Understanding current UK practice for the incidental identification of vertebral fragility fractures from CT scans: an expert elicitation study.

BACKGROUND: There is an emerging interest in using automated approaches to enable the incidental identification of vertebral fragility fractures (VFFs) on existing medical images visualising the spine. AIM: To quantify values, and the degree of uncertainty associated with them, for the incidental identification of VFFs from computed tomography (CT) scans in current practice. METHODS: An expert elicitation exercise was conducted to generate point estimates and measures of uncertainty for four values representing the probability of: VFF being correctly reported by the radiologist; the absence of VFF being correctly assessed by the radiologist; being referred for management when a VFF is identified; having a dual-energy X-ray absorptiometry (DXA) scan after general practitioner (GP) referral. Data from a sample of seven experts in the diagnosis and management of people with VFFs were pooled using mathematical aggregation. RESULTS: The estimated mean values for each probability parameter were: VFF being correctly reported by the radiologist = 0.25 (standard deviation (SD): 0.21); absence of VFF being correctly assessed by the radiologist = 0.89 (0.10); being referred for management when a VFF is identified by the radiologist = 0.15 (0.12); having a DXA scan after GP referral = 0.66 (0.28). DISCUSSION: These estimates could be used to facilitate the subsequent early economic evaluation of potential new approaches to improve the health outcomes of people with VFFs. CONCLUSION: In the absence of epidemiological studies, this study produced point estimates and measures of uncertainty for key parameters needed to describe current pathways for the incidental diagnosis of VFFs.

Measurement of synovial tissue volume in knee osteoarthritis using a semiautomated MRI-based quantitative approach.

PURPOSE: Synovitis is common in knee osteoarthritis and is associated with both knee pain and progression of disease. Semiautomated methods have been developed for quantitative assessment of structure in knee osteoarthritis. Our aims were to apply a novel semiautomated assessment method using 3D active appearance modeling for the quantification of synovial tissue volume (STV) and to compare its performance with conventional manual segmentation. METHODS: Thirty-two sagittal T1 -weighted fat-suppressed contrast-enhanced MRIs were assessed for STV by a single observer using 1) manual segmentation and 2) a semiautomated approach. We compared the STV analysis using the semiautomated and manual segmentation methods, including the time taken to complete the assessments. We also examined the reliability of STV assessment using the semiautomated method in a subset of 12 patients who had participated in a clinical trial of vitamin D therapy in knee osteoarthritis. RESULTS: There was no significant difference in STV using the semiautomated quantitative method compared to manual segmentation, mean difference = 207.2 mm3 (95% confidence interval -895.2 to 1309.7). The semiautomated method was significantly quicker than manual segmentation (18 vs. 71 min). For the semiautomated method, intraobserver agreement was excellent (intraclass correlation coefficient (3,1) = 0.99) and interobserver agreement was very good (intraclass correlation coefficient (3,1) = 0.83). CONCLUSION: We describe the application of a semiautomated method that is accurate, reliable, and quicker than manual segmentation for assessment of STV. The method may help increase efficiency of image assessment in large imaging studies and may also assist investigation of treatment efficacy in knee osteoarthritis.

Comparing image analysis approaches versus expert readers: the relation of knee radiograph features to knee pain.

OBJECTIVES: The relationship between radiographic evidence of osteoarthritis and knee pain has been weak. This may be because features that best discriminate knees with pain have not been included in analyses. We tested the correlation between knee pain and radiographic features taking into account both image analysis features and manual scores. METHODS: Using data of the Multicentre Osteoarthritis Study, we tested in a cross-sectional design how well X-ray features discriminated those with frequent knee pain (one question at one time) or consistent frequent knee pain (three questions at three times during the 2 weeks prior to imaging) from those without it. We trained random forest models on features from two radiographic views for classification. RESULTS: X-rays were better at classifying those with pain using three questions compared with one. When we used all manual radiographic features, the area under the curve (AUC) was 73.9%. Using the best model from automated image analyses or a combination of these and manual grades, no improvement over manual grading was found. CONCLUSIONS: X-ray changes of OA are more strongly associated with repeated reports of knee pain than pain reported once. In addition, a fully automated system that assessed features not scored on X-ray performed no better than manual grading of features.

Synovial tissue volume: a treatment target in knee osteoarthritis (OA).

