Deep learning predicts hip fracture using confounding patient and healthcare variables

Prospects and Pitfalls of Machine Learning in Nutritional Epidemiology

Nutritional epidemiology employs observational data to discover associations between diet and disease risk. However, existing analytic methods of dietary data are often sub-optimal, with limited incorporation and analysis of the correlations between the studied variables and nonlinear behaviours in the data. Machine learning (ML) is an area of artificial intelligence that has the potential to improve modelling of nonlinear associations and confounding which are found in nutritional data. These opportunities notwithstanding, the applications of ML in nutritional epidemiology must be approached cautiously to safeguard the scientific quality of the results and provide accurate interpretations. Given the complex scenario around ML, judicious application of such tools is necessary to offer nutritional epidemiology a novel analytical resource for dietary measurement and assessment and a tool to model the complexity of dietary intake and its relation to health. This work describes the applications of ML in nutritional epidemiology and provides guidelines to avoid common pitfalls encountered in applying predictive statistical models to nutritional data. Furthermore, it helps unfamiliar readers better assess the significance of their results and provides new possible future directions in the field of ML in nutritional epidemiology.

Identification of osteoporosis using ensemble deep learning model with panoramic radiographs and clinical covariates

Osteoporosis is becoming a global health issue due to increased life expectancy. However, it is difficult to detect in its early stages owing to a lack of discernible symptoms. Hence, screening for osteoporosis with widely used dental panoramic radiographs would be very cost-effective and useful. In this study, we investigate the use of deep learning to classify osteoporosis from dental panoramic radiographs. In addition, the effect of adding clinical covariate data to the radiographic images on the identification performance was assessed. For objective labeling, a dataset containing 778 images was collected from patients who underwent both skeletal-bone-mineral density measurement and dental panoramic radiography at a single general hospital between 2014 and 2020. Osteoporosis was assessed from the dental panoramic radiographs using convolutional neural network (CNN) models, including EfficientNet-b0, -b3, and -b7 and ResNet-18, -50, and -152. An ensemble model was also constructed with clinical covariates added to each CNN. The ensemble model exhibited improved performance on all metrics for all CNNs, especially accuracy and AUC. The results show that deep learning using CNN can accurately classify osteoporosis from dental panoramic radiographs. Furthermore, it was shown that the accuracy can be improved using an ensemble model with patient covariates.

Handling Imbalanced Data: Uncertainty-guided Virtual Adversarial Training with Batch Nuclear-norm Optimization for Semi-supervised Medical Image Classification

In many clinical settings, a lot of medical image datasets suffer from imbalance problems, which makes predictions of trained models to be biased toward majority classes. Semi-supervised Learning (SSL) algorithms trained with such imbalanced datasets become more problematic since pseudo-supervision of unlabeled data are generated from the model's biased predictions. To address these issues, in this work, we propose a novel semi-supervised deep learning method, i.e., uncertainty-guided virtual adversarial training (VAT) with batch nuclear-norm (BNN) optimization, for large-scale medical image classification. To effectively exploit useful information from both labeled and unlabeled data, we leverage VAT and BNN optimization to harness the underlying knowledge, which helps to improve discriminability, diversity and generalization of the trained models. More concretely, our network is trained by minimizing a combination of four types of losses, including a supervised cross-entropy loss, a BNN loss defined on the output matrix of labeled data batch (lBNN loss), a negative BNN loss defined on the output matrix of unlabeled data batch (uBNN loss), and a VAT loss on both labeled and unlabeled data. We additionally propose to use uncertainty estimation to filter out unlabeled samples near the decision boundary when computing the VAT loss. We conduct comprehensive experiments to evaluate the performance of our method on two publicly available datasets and one in-house collected dataset. The experimental results demonstrated that our method achieved better results than state-of-the-art SSL methods.

Navigating the pitfalls of applying machine learning in genomics

The scale of genetic, epigenomic, transcriptomic, cheminformatic and proteomic data available today, coupled with easy-to-use machine learning (ML) toolkits, has propelled the application of supervised learning in genomics research. However, the assumptions behind the statistical models and performance evaluations in ML software frequently are not met in biological systems. In this Review, we illustrate the impact of several common pitfalls encountered when applying supervised ML in genomics. We explore how the structure of genomics data can bias performance evaluations and predictions. To address the challenges associated with applying cutting-edge ML methods to genomics, we describe solutions and appropriate use cases where ML modelling shows great potential.

Automatic detection and classification of peri-prosthetic femur fracture

Purpose

Object classification and localization is a key task of computer-aided diagnosis (CAD) tool. Although there have been numerous generic deep learning (DL) models developed for CAD, there is no work in the literature to evaluate their effectiveness when utilized in diagnosing fractures in proximity of joint implants. In this work, we aim to assess the performance of existing classification systems on binary and multi-class problems (fracture types) using plain radiographs. In addition, we evaluated the performance of object detection systems using the one- and two-stage DL architectures.

Methods

A data set of 1272 X-ray images of Peri-prosthetic Femur Fracture PFF was collected. The fractures were annotated with bounding boxes and classified according to the Vancouver Classification System (type A, B, C) by two clinical specialists. Four classification models such as Densenet161, Resnet50, Inception, VGG and two object detection models such as Faster RCNN and RetinaNet were evaluated, and their performance compared. Six confusion matrix-based measures were reported to evaluate fracture classification. For localization of the fracture, Average Precision and localization accuracy were reported.

Results

The Resnet50 showed the best performance with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$95\%$$\end{document}95% accuracy and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$94\%$$\end{document}94% F1-score in the binary classification: fracture/normal. In addition, the Resnet50 showed \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$90\%$$\end{document}90% accuracy in multi-classification (normal, Vancouver type A, B and C).

Conclusions

A large data set of PFF images and the annotations of fracture features by two independent assessments were created to implement a DL-based approach for detecting, classifying and localizing PFFs. It was shown that this approach could be a promising diagnostic tool of fractures in proximity of joint implants.

Comparison of error rates between four pretrained DenseNet convolutional neural network models and 13 board-certified veterinary radiologists when evaluating 15 labels of canine thoracic radiographs

Convolutional neural networks (CNNs) are commonly used as artificial intelligence (AI) tools for evaluating radiographs, but published studies testing their performance in veterinary patients are currently lacking. The purpose of this retrospective, secondary analysis, diagnostic accuracy study was to compare the error rates of four CNNs to the error rates of 13 veterinary radiologists for evaluating canine thoracic radiographs using an independent gold standard. Radiographs acquired at a referral institution were used to evaluate the four CNNs sharing a common architecture. Fifty radiographic studies were selected at random. The studies were evaluated independently by three board-certified veterinary radiologists for the presence or absence of 15 thoracic labels, thus creating the gold standard through the majority rule. The labels included "cardiovascular," "pulmonary," "pleural," "airway," and "other categories." The error rates for each of the CNNs and for 13 additional board-certified veterinary radiologists were calculated on those same studies. There was no statistical difference in the error rates among the four CNNs for the majority of the labels. However, the CNN's training method impacted the overall error rate for three of 15 labels. The veterinary radiologists had a statistically lower error rate than all four CNNs overall and for five labels (33%). There was only one label ("esophageal dilation") for which two CNNs were superior to the veterinary radiologists. Findings from the current study raise numerous questions that need to be addressed to further develop and standardize AI in the veterinary radiology environment and to optimize patient care.

