Table 2. A summary of studies depicting automated diagnosis of glaucoma using OCT.
| Paper | Classifier | Number | Training and testing | Results | OCT | Glaucoma diagnosis | Database |
| Burgansky-Eliash et al 2005[54] | Multiple-take SVM | 89, 47 G, 42 N | Six fold validation, leave one out | AUROC 0.981, no CI | Stratus OCT | Anderson Patella Criteria | Recruitment of Subjects, Pennsylvania |
| Bowd et al 2008[55] | RVM | 225, 156 G, 69 N | Tenfold cross validation | AUROC 0.809, no CI | Stratus OCT | Anderson Patella Criteria | Observational Cross Sectional Study, California |
| Bizios et al 2010[62] | SVM | 152, 62 G, 90 N | Tenfold cross validation | AUROC 0.977, CI 0.959-0.999 | Stratus OCT | Anderson Patella Criteria | Observational Cross Sectional Study, Citizens of Malmo Sweden |
| Barella et al 2013[60] | RAN | 103, 57 G, 46 N | Tenfold cross validation resampling | AUROC 0.877, CI 0.810-0.944 | Cirrus SD OCT | Anderson Patella Criteria | Glaucoma Service UNICAMP, Brazil, prospective, observational cross sectional |
| Silva et al 2013[61] | RAN | 110, 62 G, 48 N | Tenfold cross validation | AUROC 0.807, CI 0.721-0.876 | Cirrus HD OCT | Anderson Patella Criteria | Glaucoma Service UNICAMP, Brazil, observational cross sectional |
| Xu et al 2013[56] | Boosted logistic regression | 192, 148 G, 44N | Normative database, Tenfold cross validation | AUROC 0.903, no CI | Cirrus HD OCT | Anderson Patella Criteria | PITT trial, Pennsylvania |
| Muhammad et al 2017[57] | CNN | 102, 57 G, 45 N | Pretrained, leave one out cross validation | AUROC 0.945, CI 0.955-0.947 | Topcon OCT | Anderson Patella Criteria | From previous study for OCT and early glaucoma diagnosis, New York |
| Asaoka et al 2019[59] | SVM | 178, 94 G, 84 N | Pre training, glaucoma OCT database | AUROC 0.937, CI 0.906-0.968 | RS 3000 | Anderson Patella Criteria | Japanese Archives of Multicentral Images of Glaucomatous OCT database, Japan |
| Christopher et al 2018[58] | PCA | 235, 179 G, 56 N | Leave one out approach | AUROC 0.95, CI 0.92-0.98 | Spectralis OCT | Ill defined | DIGS dataset, California |
| An et al 2019[50] | CNN | 357, 208 G, 149 N | Tenfold cross validation | AUROC 0.963, Mean±SD 0.029 | Topcon OCT | Anderson Patella Criteria | Observational Cross Sectional Study, Japan |
G: Glaucoma; N: Normal; AUROC: Area under the receiver operating characteristics curve; CI: Confidence interval; CNN: Convolutional neural networks; ANN: Artificial neural network; SVM: Support vector machine; LS-SVM: Least squares support vectors machine; ADAGES: African descent and glaucoma evaluation study; DIGS: Diagnostic innovations in glaucoma study; PCA: Principal component analysis; RVM: Relevance vector machine.