Review
Copyright ©The Author(s) 2022.
World J Diabetes. Oct 15, 2022; 13(10): 822-834
Published online Oct 15, 2022. doi: 10.4239/wjd.v13.i10.822
Table 1 Comparative analysis of various studies done on artificial intelligence in diabetic retinopathy[19]
Ref.
Sensitivity, specificity or accuracy of the study
Total fundus images examined
Types of AI used
Main objective
Wong et al[20] Area under the curve were 0.97 and 0.92 for microaneurysm and hemorrhages respectively143 imagesA three-layer feed forward neural networkDeals with detecting the microaneurysm and hemorrhages. Frangi filter used
Imani et al[57]Sensitivity of 75.02%-75.24%; Specificity of 97.45%-97.53%60 imagesMCADetected the exudation and blood vessel
Yazid et al[58]97.8% in sensitivity, 99% in specificity and 83.3% in predictivity for STARE database. 90.7% in sensitivity, 99.4% in specificity and 74% in predictivity for the custom database30 imagesInverse surface thresholdingDetected both hard and soft exudates
Akyol et al[59]Percentage accuracy of disc detection ranged from 90%-94.38% using different data set239 imagesKey point detection, texture analysis, and visual dictionary techniquesDetected the optic disc of fundus images
Niemeijer et al[13]Accuracy in 99.9% cases in finding the disc1000 imagesCombined k-nearest neighbor and cuesFast detection of the optic disc
Rajalakshmi et al[60], Smart phone based study 95.8% sensitivity and 80.2% specificity for detecting any DR. 99.1% sensitivity and 80.4% specificity in detecting STDRRetinal images of 296 patientsEye Art AI Dr screening software usedRetinal photography with Remidio ‘Fundus on Phone’
Eye Nuk study Sensitivity was 91.7%; Specificity was 91.5%40542 imagesEye PAC Stelescreening systemRetinal images taken with traditional desktop fundus cameras
Ting et al[61]Sensitivity and specificity for RDR was 90.5% and 91.6%; For STDR the sensitivity was 100% and the specificity was 91.1%494661 retinal imagesDeep learning systemMultiple Retinal images taken with conventional fundus cameras
IRIS Sensitivity of the IRIS algorithm in detecting STDR was 66.4% with false-negative rate of 2% and the specificity was 72.8%. Positive Predictive value of 10.8% and negative predictive value 97.8%15015 patientsIntelligent Retinal Imaging System (IRIS)Retinal screening examination and nonmydriatic fundus photography
Table 2 Summary of artificial intelligence algorithm used in age-related macular degeneration
Ref.
Sensitivity
Specificity
Diagnostic accuracy
Output
Grassman et al[62]84.2094.3063.3, Kappa of 92%Final probability value for referable vs not referable
Ting et al[61]93.2088.70Area under curve-0.932Identifying referable AMD and advanced AMD
Lee et al[26]84.6091.5087.60Prediction of binary segmentation map
Treder et al[27]1009296AMD testing score-score of 0.98 or greater adequate for diagnosis of AMD
Table 3 Summary of studies using artificial intelligence to detect progression in Glaucomatous eyes
Ref.
No. of eyes
Instrument
Approach
Comments
Lin et al[63]80SAPSupervised MLSensitivity-86%; Specificity-88%
Goldbaum et al[64]478 suspects; 150 glaucoma; 55 stable glaucomaSAPUnsupervised MLSpecificity-98.4%, AROC not available; Use of variational Byesian. Independent component analysis mixture model in indentifying patterns of glaucomatous visual field defects and its validation
Wang et al[65]11817 (method developing cohort) and 397 (clinical evaluation cohort)SAPUnsupervised MLAROC of the archetype method 0.77
Yousefi et al[16]939 Abnormal SAP and 1146 normal SAP in the cross section and 270 glaucoma in the longitudinal databaseSAPUnsupervised MLSensitivity 34.5%-63.4% at specificity 87% Comment: it took 3.5 years for ML analysis to detect progression while it took over 3.5 years for other methods to detect progression in 25% of eyes
Belghith et al27- progressing; 26-stable glaucoma and 40 healthy controlsSD OCT Supervised MLSensitivity -78% Specificity in normal eyes-93%; 94% in non-progressive eyes