Machine learning outperforms clinical experts in classification of hip fractures

Hip fractures are a major cause of morbidity and mortality in the elderly, and incur high health and social care costs. Given projected population ageing, the number of incident hip fractures is predicted to increase globally. As fracture classification strongly determines the chosen surgical treatment, differences in fracture classification influence patient outcomes and treatment costs. We aimed to create a machine learning method for identifying and classifying hip fractures, and to compare its performance to experienced human observers. We used 3659 hip radiographs, classified by at least two expert clinicians. The machine learning method was able to classify hip fractures with 19% greater accuracy than humans, achieving overall accuracy of 92%.

Surgical data science - from concepts toward clinical translation

Recent developments in data science in general and machine learning in particular have transformed the way experts envision the future of surgery. Surgical Data Science (SDS) is a new research field that aims to improve the quality of interventional healthcare through the capture, organization, analysis and modeling of data. While an increasing number of data-driven approaches and clinical applications have been studied in the fields of radiological and clinical data science, translational success stories are still lacking in surgery. In this publication, we shed light on the underlying reasons and provide a roadmap for future advances in the field. Based on an international workshop involving leading researchers in the field of SDS, we review current practice, key achievements and initiatives as well as available standards and tools for a number of topics relevant to the field, namely (1) infrastructure for data acquisition, storage and access in the presence of regulatory constraints, (2) data annotation and sharing and (3) data analytics. We further complement this technical perspective with (4) a review of currently available SDS products and the translational progress from academia and (5) a roadmap for faster clinical translation and exploitation of the full potential of SDS, based on an international multi-round Delphi process.

Detecting Distal Radial Fractures from Wrist Radiographs Using a Deep Convolutional Neural Network with an Accuracy Comparable to Hand Orthopedic Surgeons

In recent years, fracture image diagnosis using a convolutional neural network (CNN) has been reported. The purpose of the present study was to evaluate the ability of CNN to diagnose distal radius fractures (DRFs) using frontal and lateral wrist radiographs. We included 503 cases of DRF diagnosed by plain radiographs and 289 cases without fracture. We implemented the CNN model using Keras and Tensorflow. Frontal and lateral views of wrist radiographs were manually cropped and trained separately. Fine-tuning was performed using EfficientNets. The diagnostic ability of CNN was evaluated using 150 images with and without fractures from anteroposterior and lateral radiographs. The CNN model diagnosed DRF based on three views: frontal view, lateral view, and both frontal and lateral view. We determined the sensitivity, specificity, and accuracy of the CNN model, plotted a receiver operating characteristic (ROC) curve, and calculated the area under the ROC curve (AUC). We further compared performances between the CNN and three hand orthopedic surgeons. EfficientNet-B2 in the frontal view and EfficientNet-B4 in the lateral view showed highest accuracy on the validation dataset, and these models were used for combined views. The accuracy, sensitivity, and specificity of the CNN based on both anteroposterior and lateral radiographs were 99.3, 98.7, and 100, respectively. The accuracy of the CNN was equal to or better than that of three orthopedic surgeons. The AUC of the CNN on the combined views was 0.993. The CNN model exhibited high accuracy in the diagnosis of distal radius fracture with a plain radiograph.

Musculoskeletal trauma and artificial intelligence: current trends and projections

Musculoskeletal trauma accounts for a significant fraction of emergency department visits and patients seeking urgent care, with a high financial cost to society. Diagnostic imaging is indispensable in the workup and management of trauma patients. However, diagnostic imaging represents a complex multifaceted system, with many aspects of its workflow prone to inefficiencies or human error. Recent technological innovations in artificial intelligence and machine learning have shown promise to revolutionize our systems for providing medical care to patients. This review will provide a general overview of the current state of artificial intelligence and machine learning applications in different aspects of trauma imaging and provide a vision for how such applications could be leveraged to enhance our diagnostic imaging systems and optimize patient outcomes.

Generalization Challenges in Drug-Resistant Tuberculosis Detection from Chest X-rays

Classification of drug-resistant tuberculosis (DR-TB) and drug-sensitive tuberculosis (DS-TB) from chest radiographs remains an open problem. Our previous cross validation performance on publicly available chest X-ray (CXR) data combined with image augmentation, the addition of synthetically generated and publicly available images achieved a performance of 85% AUC with a deep convolutional neural network (CNN). However, when we evaluated the CNN model trained to classify DR-TB and DS-TB on unseen data, significant performance degradation was observed (65% AUC). Hence, in this paper, we investigate the generalizability of our models on images from a held out country's dataset. We explore the extent of the problem and the possible reasons behind the lack of good generalization. A comparison of radiologist-annotated lesion locations in the lung and the trained model's localization of areas of interest, using GradCAM, did not show much overlap. Using the same network architecture, a multi-country classifier was able to identify the country of origin of the X-ray with high accuracy (86%), suggesting that image acquisition differences and the distribution of non-pathological and non-anatomical aspects of the images are affecting the generalization and localization of the drug resistance classification model as well. When CXR images were severely corrupted, the performance on the validation set was still better than 60% AUC. The model overfitted to the data from countries in the cross validation set but did not generalize to the held out country. Finally, we applied a multi-task based approach that uses prior TB lesions location information to guide the classifier network to focus its attention on improving the generalization performance on the held out set from another country to 68% AUC.

A Study on 3D Deep Learning-Based Automatic Diagnosis of Nasal Fractures

This paper reported a study on the 3-dimensional deep-learning-based automatic diagnosis of nasal fractures. (1) Background: The nasal bone is the most protuberant feature of the face; therefore, it is highly vulnerable to facial trauma and its fractures are known as the most common facial fractures worldwide. In addition, its adhesion causes rapid deformation, so a clear diagnosis is needed early after fracture onset. (2) Methods: The collected computed tomography images were reconstructed to isotropic voxel data including the whole region of the nasal bone, which are represented in a fixed cubic volume. The configured 3-dimensional input data were then automatically classified by the deep learning of residual neural networks (3D-ResNet34 and ResNet50) with the spatial context information using a single network, whose performance was evaluated by 5-fold cross-validation. (3) Results: The classification of nasal fractures with simple 3D-ResNet34 and ResNet50 networks achieved areas under the receiver operating characteristic curve of 94.5% and 93.4% for binary classification, respectively, both indicating unprecedented high performance in the task. (4) Conclusions: In this paper, it is presented the possibility of automatic nasal bone fracture diagnosis using a 3-dimensional Resnet-based single classification network and it will improve the diagnostic environment with future research.

Evaluation System of Smart Logistics Comprehensive Management Based on Hospital Data Fusion Technology

With the acceleration of the informatization process, but because of the late start of the informatization construction of logistics management, the current digital system construction of logistics management has not been popularized, and the intelligent logistics integrated management evaluation system is also extremely lacking. In order to solve the lack of existing intelligent logistics comprehensive management evaluation system, this paper introduces the research of intelligent logistics comprehensive management evaluation system based on hospital data fusion technology. This paper analyzes and utilizes the Kalman filter and adaptive weighted data fusion technology in data fusion technology and then analyzes the evaluation index and system design principles of the intelligent logistics comprehensive management evaluation system and then designs the application layer from the application layer. Design the application layer from the application layer. Then design the framework of the intelligent logistics comprehensive management evaluation system at the network layer and the data layer. The system is finally tested, and the test results show that the evaluation accuracy of the system reaches 80%.

