Abstract
Abstract: :
Purpose: Tilted optic discs often have indistinct borders which can be difficult to identify clinically and on the image generated by confocal scanning laser ophthalmoscopy (CSLO). We assessed whether the use of colour stereo disc photos improves the reliability of CSLO measurements in patients with tilted optic discs. Methods: 25 eyes with tilted optic discs and 25 eyes with normal, non–glaucomatous optic discs underwent stereoscopic colour disc photography and CSLO using the Heidelberg Retinal Tomogram II (HRT). The resulting scans were then randomly selected and the optic disc contour drawn on the HRT screen image as is usual in clinical practice. This process was repeated until each scan had been assessed 3 times each on separate occasions by three observers. All scans were then re–analysed using corresponding colour disc photos, viewed stereoscopically, to aid identification of the optic disc edge. This was also repeated 3 times by each of the three observers. Results: There were no significant differences between measurements made by the same observer on separate occasions (intra–observer variability) in either the normal or tilted disc group. Similarly there were no significant differences in measurements made by different observers (inter–observer variability) for the tilted and normal disc group. However analysis showed that by not using stereo discs as a guide, the disc rim area was significantly under–estimated by 17% (+/– 15sd) (p=0.006) in patients with normal discs and 19% (+/– 14% sd) (p=0.001) in tilted discs. Conclusion: The optic disc margin in normal and tilted discs is reliably identified by different observes and on separate occasions using the HRT generated screen image. However this may be an underestimate of the true disc size. The use of colour stereo photographs at the baseline HRT examination may help to more accurately identify optic disc neural tissue so that future changes can be monitored more efficiently.
Keywords: imaging/image analysis: clinical • optic disc