Use of the structure-function relationship in detecting glaucoma progression in early glaucoma

We performed a retrospective review of patients with glaucoma who underwent OCT measurements of RNFL thickness and VF testing at Kagawa University Hospital, Kagawa, Japan, between October 2008 and January 2014. Each of the eyes underwent at least four serial RNFL measurements, with the first and last measurements separated by at least 3 years. All eligible subjects received a detailed explanation of the study and signed an informed consent form in accordance with the principles embodied in the Declaration of Helsinki. This study was approved by the Institutional Review Board at Kagawa University Hospital. At the baseline examination, all subjects underwent a complete ophthalmic evaluation that included visual acuity testing with refraction, intraocular pressure, and dilated fundus examination with stereoscopic biomicroscopy of the optic nerve head using slit-lamp and indirect ophthalmoscopy. To be included in the study, all subjects had to have a best-corrected visual acuity of 20/40 or better, a spherical error within a range between +4.0 and -6.0 diopters, and a cylinder within ± 2.0 diopters. Exclusion criteria included a history of any kind of retinal pathology, retinal laser procedure, retinal surgery, or neurologic disease. In addition, we also excluded patients with advanced hemifield VF defect who had an initial TD in the superior or inferior hemifield under -6 dB. However, if the superior and the inferior hemifield were each above -6 dB, we included both hemifields. In addition, since we assessed each hemifield in these patients, the global TD was not included in the exclusion criteria. Glaucomatous eyes were defined as eyes exhibiting structural glaucomatous changes (vertical cup-disc asymmetry between fellow eyes of ≥ 0.2, a cup-disc ratio of ≥ 0.6, and a neuroretinal rim narrowing, notches, localized pallor, or RNFL defects with glaucomatous VF loss in the corresponding hemifield). A glaucomatous VF was defined as a glaucoma hemifield test outside of the normal limits on at least two consecutive baseline tests and the presence of at least three contiguous test points within the same hemifield on the pattern deviation plot at P < 1%, with at least one at P < 0.5% (excluding those points on the edge of the field or those directly above and below the blind spot). VF testing and RNFL imaging were performed during the same visit.

Cirrus HD-OCT uses spectral domain technology (Carl Zeiss Meditec, Dublin, CA). This technique utilizes an optic disc cube that is generated from a 3-dimensional data set composed of 200 A-scans from each of 200 B-scans that cover a 6 mm² area centered on the optic disc. After creating an RNFL thickness map from the cube, the software automatically determines the center of the disc and then extracts a circumpapillary circle (1.73 mm radius) from the data set. All of the images obtained in the current study had signal strengths of at least 6. The RNFL thickness deviation and the RNFL thickness change maps were automatically generated by the OCT instrument and exported to a computer that analyzed the progression pattern of the RNFL defects. The RNFL defects were visualized in the RNFL thickness deviation map, which was composed of 50 × 50 pixels. A pixel was coded in yellow if the RNFL measurement was below the lower 95% and coded in red if below 99% of the percentile ranges for that particular pixel.

The RNFL thickness change map was a component of the Guided Progression Analysis (GPA) software (Carl Zeiss Meditec), which provided both event- and trend-based analysis of the RNFL progression based on the serial RNFL thickness maps. This software automatically aligned and registered the baseline and follow-up OCT images so that the same pixel locations could be measured for change. In order to be able to generate a GPA report, at least four patient visits were necessary. The GPA program then overlaid and compared the serial RNFL thickness against the duration of the follow-up.

Standard visual field testing was performed using static automated white-on-white threshold perimetry (Humphrey Field Analyzer II (HFA); Carl Zeiss Meditec, Dublin, CA), using the 30–2 Swedish Interactive Threshold Algorithm Standard test. The visual field was defined as reliable only when the fixation losses and the false-positive and false-negative rates were less than 20%. Only reliable test data were used in all of our analyses.

The GPA software compared the patient’s baseline visual field to each of the subsequent visual fields determined during the follow-up examinations. All baseline values were obtained by averaging the data from the first two exams. The progression evaluation was performed relative to the baseline. The evaluation of the progression was carried out by comparing the threshold modifications to a database of stable glaucoma patients who were tested over a very short period of time. These evaluations took into account the fluctuations related to eccentricity and advancing disease. Each of the consecutive tests examined the same locations (≥3) and determined if there was progression during any of the consecutive VF tests. Based on the findings of a previous study [10], “possible progression” was considered to be present when the GPA printouts reported finding progression in 2 locations, while “likely progression” was considered to be present when progression was found in 3 locations.

Using the GPA software, RNFL thickness progression was assessed by event analysis. Event analysis defined progression as a change that exceeded a predefined limit when compared to the baseline value. The baseline values were obtained by averaging the data from the first two exams. After completing a series of RNFL thickness measurements, the GPA software then compared the baseline RNFL thickness value to the final measurement value. A “possible loss” was identified when the differences between the baseline and final measurement values exceeded the test-retest variability. If the “possible loss” criterion was met on two successive visits, the patient was considered to show a “likely loss”. In the current study, if the RNFL Thickness Map Progression indicated that there was a “likely loss” or “possible loss”, we defined the RNFL thickness as having progressed. The VF GPA classification indicated “likely progression” or “possible progression”, this was defined as progression of the glaucoma.

All statistical values are presented as mean ± standard deviation (SD), with P values < 0.05 considered statistically significant. Data were analyzed using a paired t-test. Statistical analyses were performed using SPSS version 19.0 (IBM, New York).