Abstract
Purpose: :
To compare the RNFL thickness obtained with frequency domain (fd) and time domain (td) optical coherence tomography (OCT).
Methods: :
43 eyes of 27 patients with glaucoma (G), 34 eyes of 25 glaucoma suspects (S), and 25 eyes of 16 controls (C) were tested with tdOCT (Fast RNFL scan with signal strength >= 5, OCT3, Zeiss Meditech) and fdOCT (Disc 3D-Scan with Qfactor >= 40, 3DOCT-1000, Topcon). Glaucoma patients had abnormal discs and 24-2 Humphrey visual fields (abnormal GHT or PSD); the mean MD was -5.9±6.8 dB. Suspects had abnormal discs but normal visual fields. For the tdOCT, the RNFL thickness profiles were exported. For the fdOCT, the RNFL thickness was determined around a circle 3.4mm in dia., centered on the optic disc, which is comparable to that used in the tdOCT scan. For both fdOCT and tdOCT, the overall average RNFL thickness was computed, as well as the average thickness for 6 optic disc sectors.[1] Pearson correlations and Bland-Altman (B-A) plots were generated.
Results: :
The agreement between the fdOCT and tdOCT was good. The overall average thicknesses were similar for the fdOCT [C: 96.6; S: 90.2; G: 74.8] and tdOCT [C: 98.6; S: 92.5; G: 71.9], and the correlation between these measures high (r=0.91). For the 6 sectors, the correlations range from 0.76 to 0.91. The B-A plots showed no evidence of a systematic error between tdOCT and fdOCT measures and little or no evidence of a proportional error. That is, the average offset and the slopes of the B-A plots were close to zero, although there was a trend for the patients with more extensive RNFL loss to show slightly larger fdOCT values, probably due to an underestimate of blood vessel thickness by the tdOCT. [2,3] An analysis of the most discrepant points (greater than 1.75 SD on B-A plots) suggested that the algorithms determining RNFL thickness play a role. For example, the algorithm with more spatial averaging (tdOCT), can yield slightly larger RNFL thicknesses in controls, but smaller RNFL thicknesses in patients with more extensive RNFL loss.
Conclusions: :
RNFL thickness obtained with fdOCT can be compared to data previously obtained with tdOCT. The RNFL thickness values were similar and individual values showed a good correlation. However, differences in algorithms can produce differences in RNFL thickness measurements in particular individuals. 1. Garway-Heath et al, Ophthal (2000); 2. Hood & Kardon, PRER (2007); 3. Hood et al, J. Glau. In press.
Keywords: nerve fiber layer • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)