Comparison of Humphrey versus Compass Perimetry for Hemianopsia

Up to 57% of stroke patients experience visual defects. Visual restitution therapy post-stroke remains controversial, with some attributing improvements to adaptive eye scanning movements rather than true field augmentation. We compare Compass fundus-tracking perimetry (CMP), which compensates for eye movements, to the Humphrey Field Analyzer (HFA) for homonymous hemianopsia. Nine patients (mean age: 47) with homonymous hemianopsia on HFA and corresponding neuroimaging defect were prospectively tested on the same day using the HFA (24–2 SITA Fast) and CMP (24–2 ZEST fast). Reliability indices, mean deviation (MD), and visual field index (VFI for HFA; FDPI for Compass) were compared via median differences and the Wilcoxon signed-rank test. The CMP had a significantly lower MD (median difference: −0.33 dB, <i>p</i> = .02), significantly greater false negative rate (median difference: 27%, <i>p</i> = .04), and a significantly longer test duration (median difference: 87 seconds, <i>p</i> = .01) than HFA. However, no between-analyzer difference occurred for visual field index (median difference: 6.8%, <i>p</i> = .65), false positive rate (median difference: −2.8%, <i>p</i> = .18), CMP blind spot index and HFA fixation losses (median difference: 0, <i>p</i> = .79). Bland-Altman plots showed acceptable agreement, with a + 2.42 dB bias in MD favoring CMP. CMP offers real-time compensation for fixation losses but did not show a clinically significant advantage over HFA for homonymous hemianopsia.