advertisement
Editors Selection IGR 15-2
OCT: Histology of the RNFL and imaging
Comment by George Lambrou on:
22504 Retinal nerve fibre layer imaging compared with histological measurements in a human eye, Blumenthal EZ; Parikh RS; Pe'er J et al., Eye, 2009; 23: 171-175
In recent years, the development of novel and ever-more powerful fundus imaging modalities has changed the way that we assess glaucoma patients. The reliance on the C/D ratio as sole quantitative parameter to evaluate and monitor structural damage is nowadays complemented by a hoist of computer-generated disc and RNFL parameters based on innovative imaging technologies such as Confocal Scanning Laser Ophthalmoscopy, Optical Coherence Tomography (OCT) and Scanning Laser Polarimetry (SLP). When it comes to RNFL, these parameters are usually expressed as 'RNFL thickness measurements', although this is in reality a mental shortcut: what is actually being measured is an optical variable (such as laser light reflectance, coherence or degrees of retardation) which is then translated into RNFL thickness units ('microns') via a conversion factor obtained from initial histologyimaging comparisons, usually based on a limited number of human or primate eyes. In this study, Blumenthal et al. (47) have conducted a post-validation imaging/histology comparison using a novel method of histological preparation and sectioning (the 'umbrella' technique) which permits to reconstruct hi-resolution maps of the peri-papillary RNFL thickness. They have compared thickness profiles measured histologically along disc-centered rings with SLP-estimated thickness profiles along rings of the same diameters (3.0, 3.5, 4.0 and 4.5 mm) and an OCT-estimated profile along a 3.4-mm ring. The overall form of the histology and GDx profiles were roughly similar with decreasing thickness in more distant rings and an expected double hump in the superior and ‐ somewhat less ‐ the inferior quadrants, but the similarities stopped there. RNFL thicknesses displayed on imaging system reports are approximate estimates and their value resides essentially in comparisons to individual or population-based values obtained with the same instrument
The fine details of the thickness profiles were markedly different and so where the actual micron thickness values. Does this mean that the RNFL 'measurements' we read daily on our sophisticated machines are unreliable? This one study, based on data from a single enucleated eye provides no evidence whatsoever for such a blunt statement. More studies on many more eyes imaged shortly before enucleation would be required to conduct a robust quantitative histology-imaging comparison. Moreover ‐ as the authors themselves point out ‐ such a comparison would be limited by artifacts inherent to any histological technique (like post-mortem tissue swelling, shrinkage or distortion) and any imaging modality (like scan location variation, magnification compensation or errors in conversion factors), so that a micron-to-micron RNFL thickness correlation between techniques is probably illusory. This is not a new finding: comparing imaging modalities to one another can reveal significant thickness estimate discrepancies, with differences that can reach a factor of two. Interestingly, in this study the histologically-measured thickness values were located midway between the GDx and OCT estimates. As a take-home message from this study, we should always keep in mind that RNFL thicknesses displayed on imaging system reports are approximate estimates and their value resides essentially in comparisons to individual or population-based values obtained with the same instrument, and namely with the normal confidence intervals usually displayed on the same reports.
Comments
The comment section on the IGR website is restricted to WGA#One members only. Please log-in through your WGA#One account to continue.Log-in through WGA#One