Machine Learning Approach to Classify Cardiovascular Disease in Patients With Nonalcoholic Fatty Liver Disease in the UK Biobank Cohort

Background Nonalcoholic fatty liver disease (NAFLD) is the most prevalent liver disease worldwide. Cardiovascular disease (CVD) is the leading cause of mortality among patients with NAFLD. The aim of our study was to develop a machine learning algorithm integrating clinical, lifestyle, and genetic risk factors to identify CVD in patients with NAFLD. Methods and Results We created a cohort of patients with NAFLD from the UK Biobank, diagnosed according to proton density fat fraction from magnetic resonance imaging data sets. A total of 400 patients with NAFLD with subclinical atherosclerosis or clinical CVD, defined by disease codes, constituted cases and 446 NAFLD cases with no CVD constituted controls. We evaluated 7 different supervised machine learning approaches on clinical, lifestyle, and genetic variables for identifying CVD in patients with NAFLD. The most significant clinical and lifestyle variables observed by the predictive modeling were age (59 years [54.00-63.00 years]), hypertension (145 mm Hg [134.0-156.0 mm Hg] and 85 mm Hg [79.00-93.00 mm Hg]), waist circumference (98 cm [95.00-105.00 cm]), and sedentary lifestyle, defined as time spent watching TV >4 h/d. In the genetic data, single-nucleotide polymorphisms in IL16 and ANKLE1 gene were most significant. Our proposed ensemble-based integrative machine learning model achieved an area under the curve of 0.849 using the random forest modeling for CVD prediction. Conclusions We propose a machine learning algorithm that identifies CVD in patients with NAFLD through integration of significant clinical, lifestyle, and genetic risk factors. These patients with NAFLD at higher risk of CVD should be flagged for screening and aggressive treatment of their cardiometabolic risk factors to prevent cardiovascular morbidity and mortality.

Expectations for Artificial Intelligence (AI) in Psychiatry

PURPOSE OF REVIEW

Artificial intelligence (AI) is often presented as a transformative technology for clinical medicine even though the current technology maturity of AI is low. The purpose of this narrative review is to describe the complex reasons for the low technology maturity and set realistic expectations for the safe, routine use of AI in clinical medicine.

RECENT FINDINGS

For AI to be productive in clinical medicine, many diverse factors that contribute to the low maturity level need to be addressed. These include technical problems such as data quality, dataset shift, black-box opacity, validation and regulatory challenges, and human factors such as a lack of education in AI, workflow changes, automation bias, and deskilling. There will also be new and unanticipated safety risks with the introduction of AI. The solutions to these issues are complex and will take time to discover, develop, validate, and implement. However, addressing the many problems in a methodical manner will expedite the safe and beneficial use of AI to augment medical decision making in psychiatry.

Imaging Manifestations and Evaluation of Postoperative Complications of Bone and Joint Infections under Deep Learning

To explore and evaluate the imaging manifestations of postoperative complications of bone and joint infections based on deep learning, a retrospective study was performed on 40 patients with bone and joint infections in the Department of Orthopedics of Orthopedics Hospital of Henan Province of Luoyang City. Sensitivity and Dice similarity coefficient (DSC) were used to evaluate the image results by convolutional neural network (CNN) algorithm. Imaging features of postoperative complications in 40 patients were analyzed. Then, three imaging methods were used to diagnose the features. Sensitivity and specificity were used to evaluate the diagnostic performance of three imaging methods for imaging features. Compared with professional doctors and biomarker algorithms, the sensitivity of CNN algorithm proposed was 90.6%, and DSC was 84.1%. Compared with traditional methods, the CNN algorithm has higher image resolution and wider and more accurate lesion area recognition and division. The three manifestations of soft tissue abscess, periosteum swelling, and bone damage were postoperative imaging features of bone and joint infections. In addition, compared with X-ray, CT examination and MRI examination were better for the examination of imaging characteristics. CT and MRI had higher sensitivity and specificity than X-ray. The experimental results show that CNN algorithm can effectively identify and divide pathological images and assist doctors to diagnose the images more efficiently in clinic.

Augmenting Osteoporosis Imaging with Machine Learning

PURPOSE OF REVIEW

In this paper, we discuss how recent advancements in image processing and machine learning (ML) are shaping a new and exciting era for the osteoporosis imaging field. With this paper, we want to give the reader a basic exposure to the ML concepts that are necessary to build effective solutions for image processing and interpretation, while presenting an overview of the state of the art in the application of machine learning techniques for the assessment of bone structure, osteoporosis diagnosis, fracture detection, and risk prediction.

RECENT FINDINGS

ML effort in the osteoporosis imaging field is largely characterized by "low-cost" bone quality estimation and osteoporosis diagnosis, fracture detection, and risk prediction, but also automatized and standardized large-scale data analysis and data-driven imaging biomarker discovery. Our effort is not intended to be a systematic review, but an opportunity to review key studies in the recent osteoporosis imaging research landscape with the ultimate goal of discussing specific design choices, giving the reader pointers to possible solutions of regression, segmentation, and classification tasks as well as discussing common mistakes.

Artificial Neural Networks Predict 30-Day Mortality After Hip Fracture: Insights From Machine Learning

OBJECTIVES

Accurately stratifying patients in the preoperative period according to mortality risk informs treatment considerations and guides adjustments to bundled reimbursements. We developed and compared three machine learning models to determine which best predicts 30-day mortality after hip fracture.

METHODS

The 2016 to 2017 National Surgical Quality Improvement Program for hip fracture (AO/OTA 31-A-B-C) procedure-targeted data were analyzed. Three models-artificial neural network, naive Bayes, and logistic regression-were trained and tested using independent variables selected via backward variable selection. The data were split into 80% training and 20% test sets. Predictive accuracy between models was evaluated using area under the curve receiver operating characteristics. Odds ratios were determined using multivariate logistic regression with P < 0.05 for significance.

RESULTS

The study cohort included 19,835 patients (69.3% women). The 30-day mortality rate was 5.3%. In total, 47 independent patient variables were identified to train the testing models. Area under the curve receiver operating characteristics for 30-day mortality was highest for artificial neural network (0.92), followed by the logistic regression (0.87) and naive Bayes models (0.83).

DISCUSSION

Machine learning is an emerging approach to develop accurate risk calculators that account for the weighted interactions between variables. In this study, we developed and tested a neural network model that was highly accurate for predicting 30-day mortality after hip fracture. This was superior to the naive Bayes and logistic regression models. The role of machine learning models to predict orthopaedic outcomes merits further development and prospective validation but shows strong promise for positively impacting patient care.

Real-world analysis of artificial intelligence in musculoskeletal trauma

Musculoskeletal trauma accounts for a large percentage of emergency room visits and is amongst the top causes of unscheduled patient visits to the emergency room. Musculoskeletal trauma results in expenditure of billions of dollars and protracted losses of quality-adjusted life years. New and innovative methods are needed to minimise the impact by ensuring quick and accurate assessment. However, each of the currently utilised radiological procedures, such as radiography, ultrasonography, computed tomography, and magnetic resonance imaging, has resulted in implosion of medical imaging data. Deep learning, a recent advancement in artificial intelligence, has demonstrated the potential to analyse medical images with sensitivity and specificity at par with experts. In this review article, we intend to summarise and showcase the various developments which have occurred in the dynamic field of artificial intelligence and machine learning and how their applicability to different aspects of imaging in trauma can be explored to improvise our existing reporting systems and improvise on patient outcomes.

