Subchondral Bone Length in Knee Osteoarthritis: A Deep Learning-Derived Imaging Measure and Its Association With Radiographic and Clinical Outcomes.

OBJECTIVE: To develop a bone shape measure that reflects the extent of cartilage loss and bone flattening in knee osteoarthritis (OA) and test it against estimates of disease severity. METHODS: A fast region-based convolutional neural network was trained to crop the knee joints in sagittal dual-echo steady-state magnetic resonance imaging sequences obtained from the Osteoarthritis Initiative (OAI). Publicly available annotations of the cartilage and menisci were used as references to annotate the tibia and the femur in 61 knees. Another deep neural network (U-Net) was developed to learn these annotations. Model predictions were compared to radiologist-driven annotations on an independent test set (27 knees). The U-Net was applied to automatically extract the knee joint structures on the larger OAI data set (n = 9,434 knees). We defined subchondral bone length (SBL), a novel shape measure characterizing the extent of overlying cartilage and bone flattening, and examined its relationship with radiographic joint space narrowing (JSN), concurrent pain and disability (according to the Western Ontario and McMaster Universities Osteoarthritis Index), as well as subsequent partial or total knee replacement. Odds ratios (ORs) and 95% confidence intervals (95% CIs) for each outcome were estimated using relative changes in SBL from the OAI data set stratified into quartiles. RESULTS: The mean SBL values for knees with JSN were consistently different from knees without JSN. Greater changes of SBL from baseline were associated with greater pain and disability. For knees with medial or lateral JSN, the ORs for future knee replacement between the lowest and highest quartiles corresponding to SBL changes were 5.68 (95% CI 3.90-8.27) and 7.19 (95% CI 3.71-13.95), respectively. CONCLUSION: SBL quantified OA status based on JSN severity and shows promise as an imaging marker in predicting clinical and structural OA outcomes.

Correction to: Assessment of knee pain from MR imaging using a convolutional Siamese network.

The original version of this article, published on 13 February 2020, unfortunately contained a mistake.

Development and validation of an interpretable deep learning framework for Alzheimer's disease classification.

Alzheimer's disease is the primary cause of dementia worldwide, with an increasing morbidity burden that may outstrip diagnosis and management capacity as the population ages. Current methods integrate patient history, neuropsychological testing and MRI to identify likely cases, yet effective practices remain variably applied and lacking in sensitivity and specificity. Here we report an interpretable deep learning strategy that delineates unique Alzheimer's disease signatures from multimodal inputs of MRI, age, gender, and Mini-Mental State Examination score. Our framework linked a fully convolutional network, which constructs high resolution maps of disease probability from local brain structure to a multilayer perceptron and generates precise, intuitive visualization of individual Alzheimer's disease risk en route to accurate diagnosis. The model was trained using clinically diagnosed Alzheimer's disease and cognitively normal subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset (n = 417) and validated on three independent cohorts: the Australian Imaging, Biomarker and Lifestyle Flagship Study of Ageing (AIBL) (n = 382), the Framingham Heart Study (n = 102), and the National Alzheimer's Coordinating Center (NACC) (n = 582). Performance of the model that used the multimodal inputs was consistent across datasets, with mean area under curve values of 0.996, 0.974, 0.876 and 0.954 for the ADNI study, AIBL, Framingham Heart Study and NACC datasets, respectively. Moreover, our approach exceeded the diagnostic performance of a multi-institutional team of practicing neurologists (n = 11), and high-risk cerebral regions predicted by the model closely tracked post-mortem histopathological findings. This framework provides a clinically adaptable strategy for using routinely available imaging techniques such as MRI to generate nuanced neuroimaging signatures for Alzheimer's disease diagnosis, as well as a generalizable approach for linking deep learning to pathophysiological processes in human disease.

Evolutionary Selection and Constraint on Human Knee Chondrocyte Regulation Impacts Osteoarthritis Risk.

During human evolution, the knee adapted to the biomechanical demands of bipedalism by altering chondrocyte developmental programs. This adaptive process was likely not without deleterious consequences to health. Today, osteoarthritis occurs in 250 million people, with risk variants enriched in non-coding sequences near chondrocyte genes, loci that likely became optimized during knee evolution. We explore this relationship by epigenetically profiling joint chondrocytes, revealing ancient selection and recent constraint and drift on knee regulatory elements, which also overlap osteoarthritis variants that contribute to disease heritability by tending to modify constrained functional sequence. We propose a model whereby genetic violations to regulatory constraint, tolerated during knee development, lead to adult pathology. In support, we discover a causal enhancer variant (rs6060369) present in billions of people at a risk locus (GDF5-UQCC1), showing how it impacts mouse knee-shape and osteoarthritis. Overall, our methods link an evolutionarily novel aspect of human anatomy to its pathogenesis.

Assessment of knee pain from MR imaging using a convolutional Siamese network.

