The accuracy of the inferior>superior>nasal>temporal neuroretinal rim area rule for diagnosing glaucomatous optic disc damage

Abstract

Purpose: To determine the accuracy with which the optic disc can be diagnosed as normal or glaucomatous according to the ISNT rule, whereby, in the normal eye, the neuroretinal rim area follows the order inferior (I) > superior (S) > nasal (N) > temporal (T).

Design: Prospective, cross-sectional, observational, case series.

Participants: Fifty-one normal individuals and 78 individuals with open-angle glaucoma exhibiting field loss (median mean deviation, -4.37 dB; interquartile range [IQR], -2.10 to -7.96 dB; median pattern standard deviation, 5.65 dB; IQR, 2.94 to 8.56 dB). The reference diagnosis was made by 2 experts on the basis of the appearance of the optic disc and of the corresponding visual field.

Methods: Stereoscopic optic disc photographs, acquired for each individual, were digitized at high resolution and analyzed using a digital, quad-buffered, stereoscopic viewing system in which a Z screen was used to dissociate the images to the 2 eyes of the observer. Three expert observers, trained to fellowship standard in glaucoma, independently undertook planimetry of the neuroretinal rim and of the disc margin from 1 eye of each individual, using a cursor moving in stereoscopic space to minimize parallax errors. Software automatically calculated the neuroretinal rim area in 10°, 30°, 40°, and 90° segments. For the ISNT rule to be obeyed, the 3 Boolean comparisons of the neuroretinal rim area, I>S, S>N, and N>T, had to be true. If any of the comparisons returned false, the rule was considered not to have been obeyed. Values were compared at a precision of 0.0001 mm(2).

Main outcome measures: The outcome of the ISNT rule in terms of the 3 Boolean comparisons of the neuroretinal rim area was specified in terms of the sensitivity, specificity, and hence, the positive and negative likelihood ratios.

Results: Based on the ISNT rule being obeyed for 10° segments, the positive likelihood ratio among the 3 observers was 1.11 (95% confidence interval [CI], 0.99-1.25), 1.07 (95% CI, 0.94-1.21), and 1.06 (95% CI, 0.96-1.18), respectively. It was similar for the other segment sizes. Variants of the rule were not appreciably better.

Conclusions: The ISNT rule has limited utility in the diagnosis of open-angle glaucoma.