**5. Sensitivity and specificity**

Badala et al [4] compared the efficacy of stereoptic disc assessment and that of all three imaging modalities (OCT, GDx, HRT 3) in diagnosing glaucoma. The sensitivity at 95% specificity of the best performing parameter of each modality is: for the OCT (average RNFL thickness) 89%, for the GDx VCC (nerve fiber indicator) 78% and for the HRT 3 [Frederick S. Mikelberg (FSM) discriminant function] 70%. Optic disc stereophtographs are as accurate in detecting glaucoma as the other imaging modalities.

Retinal nerve fiber analysis with all the above modalities exhibit a characteristic double hump because the RNFL in thicker in the superotemporal and inferotemporal sectors compared to the nasal and temporal ones.

All of the above imaging modalities have been employed in the diagnosis and follow up of patients with various stages of glaucomatous optic neuropathy. Studies have shown that there is a discrepancy between the measurements of the optic disc parameters taken with OCT and HRT in glaucomatous eyes [13]. HRT II had higher values for disc and rim area while RTVue-100 OCT had higher values for cup area, cup-to-disc area ratio, and vertical and horizontal cup-to-disc ratio. Leite et al [14] compared three FD-OCT machines and reported that their performance in detecting glaucoma is similar. FD-OCT out-performed SD-OCT in detecting progression of the glaucomatous process [15] but they were comparable in detecting glaucomatous damage [16]. Lee et al [17] found that the best performing parameter for glaucoma detection of the GDx is the nerve fiber index and that for Cirrus OCT the inferior RNFL thickness. GDx was also more accurate in detecting glaucoma than the Cirrus OCT. Two recent studies [18,19] showed that the diagnostic accuracy for glaucoma of the HRT II is dependent on the disc size which is not the case for OCT and GDx.

The severity of the glaucomatous process also affects the accuracy of glaucoma diagnosis of the various imaging technologies. The more advanced the disease the more accurate the diagno‐ sis of glaucomatous optic neuropathy [20,21]. OCT and SLP performed better than CSLO in discriminating between early glaucomatous eyes with or without visual field defects [22]. In eyes with early glaucoma the most accurate parameter is the inferior RNFL thickness which performs better than the most accurate parameter of the CSLO (vertical cup-to-disc ratio). In glaucoma suspect eyes the most accurate parameter for the OCT is the average RNFL thick‐ ness, for the SLP the nerve fiber indicator and for the CSLO the vertical cup-to-disc ratio. The first two parameters performed better than the vertical cup-to-disc ratio. Leung et al [23] confirmed that SD-OCT performed better than HRT in recognizing patients with glaucoma. RNFL thickness changes performed better than optic nerve head parameters as evaluated with CSLO. The nerve fiber index of the SLP was more accurate in diagnosing glaucoma than the rim volume parameter of the CSLO [24]. SLP was also superior in detecting glaucoma progres‐ sion by analyzingRNFL thickness compared to CSLO analysis ofthe neuroretinalrim area [25].
