Sayres R, Taly A, Rahimy E, Blumer K, Coz D, Hammel N, Krause J, Narayanaswamy A, Rastegar Z, Wu D, Xu S, Barb S, Joseph A, Shumski M, Smith J, Sood AB, Corrado GS, Peng L, Webster DR.
Using a Deep Learning Algorithm and Integrated Gradients Explanation to Assist Grading for Diabetic Retinopathy. Ophthalmology 2019;126(4):552-564.
AbstractPURPOSE: To understand the impact of deep learning diabetic retinopathy (DR) algorithms on physician readers in computer-assisted settings. DESIGN: Evaluation of diagnostic technology. PARTICIPANTS: One thousand seven hundred ninety-six retinal fundus images from 1612 diabetic patients. METHODS: Ten ophthalmologists (5 general ophthalmologists, 4 retina specialists, 1 retina fellow) read images for DR severity based on the International Clinical Diabetic Retinopathy disease severity scale in each of 3 conditions: unassisted, grades only, or grades plus heatmap. Grades-only assistance comprised a histogram of DR predictions (grades) from a trained deep-learning model. For grades plus heatmap, we additionally showed explanatory heatmaps. MAIN OUTCOME MEASURES: For each experiment arm, we computed sensitivity and specificity of each reader and the algorithm for different levels of DR severity against an adjudicated reference standard. We also measured accuracy (exact 5-class level agreement and Cohen's quadratically weighted κ), reader-reported confidence (5-point Likert scale), and grading time. RESULTS: Readers graded more accurately with model assistance than without for the grades-only condition (P < 0.001). Grades plus heatmaps improved accuracy for patients with DR (P < 0.001), but reduced accuracy for patients without DR (P = 0.006). Both forms of assistance increased readers' sensitivity moderate-or-worse DR: unassisted: mean, 79.4% [95% confidence interval (CI), 72.3%-86.5%]; grades only: mean, 87.5% [95% CI, 85.1%-89.9%]; grades plus heatmap: mean, 88.7% [95% CI, 84.9%-92.5%] without a corresponding drop in specificity (unassisted: mean, 96.6% [95% CI, 95.9%-97.4%]; grades only: mean, 96.1% [95% CI, 95.5%-96.7%]; grades plus heatmap: mean, 95.5% [95% CI, 94.8%-96.1%]). Algorithmic assistance increased the accuracy of retina specialists above that of the unassisted reader or model alone; and increased grading confidence and grading time across all readers. For most cases, grades plus heatmap was only as effective as grades only. Over the course of the experiment, grading time decreased across all conditions, although most sharply for grades plus heatmap. CONCLUSIONS: Deep learning algorithms can improve the accuracy of, and confidence in, DR diagnosis in an assisted read setting. They also may increase grading time, although these effects may be ameliorated with experience.
Schoemaker D, Quiroz YT, Torrico-Teave H, Arboleda-Velasquez JF.
Clinical and research applications of magnetic resonance imaging in the study of CADASIL. Neurosci Lett 2019;698:173-179.
AbstractCerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy (CADASIL) is an inherited small vessel disease that leads to early cerebrovascular events and functional disability. It is the most common single-gene disorder leading to stroke. Magnetic resonance imaging (MRI) is a central component of the diagnosis and monitoring of CADASIL. Here we provide a descriptive review of the literature on three important aspects pertaining to the use of MRI in CADASIL. First, we review past research exploring MRI markers for this disease. Secondly, we describe results from studies investigating associations between neuroimaging abnormalities and neuropathology in CADASIL. Finally, we discuss previous findings relating MRI markers to clinical symptoms. This review thus provides a summary of the current state of knowledge regarding the use of MRI in CADASIL as well as suggestions for future research.
Silva RNE, Chiou CA, Wang M, Wang H, Shoji MK, Chou JC, D'Souza EE, Greenstein SH, Brauner SC, Alves MR, Pasquale LR, Shen LQ.
Microvasculature of the Optic Nerve Head and Peripapillary Region in Patients With Primary Open-Angle Glaucoma. J Glaucoma 2019;28(4):281-288.
AbstractPURPOSE: To assess optic nerve head (ONH) and peripapillary microvasculature in primary open-angle glaucoma (POAG) of mild to moderate severity using swept-source optical coherence tomography angiography (OCTA). MATERIALS AND METHODS: In a cross-sectional study, swept-source OCTA images were analyzed for 1 eye from each of 30 POAG patients with glaucomatous Humphrey visual field loss and 16 controls. The anatomic boundary of ONH was manually delineated based on Bruch's membrane opening and large vessels were removed from en face angiography images to measure vessel density (VD) and the integrated OCTA by ratio analysis signal (IOS), suggestive of flow, in the ONH and peripapillary region. POAG subgroup analysis was performed based on a history of disc hemorrhage (DH) matched by visual field mean deviation (MD). RESULTS: POAG (mean MD±SD, -3.3±3.0 dB) and control groups had similar demographic characteristics and intraocular pressure on the day of imaging. Groups did not differ in superficial ONH VD or flow indicated by IOS (P≥0.28). POAG eyes showed significantly lower VD (39.4%±4.0%) and flow (38.8%±5.6%) in deep ONH, peripapillary VD (37.9%±2.9%) and flow (43.6%±4.0%) compared with control eyes (44.1%±5.1%, 44.7%±6.9%, 40.7%±1.7%, 47.8%±2.5%, respectively; P≤0.007 for all). In the subgroup analysis, POAG eyes with (n=14) and without DH (n=16) had similar measured OCTA parameters (P>0.99 for all). CONCLUSIONS: The image processing methodology based on the anatomic boundary of ONH demonstrated compromised microvasculature in the deep ONH and peripapillary region in eyes with mild to moderate POAG, regardless of the history of DH.