Purchase this article with an account.
Ivania Pereira, Hemma Resch, Clemens Vass; Assessment of Retinal and Optic Nerve Head Biomarkers for Prediction of the Circumpapillary Retinal Nerve Fiber Profile. Invest. Ophthalmol. Vis. Sci. 2011;52(14):196.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To define an accurate normative database for glaucoma detection based on the correlation of TSNIT profiles with putative morphological biomarkers.
A sample of 120 healthy volunteers underwent complete ophthalmologic examination, including HR-OCT scanning (Cirrus ® Carl Zeiss Meditec Inc.) and RNFL Assessment (GDx-ECC ® Carl Zeiss Meditec Inc.). The direction of the major retinal vessels near the optic nerve head was proven to be a good predictor of the TSNIT profile. To confirm this and establish retinal and optic disc parameters as biomarkers for the TSNIT profile, SLO-OCT images centered at the optic disc will be processed for automatic segmentation of vessels and optic disc. All extracted putative biomarkers will be analysed to determine their relevance for the TSNIT profile. To test this model, a preliminary analysis was performed using a subsample of 79 subjects. TSNIT profiles from GDx were parameterized using Discrete Wavelet (DWT) and Discrete Fourier Transforms (DFT). The results were correlated with the major superior and inferior vessel angles (manually segmented).
A good correlation was obtained between retinal vessel angles and DWT coefficients. The 2nd and the 5th detail coefficients present better correlation with the superior major vessel (R =-0.47, R=-0.48, p<0.05), while 8th and the 10th approximation coefficients present a better correlation with the inferior major vessel (R =-0.38, R=-0.40, p<0.05). The 2nd and 4th values of amplitude of the DFT coefficients present good correlations with the superior and inferior vessel angles (0.38<R<0.55, p<0.05), as well as the 2nd and 4th values of power (0.35<R<0.50, p<0.05).
The results confirm a relation between the angles of the retinal blood vessels and TSNIT profile. This validates Wavelet and Discrete Fourier transforms as approaches to determine possible morphological biomarkers for the prediction of the TSNIT profile. This study is considered a step forwards on individualized normative data concerning RNFL. Determining independent biomarkers may compensate for the intersubject variability of RNFL measurements and might render the normative values of retinal nerve fiber measurements more accurate.
This PDF is available to Subscribers Only