BACKGROUND: Synovitis occurring frequently in osteoarthritis (OA) may be a targeted outcome. There are no data examining whether synovitis changes following intra-articular intervention. METHODS: Persons aged 40 years and older with painful knee OA participated in an open label trial of intra-articular steroid therapy. At all time points they completed the Knee Injury and Osteoarthritis Outcome Score (KOOS) questionnaire. They had a contrast-enhanced (CE) MRI immediately prior to an intra-articular steroid injection with a repeat scan within 20 days. Response status was assessed using the Osteoarthritis Research Society International (OARSI) response criteria. OARSI responders were followed until their pain relapsed either within 20% of baseline or 6 months, shortly after which a third MRI was performed. Synovial tissue volume (STV) was measured on postcontrast knee images. We looked at changes in the STV and in pain, and their association. RESULTS: 120 subjects with preinjection and postinjection CE MRI were followed. Their mean age was 62.3 years (SD=10.3) and 62 (52%) were women. The median time between injection and follow-up scan was 8 days (IQR 7-14 days). 85/120 (71%) were OARSI responders. Pain decreased (mean change in KOOS=+23.9; 95% CI 20.1 to 27.8, p<0.001) following steroid injection, as did mean STV (mean change=-1071 mm(3); 95% CI -1839 mm(3) to -303 mm(3), p=0.01). Of the 80 who returned for a third MRI, pain relapsed in 57, and in the 48 of those with MRI data, STV increased between follow-up and final visit (+1220 mm(3); 95% CI 25 mm(3) to 2414 mm(3), p=0.05). 23 were persistent responders at 6 months and, in these, STV did not increase (mean change=-202 mm(3); 95% CI -2008 mm(3) to 1604 mm(3), p=0.83). Controlling for variation over time, there was a significant association between synovitis volume and KOOS pain (b coefficient-change in KOOS pain score per 1000 mm(3) change in STV=-1.13; 95% CI -1.87 to -0.39, p=0.003), although STV accounted for only a small proportion of the variance in change in pain. CONCLUSIONS: Synovial tissue volume in knee OA shrinks following steroid therapy, and rebounds in those whose pain relapses. It can be considered a treatment target in symptomatic knee OA. TRIAL REGISTRATION NUMBER: ISRCTN07329370.

Statistical shape and appearance models in osteoporosis.

Statistical models (SMs) of shape (SSM) and appearance (SAM) have been acquiring popularity in medical image analysis since they were introduced in the early 1990s. They have been primarily used for segmentation, but they are also a powerful tool for 3D reconstruction and classification. All these tasks may be required in the osteoporosis domain, where fracture detection and risk estimation are key to reducing the mortality and/or morbidity of this bone disease. In this article, we review the different applications of SSMs and SAMs in the context of osteoporosis, and it concludes with a discussion of their advantages and disadvantages for this application.

Low-dose GBCA administration for brain tumour dynamic contrast enhanced MRI: a feasibility study.

A key limitation of current dynamic contrast enhanced (DCE) MRI techniques is the requirement for full-dose gadolinium-based contrast agent (GBCA) administration. The purpose of this feasibility study was to develop and assess a new low GBCA dose protocol for deriving high-spatial resolution kinetic parameters from brain DCE-MRI. Nineteen patients with intracranial skull base tumours were prospectively imaged at 1.5 T using a single-injection, fixed-volume low GBCA dose, dual temporal resolution interleaved DCE-MRI acquisition. The accuracy of kinetic parameters (ve, Ktrans, vp) derived using this new low GBCA dose technique was evaluated through both Monte-Carlo simulations (mean percent deviation, PD, of measured from true values) and an in vivo study incorporating comparison with a conventional full-dose GBCA protocol and correlation with histopathological data. The mean PD of data from the interleaved high-temporal-high-spatial resolution approach outperformed use of high-spatial, low temporal resolution datasets alone (p < 0.0001, t-test). Kinetic parameters derived using the low-dose interleaved protocol correlated significantly with parameters derived from a full-dose acquisition (p < 0.001) and demonstrated a significant association with tissue markers of microvessel density (p < 0.05). Our results suggest accurate high-spatial resolution kinetic parameter mapping is feasible with significantly reduced GBCA dose.

Cohort profile: Worldwide Collaboration on OsteoArthritis prediCtion for the Hip (World COACH) - an international consortium of prospective cohort studies with individual participant data on hip osteoarthritis.

