AIM:To evaluate the ability of Cirrus spectral domain optical coherence tomography(SD-OCT)-guided progression analysis(GPA)software to detect glaucomatous progression in primary open angle glaucoma(POAG)patients.

METHODS:Longitudinal study. The study examined 45 eyes of 36 patients with POAG over a 2y period. All eyes underwent at least four serial retinal nerve fiber layer(RNFL)thickness measurements performed by Cirrus OCT, with the first and last measurement separated by at least 2y. Visual field(VF)testing was performed by using the Swedish interactive threshold algorithm(SITA)Standard 30-2 program of the Humphrey field analyzer within the same week as the optic disc/RNFL photography. Serial RNFL thickness were assessed by the GPA software program. Glaucomatous eyes were classified as either early or advanced stage according to VF severity. At the same time each eye was labeled with status of RNFL(diffuse RNFL defect; localized RNFL defect; no RNFL defect; unidentifiable RNFL status)based on baseline RNFL photographs. Reference standard of glaucoma progression was defined by expert assessment of optic disc/RNFL photographs or VF data. Sensitivity and specificity of OCT GPA, as well as agreement between OCT GPA findings and each reference standard data were estimated.

RESULTS: Eighteen eyes showed progression by optic disc/RNFL photographs or VF data, while 15 eyes by OCT GPA. When expert assessment of optic disc/RNFL photographs and VF data was used as the reference standard, the sensitivity and specificity of OCT GPA employed to detect glaucoma progression were 38.9% and 70.4%. Agreement between OCT GPA and either optic disc/RNFL photographic evaluation or VF analysis was poor(к=0.211, -0.036 respectively). When expert assessment of optic disc/RNFL photographs was used as the reference standard, 6 eyes were detected progression only by photographs, 2 eyes showed a new disc hemorrhage while 4 eyes with optic disc rim thinning. Among 9 eyes processed only by OCT GPA, 8 eyes were in early stage of POAG, of which 5 eyes had a diffuse RNFL defect and 2 eyes with no RNFL defect at baseline. VF analysis used as the reference standard, 7 eyes were detected progression only by VF testing, of which 5 eyes in advanced stage of POAG. Twelve eyes were processed only by OCT GPA, of which 10 in early stage of POAG.

CONCLUSION:The Cirrus OCT GPA is more sensitive in eyes with a diffuse RNFL defect and may be useful for progression detection in earlier stage of glaucoma to complement other reference standard strategies.