Automatic Segmentation of Wrist Bones in CT Using a Statistical Wrist Shape + Pose Model

AFSegNet: few-shot 3D ankle-foot bone segmentation via hierarchical feature distillation and multi-scale attention and fusion

Accurate segmentation of ankle and foot bones from CT scans is essential for morphological analysis. Ankle and foot bone segmentation challenges due to the blurred bone boundaries, narrow inter-bone gaps, gaps in the cortical shell, and uneven spongy bone textures. Our study endeavors to create a deep learning framework that harnesses advantages of 3D deep learning and tackles the hurdles in accurately segmenting ankle and foot bones from clinical CT scans. A few-shot framework AFSegNet is proposed considering the computational cost, which comprises three 3D deep-learning networks adhering to the principles of progressing from simple to complex tasks and network structures. Specifically, a shallow network first over-segments the foreground, and along with the foreground ground truth are used to supervise a subsequent network to detect the over-segmented regions, which are overwhelmingly inter-bone gaps. The foreground and inter-bone gap probability map are then input into a network with multi-scale attentions and feature fusion, a loss function combining region-, boundary-, and topology-based terms to get the fine-level bone segmentation. AFSegNet is applied to the 16-class segmentation task utilizing 123 in-house CT scans, which only requires a GPU with 24 GB memory since the three sub-networks can be successively and individually trained. AFSegNet achieves a Dice of 0.953 and average surface distance of 0.207. The ablation study and comparison with two basic state-of-the-art networks indicates the effectiveness of the progressively distilled features, attention and feature fusion modules, and hybrid loss functions, with the mean surface distance error decreased up to 50 %.

Automatic segmentation and labelling of wrist bones in four-dimensional computed tomography datasets via deep learning

This study developed a deep learning model for fully automatic segmentation and labelling of wrist bones from four-dimensional computed tomography (4DCT) scans. This is a crucial step towards implementing 4DCT for diagnosing wrist ligament lesions, reducing time-consuming analysis of extensive data.

Three-dimensional reconstruction of the hand from biplanar X-rays: Assessment of accuracy and reliability

BACKGROUND

Functional disorders of the hand are generally investigated first using conventional radiographic imaging. However, X-rays (two-dimensional (2D)) provide limited information and the information may be reduced by overlapping bones and projection bias. This work presents a three-dimensional (3D) hand reconstruction method from biplanar X-rays.

METHOD

This approach consists of the deformation of a generic hand model on biplanar X-rays by manual and automatic processes. The reference examination being the manual CT segmentation, the precision of the method was evaluated by a comparison between the reconstructions from biplanar X-rays and the corresponding reconstructions from the CT scan (0.3mm section thickness). To assess the reproducibility of the method, 6 healthy hands (6 subjects, 3 left, 3 men) were considered. Two operators repeated each reconstruction from biplanar X-rays three times to study inter- and intra-operator variability. Three anatomical parameters that could be calculated automatically from the reconstructions were considered from the bone surfaces: the length of the scaphoid, the depth of the distal end of the radius and the height of the trapezius.

RESULTS

Double the root mean square error (2 Root Mean Square, 2RMS) at the point/area difference between biplanar X-rays and computed tomography reconstructions ranged from 0.46mm for the distal phalanges to 1.55mm for the bones of the distal carpals. The inter-intra-observer variability showed precision with a 95% confidence interval of less than 1.32mm for the anatomical parameters, and 2.12mm for the bone centroids.

DISCUSSION

The current method allows to obtain an accurate 3D reconstruction of the hand and wrist compared to the traditional segmented CT scan. By improving the automation of the method, objective information about the position of the bones in space could be obtained quickly. The value of this method lies in the early diagnosis of certain ligament pathologies (carpal instability) and it also has implications for surgical planning and personalized finite element modeling.

LEVEL OF PROOF

Basic sciences.