PURPOSE: Hip osteoarthritis (OA) is a major cause of pain and disability worldwide. Lack of effective therapies may reflect poor knowledge on its aetiology and risk factors, and result in the management of end-stage hip OA with costly joint replacement. The Worldwide Collaboration on OsteoArthritis prediCtion for the Hip (World COACH) consortium was established to pool and harmonise individual participant data from prospective cohort studies. The consortium aims to better understand determinants and risk factors for the development and progression of hip OA, to optimise and automate methods for (imaging) analysis, and to develop a personalised prediction model for hip OA. PARTICIPANTS: World COACH aimed to include participants of prospective cohort studies with ≥200 participants, that have hip imaging data available from at least 2 time points at least 4 years apart. All individual participant data, including clinical data, imaging (data), biochemical markers, questionnaires and genetic data, were collected and pooled into a single, individual-level database. FINDINGS TO DATE: World COACH currently consists of 9 cohorts, with 38 021 participants aged 18-80 years at baseline. Overall, 71% of the participants were women and mean baseline age was 65.3±8.6 years. Over 34 000 participants had baseline pelvic radiographs available, and over 22 000 had an additional pelvic radiograph after 8-12 years of follow-up. Even longer radiographic follow-up (15-25 years) is available for over 6000 of these participants. FUTURE PLANS: The World COACH consortium offers unique opportunities for studies on the relationship between determinants/risk factors and the development or progression of hip OA, by using harmonised data on clinical findings, imaging, biomarkers, genetics and lifestyle. This provides a unique opportunity to develop a personalised hip OA risk prediction model and to optimise methods for imaging analysis of the hip.

A High-Dimensional Quantification of Mouse Defensive Behaviors Reveals Enhanced Diversity and Stimulus Specificity.

Instinctive defensive behaviors, consisting of stereotyped sequences of movements and postures, are an essential component of the mouse behavioral repertoire. Since defensive behaviors can be reliably triggered by threatening sensory stimuli, the selection of the most appropriate action depends on the stimulus property. However, since the mouse has a wide repertoire of motor actions, it is not clear which set of movements and postures represent the relevant action. So far, this has been empirically identified as a change in locomotion state. However, the extent to which locomotion alone captures the diversity of defensive behaviors and their sensory specificity is unknown. To tackle this problem, we developed a method to obtain a faithful 3D reconstruction of the mouse body that enabled to quantify a wide variety of motor actions. This higher dimensional description revealed that defensive behaviors are more stimulus specific than indicated by locomotion data. Thus, responses to distinct stimuli that were equivalent in terms of locomotion (e.g., freezing induced by looming and sound) could be discriminated along other dimensions. The enhanced stimulus specificity was explained by a surprising diversity. A clustering analysis revealed that distinct combinations of movements and postures, giving rise to at least 7 different behaviors, were required to account for stimulus specificity. Moreover, each stimulus evoked more than one behavior, revealing a robust one-to-many mapping between sensations and behaviors that was not apparent from locomotion data. Our results indicate that diversity and sensory specificity of mouse defensive behaviors unfold in a higher dimensional space, spanning multiple motor actions.

Modeling the shape and composition of the human body using dual energy X-ray absorptiometry images.

There is growing evidence that body shape and regional body composition are strong indicators of metabolic health. The purpose of this study was to develop statistical models that accurately describe holistic body shape, thickness, and leanness. We hypothesized that there are unique body shape features that are predictive of mortality beyond standard clinical measures. We developed algorithms to process whole-body dual-energy X-ray absorptiometry (DXA) scans into body thickness and leanness images. We performed statistical appearance modeling (SAM) and principal component analysis (PCA) to efficiently encode the variance of body shape, leanness, and thickness across sample of 400 older Americans from the Health ABC study. The sample included 200 cases and 200 controls based on 6-year mortality status, matched on sex, race and BMI. The final model contained 52 points outlining the torso, upper arms, thighs, and bony landmarks. Correlation analyses were performed on the PCA parameters to identify body shape features that vary across groups and with metabolic risk. Stepwise logistic regression was performed to identify sex and race, and predict mortality risk as a function of body shape parameters. These parameters are novel body composition features that uniquely identify body phenotypes of different groups and predict mortality risk. Three parameters from a SAM of body leanness and thickness accurately identified sex (training AUC = 0.99) and six accurately identified race (training AUC = 0.91) in the sample dataset. Three parameters from a SAM of only body thickness predicted mortality (training AUC = 0.66, validation AUC = 0.62). Further study is warranted to identify specific shape/composition features that predict other health outcomes.

Vertebral morphometry: semiautomatic determination of detailed shape from dual-energy X-ray absorptiometry images using active appearance models.

