Retinal nerve fiber layer assessment by scanning laser polarimetry and standardized photography

To determine whether, in a clinical setting, scanning laser polarimetry and retinal nerve fiber layer photography provide equivalent information on the retinal nerve fiber layer.

We prospectively studied 60 patients with glaucoma or ocular hypertension and 24 healthy subjects. With scanning laser polarimetry, an estimate of the cross section of the retinal nerve fiber layer was obtained. By using a photographic reference set, we quantified photographs of the retinal nerve fiber layer. Visual fields were used to relate the results of both methods to functional damage.

The scanning laser polarimetry measurements yielded reproducible cross-section values (coefficient of variation, 6.6%). Comparison of cross-section values and photograph scores gave Pearson correlation coefficients smaller than r = .4 (P < .01), improving to a maximum of r = .53 after compensation for offset. When cross-section values were compared to the mean deviation of the visual field, the Spearman correlation coefficients varied from an r of -.34 to -.53 (P < .01). Correction for offset resulted in an r of -.54 to -.65. When photograph scores and mean deviation of the visual field were compared, the Spearman correlation coefficients varied from an r of -.65 to -.71 (P < .01).

Because r was maximal at .53, the information on the retinal nerve fiber layer obtained with scanning laser polarimetry and photography seems not equivalent. This result could not have been because of lack of reproducibility. Although the results suggested possible offset in scanning laser polarimetry, other methodologic differences must be considered to explain the differences between the two techniques.