FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2903896358> ?p ?o ?g. }

• W2903896358 endingPage "564" @default.
• W2903896358 startingPage "552" @default.
• W2903896358 abstract "PurposeTo understand the impact of deep learning diabetic retinopathy (DR) algorithms on physician readers in computer-assisted settings.DesignEvaluation of diagnostic technology.ParticipantsOne thousand seven hundred ninety-six retinal fundus images from 1612 diabetic patients.MethodsTen 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 MeasuresFor 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.ResultsReaders 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.ConclusionsDeep 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. To understand the impact of deep learning diabetic retinopathy (DR) algorithms on physician readers in computer-assisted settings. Evaluation of diagnostic technology. One thousand seven hundred ninety-six retinal fundus images from 1612 diabetic patients. 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. 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. 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. 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." @default.
• W2903896358 created "2018-12-22" @default.
• W2903896358 creator A5003577779 @default.
• W2903896358 creator A5012523906 @default.
• W2903896358 creator A5015709909 @default.
• W2903896358 creator A5039795308 @default.
• W2903896358 creator A5045850122 @default.
• W2903896358 creator A5049765587 @default.
• W2903896358 creator A5050136423 @default.
• W2903896358 creator A5058399428 @default.
• W2903896358 creator A5060000122 @default.
• W2903896358 creator A5060574229 @default.
• W2903896358 creator A5060909587 @default.
• W2903896358 creator A5063867870 @default.
• W2903896358 creator A5065263704 @default.
• W2903896358 creator A5068955381 @default.
• W2903896358 creator A5069391199 @default.
• W2903896358 creator A5076216352 @default.
• W2903896358 creator A5081619977 @default.
• W2903896358 creator A5086742649 @default.
• W2903896358 creator A5087579994 @default.
• W2903896358 date "2019-04-01" @default.
• W2903896358 modified "2023-10-10" @default.
• W2903896358 title "Using a Deep Learning Algorithm and Integrated Gradients Explanation to Assist Grading for Diabetic Retinopathy" @default.
• W2903896358 cites W1863821586 @default.
• W2903896358 cites W1975780514 @default.
• W2903896358 cites W1982722705 @default.
• W2903896358 cites W1984639731 @default.
• W2903896358 cites W2010120422 @default.
• W2903896358 cites W2017036902 @default.
• W2903896358 cites W2063052894 @default.
• W2903896358 cites W2073244572 @default.
• W2903896358 cites W2101490467 @default.
• W2903896358 cites W2114536324 @default.
• W2903896358 cites W2127322315 @default.
• W2903896358 cites W2138480916 @default.
• W2903896358 cites W2148309496 @default.
• W2903896358 cites W2234307896 @default.
• W2903896358 cites W2557738935 @default.
• W2903896358 cites W2590312224 @default.
• W2903896358 cites W2738975713 @default.
• W2903896358 cites W2742365000 @default.
• W2903896358 cites W2762741128 @default.
• W2903896358 cites W2768095311 @default.
• W2903896358 cites W2772246530 @default.
• W2903896358 cites W2785593065 @default.
• W2903896358 doi "https://doi.org/10.1016/j.ophtha.2018.11.016" @default.
• W2903896358 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/30553900" @default.
• W2903896358 hasPublicationYear "2019" @default.
• W2903896358 type Work @default.
• W2903896358 sameAs 2903896358 @default.
• W2903896358 citedByCount "247" @default.
• W2903896358 countsByYear W29038963582019 @default.
• W2903896358 countsByYear W29038963582020 @default.
• W2903896358 countsByYear W29038963582021 @default.
• W2903896358 countsByYear W29038963582022 @default.
• W2903896358 countsByYear W29038963582023 @default.
• W2903896358 crossrefType "journal-article" @default.
• W2903896358 hasAuthorship W2903896358A5003577779 @default.
• W2903896358 hasAuthorship W2903896358A5012523906 @default.
• W2903896358 hasAuthorship W2903896358A5015709909 @default.
• W2903896358 hasAuthorship W2903896358A5039795308 @default.
• W2903896358 hasAuthorship W2903896358A5045850122 @default.
• W2903896358 hasAuthorship W2903896358A5049765587 @default.
• W2903896358 hasAuthorship W2903896358A5050136423 @default.
• W2903896358 hasAuthorship W2903896358A5058399428 @default.
• W2903896358 hasAuthorship W2903896358A5060000122 @default.
• W2903896358 hasAuthorship W2903896358A5060574229 @default.
• W2903896358 hasAuthorship W2903896358A5060909587 @default.
• W2903896358 hasAuthorship W2903896358A5063867870 @default.
• W2903896358 hasAuthorship W2903896358A5065263704 @default.
• W2903896358 hasAuthorship W2903896358A5068955381 @default.
• W2903896358 hasAuthorship W2903896358A5069391199 @default.
• W2903896358 hasAuthorship W2903896358A5076216352 @default.
• W2903896358 hasAuthorship W2903896358A5081619977 @default.
• W2903896358 hasAuthorship W2903896358A5086742649 @default.
• W2903896358 hasAuthorship W2903896358A5087579994 @default.
• W2903896358 hasBestOaLocation W29038963581 @default.
• W2903896358 hasConcept C11413529 @default.
• W2903896358 hasConcept C118487528 @default.
• W2903896358 hasConcept C126322002 @default.
• W2903896358 hasConcept C127413603 @default.
• W2903896358 hasConcept C134018914 @default.
• W2903896358 hasConcept C141071460 @default.
• W2903896358 hasConcept C147176958 @default.
• W2903896358 hasConcept C2776391266 @default.
• W2903896358 hasConcept C2777286243 @default.
• W2903896358 hasConcept C2779829184 @default.
• W2903896358 hasConcept C2993012660 @default.
• W2903896358 hasConcept C33923547 @default.
• W2903896358 hasConcept C44249647 @default.
• W2903896358 hasConcept C555293320 @default.
• W2903896358 hasConcept C71924100 @default.
• W2903896358 hasConceptScore W2903896358C11413529 @default.
• W2903896358 hasConceptScore W2903896358C118487528 @default.
• W2903896358 hasConceptScore W2903896358C126322002 @default.
• W2903896358 hasConceptScore W2903896358C127413603 @default.
• W2903896358 hasConceptScore W2903896358C134018914 @default.

• next