Abstract
Purpose :
The 12-line spectral domain optical coherence tomography (SD-OCT) radial scan pattern has been shown to be equivalent to the 25-line raster scan pattern for the detection of fluid in neovascular age-related macular degeneration (nAMD), but no prior studies have compared either SD-OCT scan pattern to en-face OCT. The current study compares effectiveness of the auto-segmented en-face scan pattern to the 12-line radial SD-OCT scan pattern in detecting intraretinal, subretinal, and sub-retinal pigment epithelium (sub-RPE) fluid in nAMD.
Methods :
Retrospective, observational series of patients with nAMD imaged between November 2014 and October 2015. A total of 49 eyes that underwent both en-face OCT with superficial retinal, deep retinal, and choriocapillaris auto-segmentation (Avanti; Optovue, Inc., Fremont, CA) and 12-line radial spectral domain OCT (SD-OCT) image acquisition scan patterns (Spectralis; Heidelberg Engineering, Heidelberg, Germany) were evaluated by two independent graders. The 12-line radial SD-OCT scan was considered the gold standard for fluid detection. Interobserver differences were evaluated by a third grader.
Results :
A total of 72 en-face and 12-line radial SD-OCT scans were interpreted (mean: 1.47 scans/eye). 41 scans (56.9%) had confirmed fluid on SD-OCT while 31 scans (43.1%) did not. The en-face scan pattern detected fluid in all but 10 scans (false negative rate = 24.3%; 95% CI 12.4-40.3%), resulting in a sensitivity of 75.6%. Conversely, the en-face scan pattern falsely detected fluid that was not actually present in all but 12 scans (false positive rate = 38.8%; 95% CI 21.8-57.8%), resulting in a specificity of 61.2%. The difference in fluid detection between scan patterns was statistically significant (P<0.05, McNemar test). Corroborating the en-face scan with a corresponding Avanti 3 x 3 mm, 304-line raster scan improved sensitivity and specificity of fluid detection to 87.8% and 90.3%, respectively.
Conclusions :
The en-face OCT scan protocol alone with auto-segmentation is not as sensitive as the SD-OCT 12-line radial scan in detecting the presence of intraretinal, subretinal fluid, and sub-RPE fluid in nAMD. En-face scan findings should be corroborated with additional OCT scan protocols to guide clinical decision making.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.