Dynamic multi feature-class Gaussian process models

In model-based medical image analysis, three relevant features are the shape of structures of interest, their relative pose, and image intensity profiles representative of some physical properties. Often, these features are modelled separately through statistical models by decomposing the object's features into a set of basis functions through principal geodesic analysis or principal component analysis. However, analysing articulated objects in an image using independent single object models may lead to large uncertainties and impingement, especially around organ boundaries. Questions that come to mind are the feasibility of building a unique model that combines all three features of interest in the same statistical space, and what advantages can be gained for image analysis. This study presents a statistical modelling method for automatic analysis of shape, pose and intensity features in medical images which we call the Dynamic multi feature-class Gaussian process models (DMFC-GPM). The DMFC-GPM is a Gaussian process (GP)-based model with a shared latent space that encodes linear and non-linear variations. Our method is defined in a continuous domain with a principled way to represent shape, pose and intensity feature-classes in a linear space, based on deformation fields. A deformation field-based metric is adapted in the method for modelling shape and intensity variation as well as for comparing rigid transformations (pose). Moreover, DMFC-GPMs inherit properties intrinsic to GPs including marginalisation and regression. Furthermore, they allow for adding additional pose variability on top of those obtained from the image acquisition process; what we term as permutation modelling. For image analysis tasks using DMFC-GPMs, we adapt Metropolis-Hastings algorithms making the prediction of features fully probabilistic. We validate the method using controlled synthetic data and we perform experiments on bone structures from CT images of the shoulder to illustrate the efficacy of the model for pose and shape prediction. The model performance results suggest that this new modelling paradigm is robust, accurate, accessible, and has potential applications in a multitude of scenarios including the management of musculoskeletal disorders, clinical decision making and image processing.

MsgeCNN: Multiscale geometric embedded convolutional neural network for ONFH segmentation and grading

BACKGROUND

The incidence of osteonecrosis of the femoral head (ONFH) is increasing gradually, rapid and accurate grading of ONFH is critical. The existing Steinberg staging criteria grades ONFH according to the proportion of necrosis area to femoral head area.

PURPOSE

In the clinical practice, the necrosis region and femoral head region are mainly estimated by the observation and experience of doctor. This paper proposes a two-stage segmentation and grading framework, which can be used to segment the femoral head and necrosis, as well as to diagnosis.

METHODS

The core of the proposed two-stage framework is the multiscale geometric embedded convolutional neural network (MsgeCNN), which integrates geometric information into the training process and accurately segments the femoral head region. Then, the necrosis regions are segmented by the adaptive threshold method taking femoral head as the background. The area and proportion of the two are calculated to determine the grade.

RESULTS

The accuracy of the proposed MsgeCNN for femoral head segmentation is 97.73%, sensitivity is 91.17%, specificity is 99.40%, dice score is 93.34%. And the segmentation performance is better than the existing five segmentation algorithms. The diagnostic accuracy of the overall framework is 90.80%.

CONCLUSIONS

The proposed framework can accurately segment the femoral head region and the necrosis region. The area, proportion, and other pathological information of the framework output provide auxiliary strategies for subsequent clinical treatment.

Association between bone shape and the presence of a fracture in patients with a clinically suspected scaphoid fracture

Scaphoid fractures are difficult to diagnose with current imaging modalities. It is unknown whether the shape of the scaphoid bone, assessed by statistical shape modeling, can be used to differentiate between fractured and non-fractured bones. Therefore, the aim of this study was to investigate whether the presence of a scaphoid fracture is associated with shape modes of a statistical shape model (SSM). Forty-one high-resolution peripheral quantitative computed tomography (HR-pQCT) scans were available from patients with a clinically suspected scaphoid fracture of whom 15 patients had a scaphoid fracture. The scans showed no motion artefacts affecting bone shape. The scaphoid bones were semi-automatically contoured, and the contours were converted to triangular meshes. The meshes were registered, followed by principal component analysis to determine mean shape and shape modes describing shape variance. The first five out of the forty shape modes cumulatively explained 87.8% of the shape variance. Logistic regression analysis was used to study the association between shape modes and fracture presence. The regression models were used to classify the 41 scaphoid bones as fractured or non-fractured using a cut-off value that maximized the sum of sensitivity and specificity. The classification of the models was compared with fracture diagnosis on HR-pQCT. A regression model with four shape modes had an area under the ROC-curve of 72.3% and correctly classified 75.6% of the scaphoid bones (fractured: 60.0%, non-fractured: 84.6%). To conclude, fracture presence in patients with a clinically suspected scaphoid fracture appears to be associated with the shape of the scaphoid bone.

A Novel Combined Level Set Model for Carpus Segmentation from Magnetic Resonance Images with Prior Knowledge aligned in Polar Coordinate System

BACKGROUND AND OBJECTIVE

Segmentation on carpus provides essential information for clinical applications including pathological evaluations, therapy planning, wrist biomechanical analysis, etc. Along with the acquisition procedure of magnetic resonance (MR) technique, poor quality of wrist images (e.g., occlusion, low signal-to-noise ratio, and contrast) often causes segmentation failure.

