2D/3D ultrasound diagnosis of pediatric distal radius fractures by human readers vs artificial intelligence.

Wrist trauma is common in children and generally requires radiography for exclusion of fractures, subjecting children to radiation and long wait times in the emergency department. Ultrasound (US) has potential to be a safer, faster diagnostic tool. This study aimed to determine how reliably US could detect distal radius fractures in children, to contrast the accuracy of 2DUS to 3DUS, and to assess the utility of artificial intelligence for image interpretation. 127 children were scanned with 2DUS and 3DUS on the affected wrist. US scans were then read by 7 blinded human readers and an AI model. With radiographs used as the gold standard, expert human readers obtained a mean sensitivity of 0.97 and 0.98 for 2DUS and 3DUS respectively. The AI model sensitivity was 0.91 and 1.00 for 2DUS and 3DUS respectively. Study data suggests that 2DUS is comparable to 3DUS and AI diagnosis is comparable to human experts.

Domain-aware contrastive learning for ultrasound hip image analysis.

Early diagnosis of Developmental Dysplasia of Hip (DDH) using ultrasound can result in simpler and more effective treatment options. Handheld ultrasound probes are ideally suited for such screening due to their low cost and portability. However, images from the pocket-sized probes are of lower quality than conventional probes. Image quality can be enhanced by image translation techniques that generate a pseudo-image mimicking the image quality of conventional probes. This can also help in generalizing the performance of AI-based automatic interpretation techniques to multiple probes. We develop a new domain-aware contrastive unpaired translation (D-CUT) technique for translating between images acquired from different ultrasound probes. Our approach embeds a Bone Probability Map (BPM) as part of the loss function which enforces higher structural similarity around bony regions in the image. Using the D-CUT model we translated 575 images acquired from a Philips Lumify handheld probe to generate pseudo-3D ultrasound (3DUS) images similar (Fréchet Inception Distance = 92) to those acquired from a conventional ultrasound probe (Philips iU22). The pseudo-3DUS images showed high structural similarity (SSIM = 0.68, Cosine Similarity = 0.65) with the original images and improved the contrast around the bony regions. This study establishes the feasibility of using D-CUT to improve the quality of data acquired from handheld ultrasound probes. Among other potential applications, clinical use of this tool could result in wider use of ultrasound for DDH screening programs.

Automatic Assessment Of Hip Effusion From MRI.

Joint effusion is a hallmark of osteoarthritis (OA) associated with stiffness, and may relate to pain, disability, and long-term outcomes. However, it is difficult to quantify accurately. We propose a new Deep Learning (DL) approach for automatic effusion assessment from Magnetic Resonance Imaging (MRI) using volumetric quantification measures (VQM). We developed a new multiplane ensemble convolutional neural network (CNN) approach for 1) localizing bony anatomy and 2) detecting effusion regions. CNNs were trained on femoral head and effusion regions manually segmented from 3856 images (63 patients). Upon validation on a non-overlapping set of 2040 images (34 patients) DL showed high agreement with ground-truth in terms of Dice score (0.85), sensitivity (0.86) and precision (0.83). Agreement of VQM per-patient was high for DL vs experts in term of Intraclass correlation coefficient (ICC)= 0.88[0.80,0.93]. We expect this technique to reduce inter-observer variability in effusion assessment, reducing expert time and potentially improving the quality of OA care.Clinical Relevance- Our technique for automatic assessment of hip MRI can be used for volumetric measurement of effusion. We expect this to reduce variability in OA biomarker assessment and provide more reliable indicators for disease progression.

Thyroid Nodule Segmentation and Classification Using Deep Convolutional Neural Network and Rule-based Classifiers.

Thyroid cancer has a high prevalence all over the world. Accurate thyroid nodule diagnosis can lead to effective treatment and decrease the mortality rate. Ultrasound imaging is a safe, portable, and inexpensive tool for thyroid nodule monitoring. However, the widespread use of ultrasound has also resulted in over-diagnosis and over-treatment of nodules. There is also large variability in the assessment and characterization of nodules. Thyroid nodule classification requires precise delineation of the nodule boundary which is tedious and time- consuming. Automatic segmentation of nodule boundaries is highly desirable, however, it is challenging due to the wide range of nodule appearances, shapes, and sizes. In this study, we propose an end-to-end pipeline for nodule segmentation and classification. A residual dilated UNet (resDUnet) model is proposed for nodule segmentation. The output of resDUnet is fed to two rule-based classifiers to categorize the composition and echogenicity of the segmented nodule. We evaluate our segmentation method on a large dataset of 352 ultrasound images reviewed by a certified radiologist. When compared with ground-truth, resDUnet gives a higher Dice score than the standard UNet (82% vs. 81%). Our method requires minimal user interaction and it is robust to reasonable variations in the user-specified region-of-interest. We expect the proposed method to reduce variability in thyroid nodule assessment which results in more efficient and cost-effective monitoring of thyroid cancer.

