inputs. A feature-guided progressive decoding process is developed on the decoder side, where a full-scale feature attention guidance (FFAG) module is introduced to efficiently aggregate image features and guide the decoding process at each level. In addition, a class augmentation method and a spatially-weighted cross-entropy loss function are used for network training with improved reconstruction quality for the vertebral edge region. The experimental results demonstrate the effectiveness of our model in reconstructing high-quality 3D spinal structures from bi-planar X-ray images. The code is available at https://github.com/NBU-CVMI/bx2s-net.

An Experimental Study of a 3D Bone Position Estimation System Based on Fluoroscopic Images

To compare a 3D preoperative planning image and fluoroscopic image, a 3D bone position estimation system that displays 3D images in response to changes in the position of fluoroscopic images was developed. The objective of the present study was to evaluate the accuracy of the estimated position of 3D bone images with reference to fluoroscopic images. Bone positions were estimated from reference points on a fluoroscopic image compared with those on a 3D image. The four reference markers positional relationships on the fluoroscopic image were compared with those on the 3D image to evaluate whether a 3D image may be drawn by tracking positional changes in the radius model. Intra-class correlations coefficients for reference marker distances between the fluoroscopic image and 3D image were 0.98–0.99. Average differences between measured values on the fluoroscopic image and 3D bone image for each marker corresponding to the direction of the bone model were 1.1 ± 0.7 mm, 2.4 ± 1.8 mm, 1.4 ± 0.8 mm, and 2.0 ± 1.6 mm in the anterior-posterior view, ulnar side lateral view, posterior-anterior view, and radial side lateral view, respectively. Marker positions were more accurate in the anterior-posterior and posterior-anterior views than in the radial and ulnar side lateral views. This system helps in real-time comparison of dynamic changes in preoperative 3D and intraoperative fluoroscopy images.

Reconstruction of three-dimensional lumbar vertebrae from biplanar x-rays

Objective. Vertebrae models from computer tomographic (CT) imaging are extensively used in image-guided surgical systems to deliver percutaneous orthopaedic operations with minimum risks, but patients may be exposed to excess radiation from the pre-operative CT scans. Generating vertebrae models from intra-operative x-rays for image-guided systems can reduce radiation exposure to the patient, and the surgeons can acquire the vertebrae's relative positions during the operation; therefore, we proposed a lumbar vertebrae reconstruction method from biplanar x-rays.Approach. Non-stereo-corresponding vertebral landmarks on x-rays were identified as targets for deforming a set of template vertebrae; the deformation was formulated as a minimisation problem, and was solved using the augmented Lagrangian method. Mean surface errors between the models reconstructed using the proposed method and CT scans were measured to evaluate the reconstruction accuracy.Main results. The evaluation yielded mean errors of 1.27 mm and 1.50 mm inin vitroexperiments on normal vertebrae and pathological vertebrae, respectively; the outcomes were comparable to other template-based methods.Significance. The proposed method is a viable alternative to provide digital lumbar to be used in image-guided systems, where the models can be used as a visual reference in surgical planning and image-guided applications in operations where the reconstruction error is within the allowable surgical error.

Three Dimensional Reconstruction Models for Medical Modalities: A Comprehensive Investigation and Analysis

BACKGROUND

Image reconstruction is the mathematical process which converts the signals obtained from the scanning machine into an image. The reconstructed image plays a fundamental role in the planning of surgery and research in the medical field.

DISCUSSION

This paper introduces the first comprehensive survey of the literature about medical image reconstruction related to diseases, presenting a categorical study about the techniques and analyzing advantages and disadvantages of each technique. The images obtained by various imaging modalities like MRI, CT, CTA, Stereo radiography and Light field microscopy are included. A comparison on the basis of the reconstruction technique, Imaging Modality and Visualization, Disease, Metrics for 3D reconstruction accuracy, Dataset and Execution time, Evaluation of the technique is also performed.

CONCLUSION

The survey makes an assessment of the suitable reconstruction technique for an organ, draws general conclusions and discusses the future directions.

Toward Automated 3D Spine Reconstruction from Biplanar Radiographs Using CNN for Statistical Spine Model Fitting

To date, 3D spine reconstruction from biplanar radiographs involves intensive user supervision and semi-automated methods that are time-consuming and not effective in clinical routine. This paper proposes a new, fast, and automated 3D spine reconstruction method through which a realistic statistical shape model of the spine is fitted to images using convolutional neural networks (CNN). The CNNs automatically detect the anatomical landmarks controlling the spine model deformation through a hierarchical and gradual iterative process. The performance assessment used a set of 68 biplanar radiographs, composed of both asymptomatic subjects and adolescent idiopathic scoliosis patients, in order to compare automated reconstructions with ground truths build using multiple experts-supervised reconstructions. The mean (SD) errors of landmark locations (3D Euclidean distances) were 1.6 (1.3) mm, 1.8 (1.3) mm, and 2.3 (1.4) mm for the vertebral body center, endplate centers, and pedicle centers, respectively. The clinical parameters extracted from the automated 3D reconstruction (reconstruction time is less than one minute) presented an absolute mean error between 2.8° and 4.7° for the main spinal parameters and between 1° and 2.1° for pelvic parameters. Automated and expert's agreement analysis reported that, on average, 89% of automated measurements were inside the expert's confidence intervals. The proposed automated 3D spine reconstruction method provides an important step that should help the dissemination and adoption of 3D measurements in clinical routine.