OBJECTIVES: Manual point placement for vertebral morphometry is time-consuming and imprecise. We evaluated the accuracy of semiautomatic computer determination of the detailed vertebral shape. MATERIALS AND METHODS: The shape and appearance of vertebrae on 250 lateral dual-energy x-ray absorptiometry (DXA) scans were statistically modeled using a sequence of active appearance models of vertebral triplets. The models were matched to unseen scans given an approximate initial location of the center of each vertebra. The segmentation accuracy was analyzed by fracture grade. RESULTS: Segmentation accuracy comparable to manual precision was obtained in the case of normal vertebrae, but the accuracy decreased with increasing fracture severity. We propose methods for improving the robustness for severe fractures. CONCLUSION: Vertebral morphometry measurements may be substantially automated even on noisy data with multiple fractures present. The shape and appearance parameters of the models could provide more powerful quantitative classifiers of osteoporotic vertebral fracture.

Investigation of association between hip osteoarthritis susceptibility loci and radiographic proximal femur shape.

OBJECTIVE: To test whether previously reported hip morphology or osteoarthritis (OA) susceptibility loci are associated with proximal femur shape as represented by statistical shape model (SSM) modes and as univariate or multivariate quantitative traits. METHODS: We used pelvic radiographs and genotype data from 929 subjects with unilateral hip OA who had been recruited previously for the Arthritis Research UK Osteoarthritis Genetics Consortium genome-wide association study. We built 3 SSMs capturing the shape variation of the OA-unaffected proximal femur in the entire mixed-sex cohort and for male/female-stratified cohorts. We selected 41 candidate single-nucleotide polymorphisms (SNPs) previously reported as being associated with hip morphology (for replication analysis) or OA (for discovery analysis) and for which genotype data were available. We performed 2 types of analysis for genotype-phenotype associations between these SNPs and the modes of the SSMs: 1) a univariate analysis using individual SSM modes and 2) a multivariate analysis using combinations of SSM modes. RESULTS: The univariate analysis identified association between rs4836732 (within the ASTN2 gene) and mode 5 of the female SSM (P = 0.0016) and between rs6976 (within the GLT8D1 gene) and mode 7 of the mixed-sex SSM (P = 0.0003). The multivariate analysis identified association between rs5009270 (near the IFRD1 gene) and a combination of modes 3, 4, and 9 of the mixed-sex SSM (P = 0.0004). Evidence of associations remained significant following adjustment for multiple testing. All 3 SNPs had previously been associated with hip OA. CONCLUSION: These de novo findings suggest that rs4836732, rs6976, and rs5009270 may contribute to hip OA susceptibility by altering proximal femur shape.

Using transmission electron microscopy and 3View to determine collagen fibril size and three-dimensional organization.

Collagen fibrils are the major tensile element in vertebrate tissues, in which they occur as ordered bundles in the extracellular matrix. Abnormal fibril assembly and organization results in scarring, fibrosis, poor wound healing and connective tissue diseases. Transmission electron microscopy (TEM) is used to assess the formation of the fibrils, predominantly by measuring fibril diameter. Here we describe a protocol for measuring fibril diameter as well as fibril volume fraction, mean fibril length, fibril cross-sectional shape and fibril 3D organization, all of which are major determinants of tissue function. Serial-section TEM (ssTEM) has been used to visualize fibril 3D organization in vivo. However, serial block face-scanning electron microscopy (SBF-SEM) has emerged as a time-efficient alternative to ssTEM. The protocol described below is suitable for preparing tissues for TEM and SBF-SEM (by 3View). We describe how to use 3View for studying collagen fibril organization in vivo and show how to find and track individual fibrils. The overall time scale is ~8 d from isolating the tissue to having a 3D image stack.

Automatic construction of parts+geometry models for initializing groupwise registration.

Groupwise nonrigid image registration is a powerful tool to automatically establish correspondences across sets of images. Such correspondences are widely used for constructing statistical models of shape and appearance. As existing techniques usually treat registration as an optimization problem, a good initialization is required. Although the standard initialization-affine transformation-generally works well, it is often inadequate when registering images of complex structures. In this paper we present a more sophisticated method that uses the sparse matches of a parts+geometry model as the initialization. We show that both the model and its matches can be automatically obtained, and that the matches are able to effectively initialize a groupwise nonrigid registration algorithm, leading to accurate dense correspondences. We also show that the dense mesh models constructed during the groupwise registration process can be used to accurately annotate new images. We demonstrate the efficacy of the approach on three datasets of increasing difficulty, and report on a detailed quantitative evaluation of its performance.

Initialising groupwise non-rigid registration using multiple parts+geometry models.

Groupwise non-rigid registration is an important technique in medical image analysis. Recent studies show that its accuracy can be greatly improved by explicitly providing good initialisation. This is achieved by seeking a sparse correspondence using a parts+geometry model. In this paper we show that a single parts+geometry model is unlikely to establish consistent sparse correspondence for complex objects, and that better initialisation can be achieved using a set of models. We describe how to combine the strengths of multiple models, and demonstrate that the method gives state-of-the-art performance on three datasets, with the most significant improvement on the most challenging.