Narrative Review on the Role of Imaging in DDH.

BACKGROUND: Developmental dysplasia of hip (DDH) represents a spectrum from acetabular dysplasia to fixed dislocation, giving disability through premature osteoarthritis. Most DDH cases continue to present without any known risk factors such as breech presentation, female sex, and family history. Incidence and population-based outcomes of DDH are difficult to reliably establish due to many DDH definitions and classifications using different types of examinations. PURPOSE: This review takes a historical perspective on the role of imaging in DDH. METHODS: Pelvic radiographs (X-Ray) were amongst the first medical images identifying DDH, but these have a limited role in infancy due to absent ossification. In the 1980s, ultrasound led to a large expansion in infant DDH screening. Unfortunately, even for well-trained users, DDH indices on ultrasound generally lack reproducibility, and have led to overdiagnosis of mild DDH. CT and MRI more thoroughly evaluate the 3D hip deformity in DDH, but are costly, less available and involve radiation dose and/or anaesthesia. RESULTS: Recently 3D ultrasound has been used to characterize the 3D deformity of DDH more fully, with improved inter-observer reliability, particularly amongst novice users. 3D ultrasound is also well suited to automated image analysis, but high-resolution 3D probes are costly and not widely available. CONCLUSION: Combining the latest handheld portable ultrasound probes and artificial intelligence analysis could lead to an inexpensive tool permitting practical mass population screening for DDH. Overall, our understanding of DDH is heavily influenced by the imaging tools used to visualize it and changing quickly with modern technology.

A Novel 4D Semi-Automated Algorithm for Volumetric Segmentation in Echocardiography.

Segmentation of the left ventricle (LV) in temporal 3D echocardiography sequences poses a challenge. However, it is an essential component in generating quantitative clinical measurements for the diagnosis and treatment of various cardiac diseases. Identifying the endocardial borders of the left ventricle can be difficult due to the inherent properties of ultrasound. This study proposes a 4D segmentation algorithm that segments over temporal 3D volumes that has minimal user interaction and is based on a diffeomorphic registration approach. In contrast to several existing algorithms, the proposed method does not depend on training data or make any geometrical assumptions. The algorithm was evaluated on seven patients obtained from the Mazankowski Alberta Heart Institute, Edmonton, Canada in comparison to expert manual segmentation. The proposed approach yielded Dice scores of 0.94 (0.01), 0.91 (0.03) and 0.92 (0.02) at end diastole, at end systole and over the entire cardiac cycle, respectively. The corresponding Hausdorff distance values were 4.49 (1.01) mm, 4.94 (1.41) mm, and 5.05 (0.85) mm, respectively. These results demonstrate that the proposed 4D segmentation approach for the left ventricle is robust and can potentially be used in clinical practice.

Semiautomatic classification of acetabular shape from three-dimensional ultrasound for diagnosis of infant hip dysplasia using geometric features.

PURPOSE: Developmental dysplasia of the hip (DDH) is a congenital deformity which in severe cases leads to hip dislocation and in milder cases to premature osteoarthritis. Image-aided diagnosis of DDH is partly based on Graf classification which quantifies the acetabular shape seen at two-dimensional ultrasound (2DUS), which leads to high inter-scan variance. 3D ultrasound (3DUS) is a promising alternative for more reliable DDH diagnosis. However, manual quantification of acetabular shape from 3DUS is cumbersome. METHODS: Here, we (1) propose a semiautomated segmentation algorithm to delineate 3D acetabular surface models from 3DUS using graph search; (2) propose a fully automated method to classify acetabular shape based on a random forest (RF) classifier using features derived from 3D acetabular surface models; and (3) test diagnostic accuracy on a dataset of 79 3DUS infant hip recordings (36 normal, 16 borderline, 27 dysplastic based on orthopedic surgeon assessment) in 42 patients. For each 3DUS, we performed semiautomated segmentation to produce 3D acetabular surface models and then calculated geometric features including the automatic [Formula: see text]lpha (AA), acetabular contact angle (ACA), kurtosis (K), skewness (S) and convexity (C). Mean values of features obtained from surface models were used as inputs to train a RF classifier. RESULTS: Surface models were generated rapidly (user time 46.2 s) via semiautomated segmentation and visually closely correlated with actual acetabular contours (RMS error 1.39 ± 0.7 mm). A paired nonparametric u test on of feature values in each category showed statistically significant variation (p < 0.001) for AA, ACA and convexity. The RF classifier was 100 % specific and 97.2 % sensitive in classifying normal versus dysplastic hips and yielded true positive rates of 94.4, 62.5 and 89.9 % for normal, borderline and dysplastic hips. CONCLUSIONS: The proposed technique reduces the subjectivity of image-aided DDH diagnosis and could be useful in clinical practice.

