BACKGROUND

To compare the performance of high-resolution microperimetry testing against standard automated perimetry (SAP) for detecting glaucomatous damage seen on optical coherence tomography (OCT) scans.

METHODS

250 eyes from 200 individuals underwent high-resolution microperimetry testing of a hemifield using a stimulus pattern optimised to sample typical arcuate patterns of visual field loss in glaucoma. SAP was performed using a 24-2 stimulus pattern. The presence of glaucomatous damage in each hemifield was subsequently independently assessed by two graders based on a circumpapillary OCT circle scan, 6×6 mm optic disc-centred OCT volume scan, and 15×15 mm widefield OCT volume scan.

RESULTS

The hemifield-based mean total deviation (MTD) and pointwise sensitivity standard deviation (PSD) for high-resolution microperimetry and SAP were both independently associated with the presence of glaucomatous damage in multivariable logistic regression analyses (p ≤ 0.046). Prediction models developed using these two parameters showed significantly higher performance with high-resolution microperimetry compared to SAP (partial area under the receiver operating characteristic curve [pAUC] = 0.86 and 0.75 respectively; p = 0.007). The performance of microperimetry decreased with decreasing resolution when the MTD and PSD values were derived from every second, third, and sixth test locations (pAUC = 0.85 [p = 0.16], 0.85 [p = 0.044] and 0.84 [p < 0.001] respectively), but they were all significantly better than SAP (p ≤ 0.028 for all).

CONCLUSIONS

High-resolution microperimetry enabled improved performance for detecting glaucomatous damage compared to SAP, highlighting the potential value of fundus-tracking and higher-resolution sampling for the detection of glaucomatous visual field loss.