advertisement
Editors Selection IGR 21-3
Visual Function and Structure: Case finding with HRT
Comment by David Crabb on:
23861 Sector-based analysis with the Heidelberg Retinal Tomograph 3 across disc sizes and glaucoma stages: a multicenter study, Oddone F; Centofanti M; Iester M et al., Ophthalmology, 2009; 116: 1106-1111
Oddone et al. (671) examined how well patients with reliable, repeatable visual field defects could be separated from age-matched healthy controls using the Heidelberg Retina Tomograph (HRT). Specifically, they investigated the HRT Moorfields Regression Analysis (MRA) and Glaucoma Probability Score (GPS) applied to global and sector-based optic nerve head (ONH) outcome measures across a range of disc sizes. Visual field defects were classified into three stages of severity. Both visual field and HRT data were excluded on the basis of established quality metrics. The study was multi-center, conducted on a large number of eyes, and is well described. The authors conclude that small ONHs are classified more accurately by looking at the inferonasal sector, whereas large ONHs are better classified by looking at the inferotemporal sector. Results and conclusions from studies examining clinical diagnostic tools only really have merit if they follow the Standards for the Reporting of Diagnostic studies guidelines (STARD www.stard-statement.org). This study adheres to these but there are some minor exceptions (e.g., not reporting the quality of the instrument operators or describing the entire distribution of disease severity). However, the omission of statistical estimates of uncertainty associated with sensitivity/specificity calculations could question the inference about the location of optimum sectoral measures for classifying optic discs. Nevertheless, this study usefully confirms the relatively low positive predictive value of using this technology in isolation; a very important message when case finding and detecting disease. GPS is fully automated and makes it an appealing diagnostic technique, especially in the busy clinical or screening environment, but this study provides more good evidence that the algorithm's performance, like the MRA, is less than satisfactory when ONHs are very small or very large.
Comments
The comment section on the IGR website is restricted to WGA#One members only. Please log-in through your WGA#One account to continue.Log-in through WGA#One