METHODS

In this work, to address such problems, a shape prior enhanced level set model was proposed. By transferring a shape contour in Cartesian Coordinate System (COS) into a curve in Polar Coordinate System (POS), parameters describing conventional shape invariance, i.e., translations, rotation, and scale were simplified into a single parameter for phase shift, which strongly improved algorithm efficiency. Given a training set in COS, a confidence interval representing the corresponding curves in POS was utilized as the shape prior set term in the model. Integrated with an edge detector, a local intensity descriptor, and a regularization term, the proposed method further possessed abilities against noise, intensity inhomogeneity as well as re-initialization problem. Images from 15 in-vivo acquired MR-datasets of the human wrist were used for validation. The performance of the proposed method has been compared with three state-of-the-art methods.

RESULTS

We reported a Dice Similarity Coefficient of 96.88±1.20%, a Relative Volume Difference of -1.53±3.01%, a Volume Overlap Error of 6.03±2.23%, a 95% Hausdorff Distance of 1.43±0.66 mm, an Average Symmetric Surface Distance of 0.50±0.17 mm, and a Root Mean Square Distance of 0.71±0.25 mm for the proposed method. The time consumption was 36.03±19.98 s.

CONCLUSIONS

Experimental results indicated that, compared with three other methods, the proposed method achieved significant improvement in terms of accuracy and efficiency.

Anatomical fitting of a plate shape directly derived from a 3D statistical bone model of the tibia

INTRODUCTION

Intra- and inter-population variations of bone morphology have made the process of designing an anatomically well-fitting fracture fixation plate challenging. Although statistical bone models have recently been used for analysing morphological variabilities, it is not known to what extent they would also provide the basis for the design of a new plate shape. This would be particularly valuable in the case where no existing plate shape is available to start the process of fit optimisation. Therefore, this study investigated the anatomical fitting of a plate shape (statistical plate) derived from the mean shape of a statistical 3D tibia bone model in comparison to results available from two other plate shapes.

METHODS

Forty-five 3D bone models of tibiae from Japanese cadaver specimens, as well as 3D models of the plate undersurface of both a commercial and shape optimised Medial Distal Tibia Plate, were utilised from earlier studies. The mean shape of the 3D statistical bone model was generated from the tibia models utilising the Statismo framework. With reverse engineering software, the plate undersurface of the statistical plate shape was derived directly from the mean surface of the statistical 3D bone model. Through an iterative process, the statistical plate model was placed at the correct surgical position on each bone model for fit assessment.

RESULTS

The statistical plate was fitting for 20% of the tibiae compared to 13% for the commercial and 67% for the optimised plate, respectively.

CONCLUSIONS

The plate shape derived directly from a statistical bone model was fitting better than the commercial plate, but considerably inferior to that of an optimised plate. However, the results do clearly indicate that this approach provides an appropriate and solid basis for commencing shape optimisation of the statistical plate. Studies of other anatomical regions are required to confirm whether these findings can be generalised.

A novel robust kernel principal component analysis for nonlinear statistical shape modeling from erroneous data

Statistical Shape Models (SSMs) have achieved considerable success in medical image segmentation. A high quality SSM is able to approximate the main plausible variances of a given anatomical structure to guide segmentation. However, it is technically challenging to derive such a quality model because: (1) the distribution of shape variance is often nonlinear or multi-modal which cannot be modeled by standard approaches assuming Gaussian distribution; (2) as the quality of annotations in training data usually varies, heavy corruption will degrade the quality of the model as a whole. In this work, these challenges are addressed by introducing a generic SSM that is able to model nonlinear distribution and is robust to outliers in training data. Without losing generality and assuming a sparsity in nonlinear distribution, a novel Robust Kernel Principal Component Analysis (RKPCA) for statistical shape modeling is proposed with the aim of constructing a low-rank nonlinear subspace where outliers are discarded. The proposed approach is validated on two different datasets: a set of 30 public CT kidney pairs and a set of 49 MRI ankle bones volumes. Experimental results demonstrate a significantly better performance on outlier recovery and a higher quality of the proposed model as well as lower segmentation errors compared to the state-of-the-art techniques.

