Editors Selection IGR 14-4

Clinical Examination Methods: RNFL imaging

Tae-Woo Kim
Eun Ji Lee

Comment by Tae-Woo Kim & Eun Ji Lee on:

51690 Imaging of Localized Retinal Nerve Fiber Layer Defects in Preperimetric Glaucoma Using Spectral-domain Optical Coherence Tomography, Nukada M; Hangai M; Mori S et al., Journal of Glaucoma, 2014; 23: 150-159

Find related abstracts

Early diagnosis is important for effective treatment of glaucoma, which leads to irreversible optic nerve damage. It has been demonstrated that structural changes in the optic nerve head and retinal nerve fiber layer (RNFL) often precede the detectable visual field loss in glaucoma. In this regard, it is of interest to see whether the new imaging devices, such as spectral domain (SD)-optical coherence tomography (OCT), allow detection of structural damage before manifestation of visual field defect.

Nukada et al. compared diagnostic capacity of time-domain (TD)-OCT (Stratus), single-scan SD-OCT (RTVue-100) and speckle noise-reduced (i.e., multiple-scan averaged) SD-OCT (Spectralis) to detect perperimetric glaucoma with RNFL defect identified by fundus photography. They found that speckle noise-reduced SD-OCT had a higher sensitivity (47.5%) than TD-OCT (15.0%) and single scan SD-OCT (42.5%) using the criterion of at least 1 sector < 1% of normal for an age-matched database. They proposed that the difference of the sensitivity was attributed to the higher rate of correct autosegmentation by Spectralis. However, the sensitivity of Spectralis SD-OCT, which was 47.5%, may still be unsatisfactory in the clinical practice. The low sensitivity of OCT based on the sector map analysis may be attributable to the rather wide angular width of sectors (Spectralis, 45 or 95 degrees; Cirrus, 30 degrees; and Stratus, 30 degrees) compared to that of pre-perimetric RNFL defects, which is mostly less than 30 degrees. This matter may be particularly relevant in cases where the defect lies partly in 1 sector and partly in an adjacent sector.

Examining the TSNIT graph, rather than simply looking at the sector maps, may help clinicians not to overlook an early RNFL defect

Being consistent with this explanation, the authors found that TSNIT graphs were more sensitive than sector maps for detecting RNFL defects. On the basis of TSNIT thickness graphs, RTVue-100 and Spectralis had sensitivities of 57.5% and 90.0%, respectively. This finding suggests that examining the TSNIT graph, rather than simply looking at the sector maps, may help clinicians not to overlook an early RNFL defect.

Issue 14-4

Select Issue