Automated Detection of Glaucoma With Interpretable Machine Learning Using Clinical Data and Multimodal Retinal Images

PURPOSE

To develop a multimodal model to automate glaucoma detection DESIGN: Development of a machine-learning glaucoma detection model METHODS: We selected a study cohort from the UK Biobank data set with 1193 eyes of 863 healthy subjects and 1283 eyes of 771 subjects with glaucoma. We trained a multimodal model that combines multiple deep neural nets, trained on macular optical coherence tomography volumes and color fundus photographs, with demographic and clinical data. We performed an interpretability analysis to identify features the model relied on to detect glaucoma. We determined the importance of different features in detecting glaucoma using interpretable machine learning methods. We also evaluated the model on subjects who did not have a diagnosis of glaucoma on the day of imaging but were later diagnosed (progress-to-glaucoma [PTG]).

RESULTS

Results show that a multimodal model that combines imaging with demographic and clinical features is highly accurate (area under the curve 0.97). Interpretation of this model highlights biological features known to be related to the disease, such as age, intraocular pressure, and optic disc morphology. Our model also points to previously unknown or disputed features, such as pulmonary function and retinal outer layers. Accurate prediction in PTG highlights variables that change with progression to glaucoma-age and pulmonary function.

CONCLUSIONS

The accuracy of our model suggests distinct sources of information in each imaging modality and in the different clinical and demographic variables. Interpretable machine learning methods elucidate subject-level prediction and help uncover the factors that lead to accurate predictions, pointing to potential disease mechanisms or variables related to the disease.

An increasing number of convolutional neural networks for fracture recognition and classification in orthopaedics : are these externally validated and ready for clinical application?

Aims

The number of convolutional neural networks (CNN) available for fracture detection and classification is rapidly increasing. External validation of a CNN on a temporally separate (separated by time) or geographically separate (separated by location) dataset is crucial to assess generalizability of the CNN before application to clinical practice in other institutions. We aimed to answer the following questions: are current CNNs for fracture recognition externally valid?; which methods are applied for external validation (EV)?; and, what are reported performances of the EV sets compared to the internal validation (IV) sets of these CNNs?

Methods

The PubMed and Embase databases were systematically searched from January 2010 to October 2020 according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. The type of EV, characteristics of the external dataset, and diagnostic performance characteristics on the IV and EV datasets were collected and compared. Quality assessment was conducted using a seven-item checklist based on a modified Methodologic Index for NOn-Randomized Studies instrument (MINORS).

Results

Out of 1,349 studies, 36 reported development of a CNN for fracture detection and/or classification. Of these, only four (11%) reported a form of EV. One study used temporal EV, one conducted both temporal and geographical EV, and two used geographical EV. When comparing the CNN’s performance on the IV set versus the EV set, the following were found: AUCs of 0.967 (IV) versus 0.975 (EV), 0.976 (IV) versus 0.985 to 0.992 (EV), 0.93 to 0.96 (IV) versus 0.80 to 0.89 (EV), and F1-scores of 0.856 to 0.863 (IV) versus 0.757 to 0.840 (EV).

Conclusion

The number of externally validated CNNs in orthopaedic trauma for fracture recognition is still scarce. This greatly limits the potential for transfer of these CNNs from the developing institute to another hospital to achieve similar diagnostic performance. We recommend the use of geographical EV and statements such as the Consolidated Standards of Reporting Trials–Artificial Intelligence (CONSORT-AI), the Standard Protocol Items: Recommendations for Interventional Trials–Artificial Intelligence (SPIRIT-AI) and the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis–Machine Learning (TRIPOD-ML) to critically appraise performance of CNNs and improve methodological rigor, quality of future models, and facilitate eventual implementation in clinical practice.

Cite this article: Bone Jt Open 2021;2(10):879–885.

Curriculum learning for improved femur fracture classification: Scheduling data with prior knowledge and uncertainty

An adequate classification of proximal femur fractures from X-ray images is crucial for the treatment choice and the patients' clinical outcome. We rely on the commonly used AO system, which describes a hierarchical knowledge tree classifying the images into types and subtypes according to the fracture's location and complexity. In this paper, we propose a method for the automatic classification of proximal femur fractures into 3 and 7 AO classes based on a Convolutional Neural Network (CNN). As it is known, CNNs need large and representative datasets with reliable labels, which are hard to collect for the application at hand. In this paper, we design a curriculum learning (CL) approach that improves over the basic CNNs performance under such conditions. Our novel formulation reunites three curriculum strategies: individually weighting training samples, reordering the training set, and sampling subsets of data. The core of these strategies is a scoring function ranking the training samples. We define two novel scoring functions: one from domain-specific prior knowledge and an original self-paced uncertainty score. We perform experiments on a clinical dataset of proximal femur radiographs. The curriculum improves proximal femur fracture classification up to the performance of experienced trauma surgeons. The best curriculum method reorders the training set based on prior knowledge resulting into a classification improvement of 15%. Using the publicly available MNIST dataset, we further discuss and demonstrate the benefits of our unified CL formulation for three controlled and challenging digit recognition scenarios: with limited amounts of data, under class-imbalance, and in the presence of label noise. The code of our work is available at: https://github.com/ameliajimenez/curriculum-learning-prior-uncertainty.

Characterization of tissue types in basal cell carcinoma images via generative modeling and concept vectors

The promise of machine learning methods to act as decision support systems for pathologists continues to grow. However, central to their successful adoption must be interpretable implementations so that people can trust and learn from them effectively. Generative modeling, most notable in the form of adversarial generative models, is a naturally interpretable technique because the quality of the model is explicit from the quality of images it generates. Such a model can be further assessed by exploring its latent space, using human-meaningful concepts by defining concept vectors. Motivated by these ideas, we apply for the first time generative methods to histological images of basal cell carcinoma (BCC). By simultaneously learning to generate and encode realistic image patches, we extract feature rich latent vectors that correspond to various tissue morphologies, namely BCC, epidermis, keratin, papillary dermis and inflammation. We show that a logistic regression model trained on these latent vectors can achieve high classification accuracies across 6 binary tasks (86-98%). Further, by projecting the latent vectors onto learned concept vectors we can generate a score for the absence or degree of presence for a given concept, providing semantically accurate "conceptual summaries" of the various tissues types within a patch. This can be extended to generate multi-dimensional heat maps for whole-image specimens, which characterizes the tissue in a similar way to a pathologist. We additionally find that accurate concept vectors can be defined using a small labeled dataset.