Automatic Segmentation of Ulna and Radius in Forearm Radiographs

Automatic segmentation of ulna and radius (UR) in forearm radiographs is a necessary step for single X-ray absorptiometry bone mineral density measurement and diagnosis of osteoporosis. Accurate and robust segmentation of UR is difficult, given the variation in forearms between patients and the nonuniformity intensity in forearm radiographs. In this work, we proposed a practical automatic UR segmentation method through the dynamic programming (DP) algorithm to trace UR contours. Four seed points along four UR diaphysis edges are automatically located in the preprocessed radiographs. Then, the minimum cost paths in a cost map are traced from the seed points through the DP algorithm as UR edges and are merged as the UR contours. The proposed method is quantitatively evaluated using 37 forearm radiographs with manual segmentation results, including 22 normal-exposure and 15 low-exposure radiographs. The average Dice similarity coefficient of our method reached 0.945. The average mean absolute distance between the contours extracted by our method and a radiologist is only 5.04 pixels. The segmentation performance of our method between the normal- and low-exposure radiographs was insignificantly different. Our method was also validated on 105 forearm radiographs acquired under various imaging conditions from several hospitals. The results demonstrated that our method was fairly robust for forearm radiographs of various qualities.

Reconstructing a 3D heart surface with stereo-endoscope by learning eigen-shapes

An efficient approach to dynamically reconstruct a region of interest (ROI) on a beating heart from stereo-endoscopic video is developed. A ROI is first pre-reconstructed with a decoupled high-rank thin plate spline model. Eigen-shapes are learned from the pre-reconstructed data by using principal component analysis (PCA) to build a low-rank statistical deformable model for reconstructing subsequent frames. The linear transferability of PCA is proved, which allows fast eigen-shape learning. A general dynamic reconstruction framework is developed that formulates ROI reconstruction as an optimization problem of model parameters, and an efficient second-order minimization algorithm is derived to iteratively solve it. The performance of the proposed method is finally validated on stereo-endoscopic videos recorded by da Vinci robots.

3D surface voxel tracing corrector for accurate bone segmentation

PURPOSE

For extremely close bones, their boundaries are weak and diffused due to strong interaction between adjacent surfaces. These factors prevent the accurate segmentation of bone structure. To alleviate these difficulties, we propose an automatic method for accurate bone segmentation. The method is based on a consideration of the 3D surface normal direction, which is used to detect the bone boundary in 3D CT images.

METHODS

Our segmentation method is divided into three main stages. Firstly, we consider a surface tracing corrector combined with Gaussian standard deviation [Formula: see text] to improve the estimation of normal direction. Secondly, we determine an optimal value of [Formula: see text] for each surface point during this normal direction correction. Thirdly, we construct the 1D signal and refining the rough boundary along the corrected normal direction. The value of [Formula: see text] is used in the first directional derivative of the Gaussian to refine the location of the edge point along accurate normal direction. Because the normal direction is corrected and the value of [Formula: see text] is optimized, our method is robust to noise images and narrow joint space caused by joint degeneration.

RESULTS

We applied our method to 15 wrists and 50 hip joints for evaluation. In the wrist segmentation, Dice overlap coefficient (DOC) of [Formula: see text]% was obtained by our method. In the hip segmentation, fivefold cross-validations were performed for two state-of-the-art methods. Forty hip joints were used for training in two state-of-the-art methods, 10 hip joints were used for testing and performing comparisons. The DOCs of [Formula: see text], [Formula: see text]%, and [Formula: see text]% were achieved by our method for the pelvis, the left femoral head and the right femoral head, respectively.

CONCLUSION

Our method was shown to improve segmentation accuracy for several specific challenging cases. The results demonstrate that our approach achieved a superior accuracy over two state-of-the-art methods.

WRIST: A WRist Image Segmentation Toolkit for carpal bone delineation from MRI

Segmentation of the carpal bones from 3D imaging modalities, such as magnetic resonance imaging (MRI), is commonly performed for in vivo analysis of wrist morphology, kinematics, and biomechanics. This crucial task is typically carried out manually and is labor intensive, time consuming, subject to high inter- and intra-observer variability, and may result in topologically incorrect surfaces. We present a method, WRist Image Segmentation Toolkit (WRIST), for 3D semi-automated, rapid segmentation of the carpal bones of the wrist from MRI. In our method, the boundary of the bones were iteratively found using prior known anatomical constraints and a shape-detection level set. The parameters of the method were optimized using a training dataset of 48 manually segmented carpal bones and evaluated on 112 carpal bones which included both healthy participants without known wrist conditions and participants with thumb basilar osteoarthritis (OA). Manual segmentation by two expert human observers was considered as a reference. On the healthy subject dataset we obtained a Dice overlap of 93.0 ± 3.8, Jaccard Index of 87.3 ± 6.2, and a Hausdorff distance of 2.7 ± 3.4 mm, while on the OA dataset we obtained a Dice overlap of 90.7 ± 8.6, Jaccard Index of 83.0 ± 10.6, and a Hausdorff distance of 4.0 ± 4.4 mm. The short computational time of 20.8 s per bone (or 5.1 s per bone in the parallelized version) and the high agreement with the expert observers gives WRIST the potential to be utilized in musculoskeletal research.