Automatic location of vertebrae on DXA images using random forest regression.

We provide a fully automatic method of segmenting vertebrae in DXA images. This is of clinical relevance to the diagnosis of osteoporosis by vertebral fracture, and to grading fractures in clinical trials. In order to locate the vertebrae we train detectors for the upper and lower vertebral endplates. Each detector uses random forest regressor voting applied to Haar-like input features. The regressors are applied at a grid of points across the image, and each tree votes for an endplate centre position. Modes in the smoothed vote image are endplate candidates, some of which are the neighbouring vertebrae of the one sought. The ambiguity is resolved by applying geometric constraints to the connections between vertebrae, although there can be some ambiguity about where the sequence starts (e.g., is the lowest vertebra L4 or L5, fig 2a). The endplate centres are used to initialise a final phase of active appearance model search for a detailed solution. The method is applied to a dataset of 320 DXA images. Accuracy is comparable to manually initialised AAM segmentation in 91% of images, but multiple grade 3 fractures can cause some edge confusion in severely osteoporotic cases.

Accurate fully automatic femur segmentation in pelvic radiographs using regression voting.

Extraction of bone contours from radiographs plays an important role in disease diagnosis, pre-operative planning, and treatment analysis. We present a fully automatic method to accurately segment the proximal femur in anteroposterior pelvic radiographs. A number of candidate positions are produced by a global search with a detector. Each is then refined using a statistical shape model together with local detectors for each model point. Both global and local models use Random Forest regression to vote for the optimal positions, leading to robust and accurate results. The performance of the system is evaluated using a set of 519 images. We show that the fully automated system is able to achieve a mean point-to-curve error of less than 1 mm for 98% of all 519 images. To the best of our knowledge, this is the most accurate automatic method for segmenting the proximal femur in radiographs yet reported.

Automatic part selection for groupwise registration.

Groupwise non-rigid image registration plays an important role in medical image analysis. As local optimisation is largely used in such techniques, a good initialisation is required to avoid local minima. Although the traditional approach to initialisation--affine transformation--generally works well, recent studies have shown that it is inadequate when registering images of complex structures. In this paper we present a more sophisticated method that uses the sparse matches of a parts+geometry model as the initialisation. The choice of parts is made by a voting scheme. We generate a large number of candidate parts, randomly construct many different parts+geometry models and then use the models to select the parts with good localisability. We show that the algorithm can achieve better results than the state of the art on three different datasets of increasing difficulty. We also show that dense mesh models constructed during the groupwise registration process can be used to accurately annotate new images.

Automatic learning sparse correspondences for initialising groupwise registration.

We seek to automatically establish dense correspondences across groups of images. Existing non-rigid registration methods usually involve local optimisation and thus require accurate initialisation. It is difficult to obtain such initialisation for images of complex structures, especially those with many self-similar parts. In this paper we show that satisfactory initialisation for such images can be found by a parts+geometry model. We use a population based optimisation strategy to select the best parts from a large pool of candidates. The best matches of the optimal model are used to initialise a groupwise registration algorithm, leading to dense, accurate results. We demonstrate the efficacy of the approach on two challenging datasets, and report on a detailed quantitative evaluation of its performance.

Building 3-D statistical shape models by direct optimization.

Statistical shape models are powerful tools for image interpretation and shape analysis. A simple, yet effective, way of building such models is to capture the statistics of sampled point coordinates over a training set of example shapes. However, a major drawback of this approach is the need to establish a correspondence across the training set. In 2-D, a correspondence is often defined using a set of manually placed 'landmarks' and linear interpolation to sample the shape in between. Such annotation is, however, time-consuming and subjective, particularly when extended to 3-D. In this paper, we show that it is possible to establish a dense correspondence across the whole training set automatically by treating correspondence as an optimization problem. The objective function we use for the optimization is based on the minimum description length principle, which we argue is a criterion that leads to models with good compactness, specificity, and generalization ability. We manipulate correspondence by reparameterizing each training shape. We describe an explicit representation of reparameterization for surfaces in 3-D that makes it impossible to generate an illegal (i.e., not one-to-one) correspondence. We also describe several large-scale optimization strategies for model building, and perform a detailed analysis of each approach. Finally, we derive quantitative measures of model quality, allowing meaningful comparison between models built using different methods. Results are given for several different training sets of 3-D shapes, which show that the minimum description length models perform significantly better than other approaches.