Toward automated classification of acetabular shape in ultrasound for diagnosis of DDH: Contour alpha angle and the rounding index.

BACKGROUND AND OBJECTIVES: The diagnosis of Developmental Dysplasia of the Hip (DDH) in infants is currently made primarily by ultrasound. However, two-dimensional ultrasound (2DUS) images capture only an incomplete portion of the acetabular shape, and the alpha and beta angles measured on 2DUS for the Graf classification technique show high inter-scan and inter-observer variability. This variability relates partly to the manual determination of the apex point separating the acetabular roof from the ilium during index measurement. This study proposes a new 2DUS image processing technique for semi-automated tracing of the bony surface followed by automatic calculation of two indices: a contour-based alpha angle (αA), and a new modality-independent quantitative rounding index (M). The new index M is independent of the apex point, and can be directly extended to 3D surface models. METHODS: We tested the proposed indices on a dataset of 114 2DUS scans of infant hips aged between 4 and 183 days scanned using a 12MHz linear transducer. We calculated the manual alpha angle (αM), coverage, contour-based alpha angle and rounding index for each of the recordings and statistically evaluated these indices based on regression analysis, area under the receiver operating characteristic curve (AUC) and analysis of variance (ANOVA). RESULTS: Processing time for calculating αA and M was similar to manual alpha angle measurement, ∼30s per image. Reliability of the new indices was high, with inter-observer intraclass correlation coefficients (ICC) 0.90 for αA and 0.89 for M. For a diagnostic test classifying hips as normal or dysplastic, AUC was 93.0% for αA vs. 92.7% for αM, 91.6% for M alone, and up to 95.7% for combination of M with αM, αA or coverage. CONCLUSIONS: The rounding index provides complimentary information to conventional indices such as alpha angle and coverage. Calculation of the contour-based alpha angle and rounding index is rapid, shows potential to improve the reliability and accuracy of DDH diagnosis from 2DUS, and could be extended to 3D ultrasound in future.

An index for diagnosing infant hip dysplasia using 3-D ultrasound: the acetabular contact angle.

BACKGROUND: Developmental dysplasia of the hip (DDH) is a common condition that is highly treatable in infancy but can lead to the lifelong morbidity of premature osteoarthritis if left untreated. Current diagnostic methods lack reliability, which may be improved by using 3-D ultrasound. OBJECTIVE: Conventional 2-D US assessment of DDH has limitations, including high inter-scan variability. We quantified DDH on 3-D US using the acetabular contact angle (ACA), a property of the 3-D acetabular shape. We assessed ACA reliability and diagnostic utility. MATERIALS AND METHODS: We prospectively collected data from January 2013 to December 2014, including 114 hips in 85 children divided into three clinical diagnostic groups: (1) normal, (2) initially borderline but ultimately normal without treatment and (3) dysplastic requiring treatment. Using custom software, two observers each traced acetabula twice on two 3-D US scans of each hip, enabling automated generation of 3-D surface models and ACA calculation. We computed inter-observer and inter-scan variability of repeatability coefficients and generated receiver operating characteristic (ROC) curves. RESULTS: The 3-D US acetabular contact angle was reproduced 95% of the time within 6° in the same scan and within 9° in different scans of the same hip, vs. 9° and 14° for the 2-D US alpha angle (P < 0.001). Areas under ROC curves for diagnosis of developmental dysplasia of the hip were 0.954 for ACA and 0.927 for alpha angle. CONCLUSION: The 3-D US ACA was significantly more reliable than 2-D US alpha angle, and the 3-D US measurement predicted the presence of DDH with slightly higher accuracy. The ACA therefore shows promising initial diagnostic utility. Our findings call for further study of 3-D US in the diagnosis and longer-term follow-up of infant hip dysplasia.

Hip segmentation from MRI volumes in infants for DDH diagnosis and treatment planning.