Presenting artificial intelligence, deep learning, and machine learning studies to clinicians and healthcare stakeholders: an introductory reference with a guideline and a Clinical AI Research (CAIR) checklist proposal

Background and purpose - Artificial intelligence (AI), deep learning (DL), and machine learning (ML) have become common research fields in orthopedics and medicine in general. Engineers perform much of the work. While they gear the results towards healthcare professionals, the difference in competencies and goals creates challenges for collaboration and knowledge exchange. We aim to provide clinicians with a context and understanding of AI research by facilitating communication between creators, researchers, clinicians, and readers of medical AI and ML research.Methods and results - We present the common tasks, considerations, and pitfalls (both methodological and ethical) that clinicians will encounter in AI research. We discuss the following topics: labeling, missing data, training, testing, and overfitting. Common performance and outcome measures for various AI and ML tasks are presented, including accuracy, precision, recall, F1 score, Dice score, the area under the curve, and ROC curves. We also discuss ethical considerations in terms of privacy, fairness, autonomy, safety, responsibility, and liability regarding data collecting or sharing.Interpretation - We have developed guidelines for reporting medical AI research to clinicians in the run-up to a broader consensus process. The proposed guidelines consist of a Clinical Artificial Intelligence Research (CAIR) checklist and specific performance metrics guidelines to present and evaluate research using AI components. Researchers, engineers, clinicians, and other stakeholders can use these proposal guidelines and the CAIR checklist to read, present, and evaluate AI research geared towards a healthcare setting.

Deep Learning for the Diagnosis of Stage in Retinopathy of Prematurity: Accuracy and Generalizability across Populations and Cameras

PURPOSE

Stage is an important feature to identify in retinal images of infants at risk of retinopathy of prematurity (ROP). The purpose of this study was to implement a convolutional neural network (CNN) for binary detection of stages 1, 2, and 3 in ROP and to evaluate its generalizability across different populations and camera systems.

DESIGN

Diagnostic validation study of CNN for stage detection.

PARTICIPANTS

Retinal fundus images obtained from preterm infants during routine ROP screenings.

METHODS

Two datasets were used: 5943 fundus images obtained by RetCam camera (Natus Medical, Pleasanton, CA) from 9 North American institutions and 5049 images obtained by 3nethra camera (Forus Health Incorporated, Bengaluru, India) from 4 hospitals in Nepal. Images were labeled based on the presence of stage by 1 to 3 expert graders. Three CNN models were trained using 5-fold cross-validation on datasets from North America alone, Nepal alone, and a combined dataset and were evaluated on 2 held-out test sets consisting of 708 and 247 images from the Nepali and North American datasets, respectively.

MAIN OUTCOME MEASURES

Convolutional neural network performance was evaluated using area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), sensitivity, and specificity.

RESULTS

Both the North American- and Nepali-trained models demonstrated high performance on a test set from the same population: AUROC, 0.99; AUPRC, 0.98; sensitivity, 94%; and AUROC, 0.97; AUPRC, 0.91; and sensitivity, 73%; respectively. However, the performance of each model decreased to AUROC of 0.96 and AUPRC of 0.88 (sensitivity, 52%) and AUROC of 0.62 and AUPRC of 0.36 (sensitivity, 44%) when evaluated on a test set from the other population. Compared with the models trained on individual datasets, the model trained on a combined dataset achieved improved performance on each respective test set: sensitivity improved from 94% to 98% on the North American test set and from 73% to 82% on the Nepali test set.

CONCLUSIONS

A CNN can identify accurately the presence of ROP stage in retinal images, but performance depends on the similarity between training and testing populations. We demonstrated that internal and external performance can be improved by increasing the heterogeneity of the training dataset features of the training dataset, in this case by combining images from different populations and cameras.

Radiomics in PET Imaging:: A Practical Guide for Newcomers

Radiomics has undergone considerable development in recent years. In PET imaging, very promising results concerning the ability of handcrafted features to predict the biological characteristics of lesions and to assess patient prognosis or response to treatment have been reported in the literature. This article presents a checklist for designing a reliable radiomic study, gives an overview of the steps of the pipeline, and outlines approaches for data harmonization. Tips are provided for critical reading of the content of articles. The advantages and limitations of handcrafted radiomics compared with deep-learning approaches for the characterization of PET images are also discussed.

Applications of Machine Learning in Bone and Mineral Research

In this unprecedented era of the overwhelming volume of medical data, machine learning can be a promising tool that may shed light on an individualized approach and a better understanding of the disease in the field of osteoporosis research, similar to that in other research fields. This review aimed to provide an overview of the latest studies using machine learning to address issues, mainly focusing on osteoporosis and fractures. Machine learning models for diagnosing and classifying osteoporosis and detecting fractures from images have shown promising performance. Fracture risk prediction is another promising field of research, and studies are being conducted using various data sources. However, these approaches may be biased due to the nature of the techniques or the quality of the data. Therefore, more studies based on the proposed guidelines are needed to improve the technical feasibility and generalizability of artificial intelligence algorithms.

Effect of Patient Clinical Variables in Osteoporosis Classification Using Hip X-rays in Deep Learning Analysis

Background and Objectives: A few deep learning studies have reported that combining image features with patient variables enhanced identification accuracy compared with image-only models. However, previous studies have not statistically reported the additional effect of patient variables on the image-only models. This study aimed to statistically evaluate the osteoporosis identification ability of deep learning by combining hip radiographs with patient variables. Materials andMethods: We collected a dataset containing 1699 images from patients who underwent skeletal-bone-mineral density measurements and hip radiography at a general hospital from 2014 to 2021. Osteoporosis was assessed from hip radiographs using convolutional neural network (CNN) models (ResNet18, 34, 50, 101, and 152). We also investigated ensemble models with patient clinical variables added to each CNN. Accuracy, precision, recall, specificity, F1 score, and area under the curve (AUC) were calculated as performance metrics. Furthermore, we statistically compared the accuracy of the image-only model with that of an ensemble model that included images plus patient factors, including effect size for each performance metric. Results: All metrics were improved in the ResNet34 ensemble model compared with the image-only model. The AUC score in the ensemble model was significantly improved compared with the image-only model (difference 0.004; 95% CI 0.002–0.0007; p = 0.0004, effect size: 0.871). Conclusions: This study revealed the additional effect of patient variables in identification of osteoporosis using deep CNNs with hip radiographs. Our results provided evidence that the patient variables had additive synergistic effects on the image in osteoporosis identification.

A comparison of attentional neural network architectures for modeling with electronic medical records

Objective

Attention networks learn an intelligent weighted averaging mechanism over a series of entities, providing increases to both performance and interpretability. In this article, we propose a novel time-aware transformer-based network and compare it to another leading model with similar characteristics. We also decompose model performance along several critical axes and examine which features contribute most to our model's performance.

Materials and methods

Using data sets representing patient records obtained between 2017 and 2019 by the Kaiser Permanente Mid-Atlantic States medical system, we construct four attentional models with varying levels of complexity on two targets (patient mortality and hospitalization). We examine how incorporating transfer learning and demographic features contribute to model success. We also test the performance of a model proposed in recent medical modeling literature. We compare these models with out-of-sample data using the area under the receiver-operator characteristic (AUROC) curve and average precision as measures of performance. We also analyze the attentional weights assigned by these models to patient diagnoses.

Results

We found that our model significantly outperformed the alternative on a mortality prediction task (91.96% AUROC against 73.82% AUROC). Our model also outperformed on the hospitalization task, although the models were significantly more competitive in that space (82.41% AUROC against 80.33% AUROC). Furthermore, we found that demographic features and transfer learning features which are frequently omitted from new models proposed in the EMR modeling space contributed significantly to the success of our model.

Discussion

We proposed an original construction of deep learning electronic medical record models which achieved very strong performance. We found that our unique model construction outperformed on several tasks in comparison to a leading literature alternative, even when input data was held constant between them. We obtained further improvements by incorporating several methods that are frequently overlooked in new model proposals, suggesting that it will be useful to explore these options further in the future.

