Abstract
Purpose :
Microperimetry (MP) measures retinal sensitivity at several locations throughout the macula and could provide a functional outcome measure to evaluate treatments for retinitis pigmentosa (RP). In this prospective observational study, we assessed test-retest reliability in patients with RP using two different MP devices, the Nidek Micro-Perimeter 1 (MP1) and the CenterVue Macular Integrity Assessment (MAIA).
Methods :
Data were collected as part of an ongoing clinical trial in which patients with RP having some central field loss underwent MP with MP1 and MAIA at two visits prior to initiation of treatment. Both tests were performed using a 10-2 pattern comprising 68 points across the macula. Retinal sensitivity at each point was analyzed between visits and between devices. Test-retest reliability was calculated using Intraclass Correlation Coefficient (ICC) and the Coefficient of Repeatability (COR). A mixed effects model was used to account for correlation between eyes of the same patient. To ensure that the reliability measures are not misrepresented, analysis was done both with and without the inclusion of deep scotomas (points which had a sensitivity of 0 on both visits).
Results :
A total of 31 eyes from 16 patients were analyzed (one eye excluded due to CRVO). Mean (± SEM) age was 49 ± 4 years and 50% were male. Averaging across both visits, mean sensitivity was 2.4 ± 0.1 dB with MP1 and 8.2 ± 0.2 dB with MAIA. MP1 had a significant floor effect as there were 1435 (68%) deep scotomas. With MAIA, only 664 (31%) of these points were deep scotomas while the rest had a mean sensitivity of 7.0 ± 0.2 dB. Test-retest reliability was worse for both devices after the exclusion of deep scotomas and this is reported. MP1 had an ICC (95% CI) of 0.69 (0.63-0.75) and a COR (95% CI) of 9.2 (7.8-10.7) dB. MAIA had an ICC (95% CI) of 0.84 (0.80-0.87) and a COR (95% CI) of 12.2 (9.7-14.8) dB.
Conclusions :
The greater number of points with measurable sensitivity using MAIA provides greater power for outcome assessment in patients with RP. Overall test-retest repeatability was good, but some points in all patients showed high variability. This suggests that change in mean sensitivity is not a useful outcome for RP, but rather changes in sensitivity should be assessed at each individual point, taking into account the test-retest reliability at each point prior to treatment onset.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.