OBJECTIVES: It remains difficult to characterize the source of pain in knee joints either using radiographs or magnetic resonance imaging (MRI). We sought to determine if advanced machine learning methods such as deep neural networks could distinguish knees with pain from those without it and identify the structural features that are associated with knee pain. METHODS: We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral knee pain comparing the knee with frequent pain to the contralateral knee without pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions most associated with knee pain. The MRI scans and the CAMs of each subject were reviewed by an expert radiologist to identify the presence of abnormalities within the model-predicted regions of high association. RESULTS: Using 10-fold cross-validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose knee WOMAC pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with pain. CONCLUSIONS: This study demonstrates a proof of principle that deep learning can be applied to assess knee pain from MRI scans. KEY POINTS: • Our article is the first to leverage a deep learning framework to associate MR images of the knee with knee pain. • We developed a convolutional Siamese network that had the ability to fuse information from multiple two-dimensional (2D) MRI slices from the knee with pain and the contralateral knee of the same individual without pain to predict unilateral knee pain. • Our model achieved an area under curve (AUC) value of 0.808. When individuals who had WOMAC pain scores that were not discordant for knees (pain discordance < 3) were excluded, model performance increased to 0.853.

Intrinsic coating morphology modulates acute drug transfer in drug-coated balloon therapy.

The hallmark of drug-coated balloon (DCB) therapy for the treatment of peripheral vascular disease is that it allows for reopening of the narrowed lumen and local drug delivery without the need for a permanent indwelling metal implant such as a stent. Current DCB designs rely on transferring drugs such as paclitaxel to the arterial vessel using a variety of biocompatible excipients coated on the balloons. Inherent procedural challenges, along with limited understanding of the interactions between the coating and the artery, interactions between the coating and the balloon as well as site-specific differences, have led to DCB designs with poor drug delivery efficiency. Our study is focused on two clinically significant DCB excipients, urea and shellac, and uses uniaxial mechanical testing, scanning electron microscopy (SEM), and biophysical modeling based on classic Hertz theory to elucidate how coating microstructure governs the transmission of forces at the coating-artery interface. SEM revealed shellac-based coatings to contain spherical-shaped microstructural elements whereas urea-based coatings contained conical-shaped microstructural elements. Our model based on Hertz theory showed that the interactions between these intrinsic coating elements with the arterial wall were fundamentally different, even when the same external force was applied by the balloon on the arterial wall. Using two orthogonal cell-based assays, our study also found differential viability when endothelial cells were exposed to titrated concentrations of urea and shellac, further highlighting the need to maximize coating transfer efficiency in the context of DCB therapies. Our results underscore the significance of the excipient in DCB design and suggest that coating microstructure modulates acute drug transfer during device deployment.

Fusion of deep learning models of MRI scans, Mini-Mental State Examination, and logical memory test enhances diagnosis of mild cognitive impairment.

INTRODUCTION: Our aim was to investigate if the accuracy of diagnosing mild cognitive impairment (MCI) using the Mini-Mental State Examination (MMSE) and logical memory (LM) test could be enhanced by adding MRI data. METHODS: Data of individuals with normal cognition and MCI were obtained from the National Alzheimer Coordinating Center database (n = 386). Deep learning models trained on MRI slices were combined to generate a fused MRI model using different voting techniques to predict normal cognition versus MCI. Two multilayer perceptron (MLP) models were developed with MMSE and LM test results. Finally, the fused MRI model and the MLP models were combined using majority voting. RESULTS: The fusion model was superior to the individual models alone and achieved an overall accuracy of 90.9%. DISCUSSION: This study is a proof of principle that multimodal fusion of models developed using MRI scans, MMSE, and LM test data is feasible and can better predict MCI.

Uremic Solute-Aryl Hydrocarbon Receptor-Tissue Factor Axis Associates with Thrombosis after Vascular Injury in Humans.

Individuals with CKD are particularly predisposed to thrombosis after vascular injury. Using mouse models, we recently described indoxyl sulfate, a tryptophan metabolite retained in CKD and an activator of tissue factor (TF) through aryl hydrocarbon receptor (AHR) signaling, as an inducer of thrombosis across the CKD spectrum. However, the translation of findings from animal models to humans is often challenging. Here, we investigated the uremic solute-AHR-TF thrombosis axis in two human cohorts, using a targeted metabolomics approach to probe a set of tryptophan products and high-throughput assays to measure AHR and TF activity. Analysis of baseline serum samples was performed from 473 participants with advanced CKD from the Dialysis Access Consortium Clopidogrel Prevention of Early AV Fistula Thrombosis trial. Participants with subsequent arteriovenous thrombosis had significantly higher levels of indoxyl sulfate and kynurenine, another uremic solute, and greater activity of AHR and TF, than those without thrombosis. Pattern recognition analysis using the components of the thrombosis axis facilitated clustering of the thrombotic and nonthrombotic groups. We further validated these findings using 377 baseline samples from participants in the Thrombolysis in Myocardial Infarction II trial, many of whom had CKD stage 2-3. Mechanistic probing revealed that kynurenine enhances thrombosis after vascular injury in an animal model and regulates thrombosis in an AHR-dependent manner. This human validation of the solute-AHR-TF axis supports further studies probing its utility in risk stratification of patients with CKD and exploring its role in other diseases with heightened risk of thrombosis.