External Validation of Deep Learning Algorithm for Detecting and Visualizing Femoral Neck Fracture Including Displaced and Non-displaced Fracture on Plain X-ray

This study aimed to develop a method for detection of femoral neck fracture (FNF) including displaced and non-displaced fractures using convolutional neural network (CNN) with plain X-ray and to validate its use across hospitals through internal and external validation sets. This is a retrospective study using hip and pelvic anteroposterior films for training and detecting femoral neck fracture through residual neural network (ResNet) 18 with convolutional block attention module (CBAM) +  + . The study was performed at two tertiary hospitals between February and May 2020 and used data from January 2005 to December 2018. Our primary outcome was favorable performance for diagnosis of femoral neck fracture from negative studies in our dataset. We described the outcomes as area under the receiver operating characteristic curve (AUC), accuracy, Youden index, sensitivity, and specificity. A total of 4,189 images that contained 1,109 positive images (332 non-displaced and 777 displaced) and 3,080 negative images were collected from two hospitals. The test values after training with one hospital dataset were 0.999 AUC, 0.986 accuracy, 0.960 Youden index, and 0.966 sensitivity, and 0.993 specificity. Values of external validation with the other hospital dataset were 0.977, 0.971, 0.920, 0.939, and 0.982, respectively. Values of merged hospital datasets were 0.987, 0.983, 0.960, 0.973, and 0.987, respectively. A CNN algorithm for FNF detection in both displaced and non-displaced fractures using plain X-rays could be used in other hospitals to screen for FNF after training with images from the hospital of interest.

Application of convolutional neural networks for distal radio-ulnar fracture detection on plain radiographs in the emergency room

OBJECTIVE

Recent studies have suggested that deep-learning models can satisfactorily assist in fracture diagnosis. We aimed to evaluate the performance of two of such models in wrist fracture detection.

METHODS

We collected image data of patients who visited with wrist trauma at the emergency department. A dataset extracted from January 2018 to May 2020 was split into training (90%) and test (10%) datasets, and two types of convolutional neural networks (i.e., DenseNet-161 and ResNet-152) were trained to detect wrist fractures. Gradient-weighted class activation mapping was used to highlight the regions of radiograph scans that contributed to the decision of the model. Performance of the convolutional neural network models was evaluated using the area under the receiver operating characteristic curve.

RESULTS

For model training, we used 4,551 radiographs from 798 patients and 4,443 radiographs from 1,481 patients with and without fractures, respectively. The remaining 10% (300 radiographs from 100 patients with fractures and 690 radiographs from 230 patients without fractures) was used as a test dataset. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of DenseNet-161 and ResNet-152 in the test dataset were 90.3%, 90.3%, 80.3%, 95.6%, and 90.3% and 88.6%, 88.4%, 76.9%, 94.7%, and 88.5%, respectively. The area under the receiver operating characteristic curves of DenseNet-161 and ResNet-152 for wrist fracture detection were 0.962 and 0.947, respectively.

CONCLUSION

We demonstrated that DenseNet-161 and ResNet-152 models could help detect wrist fractures in the emergency room with satisfactory performance.

Multiomic Big Data Analysis Challenges: Increasing Confidence in the Interpretation of Artificial Intelligence Assessments

The need for holistic molecular measurements to better understand disease initiation, development, diagnosis, and therapy has led to an increasing number of multiomic analyses. The wealth of information available from multiomic assessments, however, requires both the evaluation and interpretation of extremely large data sets, limiting analysis throughput and ease of adoption. Computational methods utilizing artificial intelligence (AI) provide the most promising way to address these challenges, yet despite the conceptual benefits of AI and its successful application in singular omic studies, the widespread use of AI in multiomic studies remains limited. Here, we discuss present and future capabilities of AI techniques in multiomic studies while introducing analytical checks and balances to validate the computational conclusions.

Adaptive adversarial neural networks for the analysis of lossy and domain-shifted datasets of medical images

In machine learning for image-based medical diagnostics, supervised convolutional neural networks are typically trained with large and expertly annotated datasets obtained using high-resolution imaging systems. Moreover, the network's performance can degrade substantially when applied to a dataset with a different distribution. Here, we show that adversarial learning can be used to develop high-performing networks trained on unannotated medical images of varying image quality. Specifically, we used low-quality images acquired using inexpensive portable optical systems to train networks for the evaluation of human embryos, the quantification of human sperm morphology and the diagnosis of malarial infections in the blood, and show that the networks performed well across different data distributions. We also show that adversarial learning can be used with unlabelled data from unseen domain-shifted datasets to adapt pretrained supervised networks to new distributions, even when data from the original distribution are not available. Adaptive adversarial networks may expand the use of validated neural-network models for the evaluation of data collected from multiple imaging systems of varying quality without compromising the knowledge stored in the network.

Explaining deep neural networks for knowledge discovery in electrocardiogram analysis

Deep learning-based tools may annotate and interpret medical data more quickly, consistently, and accurately than medical doctors. However, as medical doctors are ultimately responsible for clinical decision-making, any deep learning-based prediction should be accompanied by an explanation that a human can understand. We present an approach called electrocardiogram gradient class activation map (ECGradCAM), which is used to generate attention maps and explain the reasoning behind deep learning-based decision-making in ECG analysis. Attention maps may be used in the clinic to aid diagnosis, discover new medical knowledge, and identify novel features and characteristics of medical tests. In this paper, we showcase how ECGradCAM attention maps can unmask how a novel deep learning model measures both amplitudes and intervals in 12-lead electrocardiograms, and we show an example of how attention maps may be used to develop novel ECG features.

Quantification of BERT Diagnosis Generalizability Across Medical Specialties Using Semantic Dataset Distance

Deep learning models in healthcare may fail to generalize on data from unseen corpora. Additionally, no quantitative metric exists to tell how existing models will perform on new data. Previous studies demonstrated that NLP models of medical notes generalize variably between institutions, but ignored other levels of healthcare organization. We measured SciBERT diagnosis sentiment classifier generalizability between medical specialties using EHR sentences from MIMIC-III. Models trained on one specialty performed better on internal test sets than mixed or external test sets (mean AUCs 0.92, 0.87, and 0.83, respectively; p = 0.016). When models are trained on more specialties, they have better test performances (p < 1e-4). Model performance on new corpora is directly correlated to the similarity between train and test sentence content (p < 1e-4). Future studies should assess additional axes of generalization to ensure deep learning models fulfil their intended purpose across institutions, specialties, and practices.

Differential diagnosis of benign and malignant vertebral fracture on CT using deep learning

OBJECTIVES

To evaluate the performance of deep learning using ResNet50 in differentiation of benign and malignant vertebral fracture on CT.

METHODS

A dataset of 433 patients confirmed with 296 malignant and 137 benign fractures was retrospectively selected from our spinal CT image database. A senior radiologist performed visual reading to evaluate six imaging features, and three junior radiologists gave diagnostic prediction. A ROI was placed on the most abnormal vertebrae, and the smallest square bounding box was generated. The input channel into ResNet50 network was 3, including the slice with its two neighboring slices. The diagnostic performance was evaluated using 10-fold cross-validation. After obtaining the malignancy probability from all slices in a patient, the highest probability was assigned to that patient to give the final diagnosis, using the threshold of 0.5.