Diagnosis and surgical management of Developmental Dysplasia of the Hip (DDH) relies on physical examination and 2D ultrasound scanning. Magnetic Resonance Imaging (MRI) can be used to complement existing techniques and could be advantageous in treatment planning due to its larger field of view. In this paper we propose a semi-automatic method to segment surface models of the acetabulum from MRI images. The method incorporates clinical knowledge in the form of intensity priors which are integrated into a Random Walker (RW) formulation. We use a modified RW framework which compensates for incomplete or blurred boundaries in the image by using information from neighboring slices in the sequence incorporated as node weights. We conducted a pilot study to evaluate the segmentation on a set of 10 infant hip MRI sequences using a 1.5 Tesla MR scanner. Contours obtained from the semi-automated segmentation were compared against manually segmented hip contours using Dice Ratio (DR), Hausdorff Distance (HD) and Root Mean Square (RMS) distance. The proposed method gave values of (DR = 0.84 ± 0.5, HD =3.0 ± 0.7, RMS =1.9 ± 0.3) and (DR=0.86 ± 0.2, HD=3.0 ± 0.1, RMS= 2.0 ± 0.6) for right and left acetabular contours respectively which was higher than the corresponding values obtained from conventional RW segmentation. The execution time of the segmentation algorithm was less than ~4 seconds on a 3.5 GHz CPU.

A technique for semiautomatic segmentation of echogenic structures in 3D ultrasound, applied to infant hip dysplasia.

PURPOSE: Automatic segmentation of anatomical structures and lesions from medical ultrasound images is a formidable challenge in medical imaging due to image noise, blur and artifacts. In this paper we present a segmentation technique with features highly suited to use in noisy 3D ultrasound volumes and demonstrate its use in modeling bone, specifically the acetabulum in infant hips. Quantification of the acetabular shape is crucial in diagnosing developmental dysplasia of the hip (DDH), a common condition associated with hip dislocation and premature osteoarthritis if not treated. The well-established Graf technique for DDH diagnosis has been criticized for high inter-observer and inter-scan variability. In our earlier work we have introduced a more reliable instability metric based on 3D ultrasound data. Visualizing and interpreting the acetabular shape from noisy 3D ultrasound volumes has been one of the major roadblocks in using 3D ultrasound as diagnostic tool for DDH. For this study we developed a semiautomated segmentation technique to rapidly generate 3D acetabular surface models and classified the acetabulum based on acetabular contact angle (ACA) derived from the models. We tested the feasibility and reliability of the technique compared with manual segmentation. METHODS: The proposed segmentation algorithm is based on graph search. We formulate segmentation of the acetabulum as an optimal path finding problem on an undirected weighted graph. Slice contours are defined as the optimal path passing through a set of user-defined seed points in the graph, and it can be found using dynamic programming techniques (in this case Dijkstra's algorithm). Slice contours are then interpolated over the 3D volume to generate the surface model. A three-dimensional ACA was calculated using normal vectors of the surface model. RESULTS: The algorithm was tested over an extensive dataset of 51 infant ultrasound hip volumes obtained from 42 subjects with normal to dysplastic hips. The contours generated by the segmentation algorithm closely matched with those obtained from manual segmentation. The average RMS errors between the semiautomated and manual segmentation for the 51 volumes were 0.28 mm/1.1 voxel (with 2 node points) and 0.24 mm/0.9 voxel (with 3 node points). The semiautomatic algorithm gave visually acceptable results on images with moderate levels of noise and was able to trace the boundary of the acetabulum even in the presence of significant shadowing. Semiautomatic contouring was also faster than manual segmentation at 37 versus 56 s per scan. It also improved the repeatability of the ACA calculation with inter-observer and intra-observer variability of 1.4 ± 0.9 degree and 1.4 ± 1.0 degree. CONCLUSION: The semiautomatic segmentation technique proposed in this work offers a fast and reliable method to delineate the contours of the acetabulum from 3D ultrasound volumes of the hip. Since the technique does not rely upon contour evolution, it is less susceptible than other methods to the frequent missing or incomplete boundaries and noise artifacts common in ultrasound images. ACA derived from the segmented 3D surface was able to accurately classify the acetabulum under the categories normal, borderline and dysplastic. The semiautomatic technique makes it easier to segment the volume and reduces the inter-observer and intra-observer variation in ACA calculation compared with manual segmentation. The method can be applied to any structure with an echogenic boundary on ultrasound (such as a ventricle, blood vessel, organ or tumor), or even to structures with a bright border on computed tomography or magnetic resonance imaging.