Improving Visibility of Stereo-Radiographic Spine Reconstruction with Geometric Inferences

Complex deformities of the spine, like scoliosis, are evaluated more precisely using stereo-radiographic 3D reconstruction techniques. Primarily, it uses six stereo-corresponding points available on the vertebral body for the 3D reconstruction of each vertebra. The wireframe structure obtained in this process has poor visualization, hence difficult to diagnose. In this paper, a novel method is proposed to improve the visibility of this wireframe structure using a deformation of a generic spine model in accordance with the 3D-reconstructed corresponding points. Then, the geometric inferences like vertebral orientations are automatically extracted from the radiographs to improve the visibility of the 3D model. Biplanar radiographs are acquired from five scoliotic subjects on a specifically designed calibration bench. The stereo-corresponding point reconstruction method is used to build six-point wireframe vertebral structures and thus the entire spine model. Using the 3D spine midline and automatically extracted vertebral orientation features, a more realistic 3D spine model is generated. To validate the method, the 3D spine model is back-projected on biplanar radiographs and the error difference is computed. Though, this difference is within the error limits available in the literature, the proposed work is simple and economical. The proposed method does not require more corresponding points and image features to improve the visibility of the model. Hence, it reduces the computational complexity. Expensive 3D digitizer and vertebral CT scan models are also excluded from this study. Thus, the visibility of stereo-corresponding point reconstruction is improved to obtain a low-cost spine model for a better diagnosis of spinal deformities.

3D–2D registration in mobile radiographs: algorithm development and preliminary clinical evaluation

An image-based 3D-2D registration method is presented using radiographs acquired in the uncalibrated, unconstrained geometry of mobile radiography. The approach extends a previous method for six degree-of-freedom (DOF) registration in C-arm fluoroscopy (namely 'LevelCheck') to solve the 9-DOF estimate of geometry in which the position of the source and detector are unconstrained. The method was implemented using a gradient correlation similarity metric and stochastic derivative-free optimization on a GPU. Development and evaluation were conducted in three steps. First, simulation studies were performed that involved a CT scan of an anthropomorphic body phantom and 1000 randomly generated digitally reconstructed radiographs in posterior-anterior and lateral views. A median projection distance error (PDE) of 0.007 mm was achieved with 9-DOF registration compared to 0.767 mm for 6-DOF. Second, cadaver studies were conducted using mobile radiographs acquired in three anatomical regions (thorax, abdomen and pelvis) and three levels of source-detector distance (~800, ~1000 and ~1200 mm). The 9-DOF method achieved a median PDE of 0.49 mm (compared to 2.53 mm for the 6-DOF method) and demonstrated robustness in the unconstrained imaging geometry. Finally, a retrospective clinical study was conducted with intraoperative radiographs of the spine exhibiting real anatomical deformation and image content mismatch (e.g. interventional devices in the radiograph that were not in the CT), demonstrating a PDE = 1.1 mm for the 9-DOF approach. Average computation time was 48.5 s, involving 687 701 function evaluations on average, compared to 18.2 s for the 6-DOF method. Despite the greater computational load, the 9-DOF method may offer a valuable tool for target localization (e.g. decision support in level counting) as well as safety and quality assurance checks at the conclusion of a procedure (e.g. overlay of planning data on the radiograph for verification of the surgical product) in a manner consistent with natural surgical workflow.

Real-scale 3D models of the scoliotic spine from biplanar radiography without calibration objects

This paper presents a new method for modelling the spines of subjects and making accurate 3D measurements using standard radiologic systems without requiring calibration objects. The method makes use of the focal distance and statistical models for estimating the geometrical parameters of the system. A dataset of 32 subjects was used to assess this method. The results show small errors for the main clinical indices, such as an RMS error of 0.49° for the Cobb angle, 0.50° for kyphosis, 0.38° for lordosis, and 2.62mm for the spinal length. This method is the first to achieve this level of accuracy without requiring the use of calibration objects when acquiring radiographs. We conclude that the proposed method allows for the evaluation of scoliosis with a much simpler setup than currently available methods.