RESULTS

Visual features such as soft tissue mass and bone destruction were highly suggestive of malignancy; the presence of a transverse fracture line was highly suggestive of a benign fracture. The reading by three radiologists with 5, 3, and 1 year of experience achieved an accuracy of 99%, 95.2%, and 92.8%, respectively. In ResNet50 analysis, the per-slice diagnostic sensitivity, specificity, and accuracy were 0.90, 0.79, and 85%. When the slices were combined to ve per-patient diagnosis, the sensitivity, specificity, and accuracy were 0.95, 0.80, and 88%.

CONCLUSION

Deep learning has become an important tool for the detection of fractures on CT. In this study, ResNet50 achieved good accuracy, which can be further improved with more cases and optimized methods for future clinical implementation.

KEY POINTS

• Deep learning using ResNet50 can yield a high accuracy for differential diagnosis of benign and malignant vertebral fracture on CT. • The per-slice diagnostic sensitivity, specificity, and accuracy were 0.90, 0.79, and 85% in deep learning using ResNet50 analysis. • The slices combined with per-patient diagnostic sensitivity, specificity, and accuracy were 0.95, 0.80, and 88% in deep learning using ResNet50 analysis.

Machine learning of EEG spectra classifies unconsciousness during GABAergic anesthesia

In current anesthesiology practice, anesthesiologists infer the state of unconsciousness without directly monitoring the brain. Drug- and patient-specific electroencephalographic (EEG) signatures of anesthesia-induced unconsciousness have been identified previously. We applied machine learning approaches to construct classification models for real-time tracking of unconscious state during anesthesia-induced unconsciousness. We used cross-validation to select and train the best performing models using 33,159 2s segments of EEG data recorded from 7 healthy volunteers who received increasing infusions of propofol while responding to stimuli to directly assess unconsciousness. Cross-validated models of unconsciousness performed very well when tested on 13,929 2s EEG segments from 3 left-out volunteers collected under the same conditions (median volunteer AUCs 0.99-0.99). Models showed strong generalization when tested on a cohort of 27 surgical patients receiving solely propofol collected in a separate clinical dataset under different circumstances and using different hardware (median patient AUCs 0.95-0.98), with model predictions corresponding with actions taken by the anesthesiologist during the cases. Performance was also strong for 17 patients receiving sevoflurane (alone or in addition to propofol) (median AUCs 0.88-0.92). These results indicate that EEG spectral features can predict unconsciousness, even when tested on a different anesthetic that acts with a similar neural mechanism. With high performance predictions of unconsciousness, we can accurately monitor anesthetic state, and this approach may be used to engineer infusion pumps to intelligibly respond to patients' neural activity.

Assessment of an AI Aid in Detection of Adult Appendicular Skeletal Fractures by Emergency Physicians and Radiologists: A Multicenter Cross-sectional Diagnostic Study

Background The interpretation of radiographs suffers from an ever-increasing workload in emergency and radiology departments, while missed fractures represent up to 80% of diagnostic errors in the emergency department. Purpose To assess the performance of an artificial intelligence (AI) system designed to aid radiologists and emergency physicians in the detection and localization of appendicular skeletal fractures. Materials and Methods The AI system was previously trained on 60 170 radiographs obtained in patients with trauma. The radiographs were randomly split into 70% training, 10% validation, and 20% test sets. Between 2016 and 2018, 600 adult patients in whom multiview radiographs had been obtained after a recent trauma, with or without one or more fractures of shoulder, arm, hand, pelvis, leg, and foot, were retrospectively included from 17 French medical centers. Radiographs with quality precluding human interpretation or containing only obvious fractures were excluded. Six radiologists and six emergency physicians were asked to detect and localize fractures with (n = 300) and fractures without (n = 300) the aid of software highlighting boxes around AI-detected fractures. Aided and unaided sensitivity, specificity, and reading times were compared by means of paired Student t tests after averaging of performances of each reader. Results A total of 600 patients (mean age ± standard deviation, 57 years ± 22; 358 women) were included. The AI aid improved the sensitivity of physicians by 8.7% (95% CI: 3.1, 14.2; P = .003 for superiority) and the specificity by 4.1% (95% CI: 0.5, 7.7; P < .001 for noninferiority) and reduced the average number of false-positive fractures per patient by 41.9% (95% CI: 12.8, 61.3; P = .02) in patients without fractures and the mean reading time by 15.0% (95% CI: -30.4, 3.8; P = .12). Finally, stand-alone performance of a newer release of the AI system was greater than that of all unaided readers, including skeletal expert radiologists, with an area under the receiver operating characteristic curve of 0.94 (95% CI: 0.92, 0.96). Conclusion The artificial intelligence aid provided a gain of sensitivity (8.7% increase) and specificity (4.1% increase) without loss of reading speed. © RSNA, 2021 Online supplemental material is available for this article.

Machine Learning Solutions for Osteoporosis-A Review

Osteoporosis and its clinical consequence, bone fracture, is a multifactorial disease that has been the object of extensive research. Recent advances in machine learning (ML) have enabled the field of artificial intelligence (AI) to make impressive breakthroughs in complex data environments where human capacity to identify high-dimensional relationships is limited. The field of osteoporosis is one such domain, notwithstanding technical and clinical concerns regarding the application of ML methods. This qualitative review is intended to outline some of these concerns and to inform stakeholders interested in applying AI for improved management of osteoporosis. A systemic search in PubMed and Web of Science resulted in 89 studies for inclusion in the review. These covered one or more of four main areas in osteoporosis management: bone properties assessment (n = 13), osteoporosis classification (n = 34), fracture detection (n = 32), and risk prediction (n = 14). Reporting and methodological quality was determined by means of a 12-point checklist. In general, the studies were of moderate quality with a wide range (mode score 6, range 2 to 11). Major limitations were identified in a significant number of studies. Incomplete reporting, especially over model selection, inadequate splitting of data, and the low proportion of studies with external validation were among the most frequent problems. However, the use of images for opportunistic osteoporosis diagnosis or fracture detection emerged as a promising approach and one of the main contributions that ML could bring to the osteoporosis field. Efforts to develop ML-based models for identifying novel fracture risk factors and improving fracture prediction are additional promising lines of research. Some studies also offered insights into the potential for model-based decision-making. Finally, to avoid some of the common pitfalls, the use of standardized checklists in developing and sharing the results of ML models should be encouraged. © 2021 American Society for Bone and Mineral Research (ASBMR).

Critical evaluation of deep neural networks for wrist fracture detection

Wrist Fracture is the most common type of fracture with a high incidence rate. Conventional radiography (i.e. X-ray imaging) is used for wrist fracture detection routinely, but occasionally fracture delineation poses issues and an additional confirmation by computed tomography (CT) is needed for diagnosis. Recent advances in the field of Deep Learning (DL), a subfield of Artificial Intelligence (AI), have shown that wrist fracture detection can be automated using Convolutional Neural Networks. However, previous studies did not pay close attention to the difficult cases which can only be confirmed via CT imaging. In this study, we have developed and analyzed a state-of-the-art DL-based pipeline for wrist (distal radius) fracture detection-DeepWrist, and evaluated it against one general population test set, and one challenging test set comprising only cases requiring confirmation by CT. Our results reveal that a typical state-of-the-art approach, such as DeepWrist, while having a near-perfect performance on the general independent test set, has a substantially lower performance on the challenging test set-average precision of 0.99 (0.99-0.99) versus 0.64 (0.46-0.83), respectively. Similarly, the area under the ROC curve was of 0.99 (0.98-0.99) versus 0.84 (0.72-0.93), respectively. Our findings highlight the importance of a meticulous analysis of DL-based models before clinical use, and unearth the need for more challenging settings for testing medical AI systems.

