Abstract
Purpose: :
To evaluate and compare the diagnostic agreement and performance for detection of glaucomatous progression of HD-OCT Guided Progression Analysis (GPA), standard automated perimetry (SAP) GPA and expert assessment of optic disc stereophotographs.
Methods: :
This Observational cohort study included 154 eyes from 108 individuals followed for an average of 29.36 ± 8 months. At baseline, 38.3% of the eyes were classified as glaucomatous, 58.4% of the eyes were classified as suspects, and 3.2% as healthy eyes. Eyes with any other ocular disease except for mild cataract were excluded. Images were obtained every 6 to 12 month with the Cirrus Optic Disc Cube algorithm 200x200, along with the optic disc stereophotographs and SAP visual fields. Progression was determined in visual fields by event and trend analysis with GPA II software, in HD-OCT images with GPA and by masked assessment of optic disc stereophotographs by expert graders. Images obtained with HD-OCT were analyzed for progression using the GPA software. Main outcome measures were sensitivity, specificity, likelihood ratio of positive test result (LR+), likelihood ratio of negative test result (LR-) and area under the receiver operating characteristic curve (AUC) for detection of glaucoma progression using OCT GPA and SAP GPA. The assessment of optic disc stereophotographs was used as reference standard. Agreement among the different methods was reported using the Kappa coefficient.
Results: :
Seven of the 154 eyes (4.5%) showed progression by stereophotographs, during the follow-up period. OCT GPA detected 3 of these eyes and SAP GPA only 2, with a sensitivity of 42.9% and 28.6%, respectively. Thirty-nine eyes showed progression only by OCT GPA with a specificity of 70.7% and 20 eyes showed progression only by SAP GPA with a specificity of 83.7%. The OCT GPA software obtained similar AUC (0.568) than the SAP GPA (0.561). OCT GPA LR+ and LR- values were 1.464 and 0.807, while SAP GPA LR+ and LR- values were 1.754 and 0.853. The measure of agreement (Kappa coefficient) between stereophotographs and OCT GPA and SAP GPA was low with values of 0.0372 and 0.0533, respectively.
Conclusions: :
OCT GPA, SAP GPA and expert-assessment of optic disc stereophotographs showed poor agreement on the detection of progression in glaucoma.
Keywords: clinical (human) or epidemiologic studies: systems/equipment/techniques • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)