Statistics from Altmetric.com
I read with interest the article by Nucci et al1 on the correlation between the diffusion tensor imaging of the optic nerve with morphological indices obtained by other instruments used in diagnosis of glaucoma. The authors have mentioned one of the criteria used to diagnose glaucoma as, ‘asymmetrical vertical cup to disc ratio exceeding 0.5’. This is unusual, as most of the textbooks use asymmetrical cupping of 0.2 or more as diagnostic criteria for glaucoma. Moreover, vertical cup disc ratio of 0.5 or more is another criterion for diagnosis of glaucoma.
The authors have mentioned that the 24-2 SITA standard programme was used for visual field testing. The first criterion for diagnosis of glaucoma mentioned is ‘corrected pattern SD less than 5% and/or glaucoma hemifield test results exceeding normal limits.’ But it is a well known fact that we cannot get Corrected Pattern SD with SITA standard programme. Moreover, the presence of only one of the two criteria as suggested by ‘and/or’ is not sufficient for diagnosing glaucoma. The third Anderson's criterion of ‘a cluster of three or more points in a location typical for glaucoma (not necessarily non-edge as it is 24-2 program), all of which are depressed at a p<5% level and one of which is depressed at a p<1% level’ is not mentioned at all.
The second criterion mentioned for diagnosis of glaucoma in visual field is ‘absence of other disease that would explain the visual field abnormality’. But this should be in the exclusion criteria of patients rather than the visual field criteria; or it should be presence of visual field abnormalities that did not meet Anderson's criteria for glaucoma.
The third criterion mentioned is fixation loss rate of 25% or less. But the standard cut-off limit for good reliability is 20% or less in case of fixation loss. Fourthly, the authors have mentioned the false positive and false negative rates below 15% as criteria. Here again, the cut-off for a reliable field is 33% or less. I would like to know whether there are specific reasons for the change in the criteria of these reliability indices by the authors.
In the statistical analysis section, the authors have mentioned that ‘a p<0.004 was considered statistically significant.’ Most of the results mentioned have p<0.01 or p=0.01. If we follow the above mentioned criterion, then all the results will become statistically insignificant, and the conclusion of the manuscript will change. Usually p<0.05 is considered statistically significant.
In the discussion, the authors have written that ‘in particular diffusion tensor imaging studies by our group have revealed altered mean diffusivity and fractional anisotropy values suggestive of axonal disruption at the level of optic radiation.10’ But the study referred to in reference no. 10 is that of Zhang et al,2 which has been conducted at the Department of Radiology, Washington University, St Louis, Missouri, USA where none of the authors are involved.
I would like to thank the authors for this excellent study which proves that diffusion tensor imaging parameters of the optic nerve show good correlation with morphological features of the optic nerve head and retinal nerve fibre layer documented with scanning laser polarimetry, Heidelberg retinal tomograph III and optical coherence tomography. However, I would like to know whether it can discriminate retinal ganglion cell axonal injury due to glaucomatous and non-glaucomatous aetiologies.
Competing interests None.
Provenance and peer review Not commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.