Better Reporting of Studies on Artificial Intelligence: CONSORT-AI and Beyond

An increasing number of studies on artificial intelligence (AI) are published in the dental and oral sciences. The reporting, but also further aspects of these studies, suffer from a range of limitations. Standards towards reporting, like the recently published Consolidated Standards of Reporting Trials (CONSORT)-AI extension can help to improve studies in this emerging field, and the Journal of Dental Research (JDR) encourages authors, reviewers, and readers to adhere to these standards. Notably, though, a wide range of aspects beyond reporting, located along various steps of the AI lifecycle, should be considered when conceiving, conducting, reporting, or evaluating studies on AI in dentistry.

Development and validation of a predictive nomogram for postoperative osteonecrosis of the femoral head with cannulated screws fixation

Background

Osteonecrosis of the femoral head (ONFH) remains a major complication of femoral neck fractures. Early interventions require preliminary prediction and detection. In this study, we aimed to evaluate the perioperative variables of postoperative ONFH in femoral neck fracture patients with closed reduction and cannulated screw fixation. We also established and validated an individualized nomogram for the prediction of postoperative ONFH.

Methods

We included 470 patients with ONFH from two hospitals [First Affiliated Hospital of University of Science and Technology of China (n=360) and Southern Branch of the First Affiliated Hospital of the University of Science and Technology of China (n=110)]. We evaluated the prognostic value of multiple perioperative variables using a Cox regression model in the training cohort. We developed a nomogram for the prediction of ONFH using a logistic regression model. We assessed the performance of this nomogram in a validation cohort and evaluated its clinical value.

Results

Of the 470 patients who met the inclusion criteria, 141 (30.0%) developed postoperative ONFH. We found alcohol use [odds ratio (OR), 1.743, 95% confidence interval (CI), 1.042-2.901, P=0.033], cerebrovascular disease (OR, 5.357, 95% CI, 2.318-13.13, P<0.001), interval to surgery (OR, 5.273, 95% CI, 2.724-10.43, P<0.001), Garden classification (OR, 23.17, 95% CI, 6.812-145.3, P<0.001), Garden index (OR, 5.935, 95% CI, 2.670-14.184, P<0.001), interval to partial weight-bearing (OR, 0.053, 95% CI, 0.006-0.296, P=0.002), and six-month Harris hip score (OR, 0.856; 95% CI, 0.792-0.919, P<0.001) were independent predictors of postoperative development of ONFH. Based on these variables, we developed a nomogram that showed good discrimination in both the training [area under the curve (AUC) =0.865] and the validation cohort (AUC =0.877). The favorable performance of this nomogram was also confirmed in the validation cohort.

Conclusions

We developed and validated an easy-to-use nomogram for predicting postoperative ONFH. This nomogram can aid decision-making of intraoperative interventions and postoperative rehabilitation plans for patients, surgeons, and osteo-rehabilitative physicians.

Informative presence and observation in routine health data: A review of methodology for clinical risk prediction

Objective

Informative presence (IP) is the phenomenon whereby the presence or absence of patient data is potentially informative with respect to their health condition, with informative observation (IO) being the longitudinal equivalent. These phenomena predominantly exist within routinely collected healthcare data, in which data collection is driven by the clinical requirements of patients and clinicians. The extent to which IP and IO are considered when using such data to develop clinical prediction models (CPMs) is unknown, as is the existing methodology aiming at handling these issues. This review aims to synthesize such existing methodology, thereby helping identify an agenda for future methodological work.

Materials and Methods

A systematic literature search was conducted by 2 independent reviewers using prespecified keywords.

Results

Thirty-six articles were included. We categorized the methods presented within as derived predictors (including some representation of the measurement process as a predictor in the model), modeling under IP, and latent structures. Including missing indicators or summary measures as predictors is the most commonly presented approach amongst the included studies (24 of 36 articles).

Discussion

This is the first review to collate the literature in this area under a prediction framework. A considerable body relevant of literature exists, and we present ways in which the described methods could be developed further. Guidance is required for specifying the conditions under which each method should be used to enable applied prediction modelers to use these methods.

Conclusions

A growing recognition of IP and IO exists within the literature, and methodology is increasingly becoming available to leverage these phenomena for prediction purposes. IP and IO should be approached differently in a prediction context than when the primary goal is explanation. The work included in this review has demonstrated theoretical and empirical benefits of incorporating IP and IO, and therefore we recommend that applied health researchers consider incorporating these methods in their work.

Artificial intelligence in radiology

The interest in artificial intelligence (AI) has ballooned within radiology in the past few years primarily due to notable successes of deep learning. With the advances brought by deep learning, AI has the potential to recognize and localize complex patterns from different radiological imaging modalities, many of which even achieve comparable performance to human decision-making in recent applications. In this chapter, we review several AI applications in radiology for different anatomies: chest, abdomen, pelvis, as well as general lesion detection/identification that is not limited to specific anatomies. For each anatomy site, we focus on introducing the tasks of detection, segmentation, and classification with an emphasis on describing the technology development pathway with the aim of providing the reader with an understanding of what AI can do in radiology and what still needs to be done for AI to better fit in radiology. Combining with our own research experience of AI in medicine, we elaborate how AI can enrich knowledge discovery, understanding, and decision-making in radiology, rather than replacing the radiologist.

Automated classification of hip fractures using deep convolutional neural networks with orthopedic surgeon-level accuracy: ensemble decision-making with antero-posterior and lateral radiographs

Background and purpose - Deep-learning approaches based on convolutional neural networks (CNNs) are gaining interest in the medical imaging field. We evaluated the diagnostic performance of a CNN to discriminate femoral neck fractures, trochanteric fractures, and non-fracture using antero-posterior (AP) and lateral hip radiographs. Patients and methods - 1,703 plain hip AP radiographs and 1,220 plain hip lateral radiographs were included in the total dataset. 150 images each of the AP and lateral views were separated out and the remainder of the dataset was used for training. The CNN made the diagnosis based on: (1) AP radiographs alone, (2) lateral radiographs alone, or (3) both AP and lateral radiographs combined. The diagnostic performance of the CNN was measured by the accuracy, recall, precision, and F1 score. We further compared the CNN's performance with that of orthopedic surgeons. Results - The average accuracy, recall, precision, and F1 score of the CNN based on both anteroposterior and lateral radiographs were 0.98, 0.98, 0.98, and 0.98, respectively. The accuracy of the CNN was comparable to, or statistically significantly better than, that of the orthopedic surgeons regardless of radiographic view used. In the CNN model, the accuracy of the diagnosis based on both views was significantly better than the lateral view alone and tended to be better than the AP view alone. Interpretation - The CNN exhibited comparable or superior performance to that of orthopedic surgeons to discriminate femoral neck fractures, trochanteric fractures, and non-fracture using both AP and lateral hip radiographs.