Positioning error of custom 3D-printed surgical guides for the radius: influence of fitting location and guide design

PURPOSE

Utilization of 3D-printed patient-specific surgical guides is a promising navigation approach for orthopedic surgery. However, navigation errors can arise if the guide is not correctly positioned at the planned bone location, compromising the surgical outcome. Quantitative measurements of guide positioning errors are rarely reported and have never been related to guide design and underlying bone anatomy. In this study, the positioning accuracy of a standard and an extended guide design with lateral extension is evaluated at different fitting locations (distal, mid-shaft and proximal) on the volar side of the radius.

METHODS

Four operators placed the surgical guides on 3D-printed radius models obtained from the CT scans of six patients. For each radius model, every operator positioned two guide designs on the three fitting locations. The residual positioning error was quantified with a CT-based image analysis method in terms of the mean target registration error (mTRE), total translation error ([Formula: see text]) and total rotation error ([Formula: see text]) by comparing the actual guide position with the preoperatively planned position. Three generalized linear regression models were constructed to evaluate if the fitting location and the guide design affected mTRE, [Formula: see text] and [Formula: see text].

RESULTS

mTRE, [Formula: see text] and [Formula: see text] were significantly higher for mid-shaft guides ([Formula: see text]) compared to distal guides. The guide extension significantly improved the target registration and translational accuracy in all the volar radius locations ([Formula: see text]). However, in the mid-shaft region, the guide extension yielded an increased total rotational error ([Formula: see text]).

CONCLUSION

Our study demonstrates that positioning accuracy depends on the fitting location and on the guide design. In distal and proximal radial regions, the accuracy of guides with lateral extension is higher than standard guides and is therefore recommended for future use.

Liver segmentation from CT images using a sparse priori statistical shape model (SP-SSM)

This study proposes a new liver segmentation method based on a sparse a priori statistical shape model (SP-SSM). First, mark points are selected in the liver a priori model and the original image. Then, the a priori shape and its mark points are used to obtain a dictionary for the liver boundary information. Second, the sparse coefficient is calculated based on the correspondence between mark points in the original image and those in the a priori model, and then the sparse statistical model is established by combining the sparse coefficients and the dictionary. Finally, the intensity energy and boundary energy models are built based on the intensity information and the specific boundary information of the original image. Then, the sparse matching constraint model is established based on the sparse coding theory. These models jointly drive the iterative deformation of the sparse statistical model to approximate and accurately extract the liver boundaries. This method can solve the problems of deformation model initialization and a priori method accuracy using the sparse dictionary. The SP-SSM can achieve a mean overlap error of 4.8% and a mean volume difference of 1.8%, whereas the average symmetric surface distance and the root mean square symmetric surface distance can reach 0.8 mm and 1.4 mm, respectively.

Automated Bone Segmentation and Surface Evaluation of a Small Animal Model of Post-Traumatic Osteoarthritis

MicroCT imaging allows for noninvasive microstructural evaluation of mineralized bone tissue, and is essential in studies of small animal models of bone and joint diseases. Automatic segmentation and evaluation of articular surfaces is challenging. Here, we present a novel method to create knee joint surface models, for the evaluation of PTOA-related joint changes in the rat using an atlas-based diffeomorphic registration to automatically isolate bone from surrounding tissues. As validation, two independent raters manually segment datasets and the resulting segmentations were compared to our novel automatic segmentation process. Data were evaluated using label map volumes, overlap metrics, Euclidean distance mapping, and a time trial. Intraclass correlation coefficients were calculated to compare methods, and were greater than 0.90. Total overlap, union overlap, and mean overlap were calculated to compare the automatic and manual methods and ranged from 0.85 to 0.99. A Euclidean distance comparison was also performed and showed no measurable difference between manual and automatic segmentations. Furthermore, our new method was 18 times faster than manual segmentation. Overall, this study describes a reliable, accurate, and automatic segmentation method for mineralized knee structures from microCT images, and will allow for efficient assessment of bony changes in small animal models of PTOA.