Abstract
Purpose: :
Doppler optical coherence tomography (OCT) has been advanced as a new technology for measuring retinal blood flow, but image quality must be optimized in order to achieve valid or reliable blood flow results. The purpose of this study was to correlate image quality parameters at the time of scan acquisition with the likelihood of obtaining a valid Doppler OCT retinal blood flow result.
Methods: :
Doppler OCT scans (Optovue RTvue FD-OCT) of 285 eyes from 274 subjects were obtained as part of the Chinese American Eye Study. Ten automated image quality parameters were generated by the machine including Acceptable scan number, Average signal strength index (Avg SSI), Average root mean square of inner limiting membrane (Avg ILM RMS), Average acceptable frames, Average continuous acceptable frames, Average maximum eye movement (Avg max EM), Average maximum frame depth shift (Avg max frame depth shift), Pooled co-efficient of variance for inner limiting membrane angle (Pooled CV of ILM), Average inner limiting membrane angle (Avg ILM angle) and Root mean square standard deviation of inner limiting membrane angle (RMS sd ILM angle). Retinal blood flow was then calculated using Doppler OCT of Retinal Circulation (DOCTORC) grading software as described in prior publications. The DOCTORC software also automatically determines whether the blood flow result is valid and reliable.The influence of the 10 image quality parameters on the validity of the blood flow measurement was analyzed using univarate and then multivariate logistic regression.
Results: :
Of the ten quality parameters, only 6 were found to significantly correlate with the validity of the Doppler blood flow measurements. Acceptable scan number (OR, 1.54; 95%CI, 1.23-1.93) was the most predictive factor followed by Average acceptable frames (OR, 1.49; 95%CI, 1.03-2.17), Average continuous acceptable frames (OR, 1.26; 95%CI, 1.002-1.59), Average ILM RMS (OR, 0.96; 95% CI, 0.93-0.99), RMS sd ILM angle (OR, 0.84; 95%CI, 0.71-0.99), and Pooled CV of ILM angle (OR, 0.19; 95%CI, 0.01-0.54). The quality parameters which did not appear to influence validity were Average ILM angle, Average max EM, Average max frame depth shift, and Average SSI.
Conclusions: :
Several image quality parameters appear to correlate with Doppler OCT scans yielding valid retinal blood flow results. These observations may be useful in developing optimized quality metrics which can assist operators in successfully obtaining adequate Doppler OCT scans.
Keywords: imaging/image analysis: non-clinical