Purchase this article with an account.
Samantha D'Amico, Brian Kim, Christopher J Brady; Meta-Analysis of Handheld Fundus Camera Validation Studies. Invest. Ophthalmol. Vis. Sci. 2019;60(11):PB044.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Fundus cameras are vital tools in addressing preventable blindness, but traditional desktop cameras are often costly, bulky, and dependent on the skills of the user. Handheld fundus cameras have reduced these barriers, but data on their efficacy are limited. We reviewed existing literature to identify validation studies of handheld fundus cameras and conducted a meta-analysis of the sensitivity and specificity of the camera.
We conducted a Pubmed search in October 2018 for “handheld fundus camera” and “portable fundus camera” and a Pubmed and Google Scholar search for “Horus scope”, “Visuscout”, “Pictor plus”, “Retinavue”, “Versacam”, “Kowa Genesis”, “Epicam”, “Smartscope”, and “Microclear Luna”. Papers not in English were excluded. A title and abstract review was conducted and non-validation studies were excluded. Full-text articles were assessed and excluded if the camera was not commercially available, the study was not a camera validation study, there was no comparison to the gold standard, or data was not quantified. Studies included in the meta-analysis validated a handheld fundus camera against a gold standard method.If not provided, sensitivity and specificity were calculated using true and false positive and negative values for each grader. True positives/negatives and false positives/negatives were averaged to generate mean values for each study if there was more than one grader. Mean sensitivity and specificity were calculated for each study. Any reported data on image quality/gradability was extracted.
Ninety-five studies were reviewed for inclusion in the review with 7 included in the meta-analysis (Figure 1). The Pictor/Pictor Plus camera was used in all but 1 study. Mean sensitivity ranged from 79.20-100% and mean specificity from 72.09-93.85% (Figure 2). The sensitivity and specificity for all studies calculated using a weighted average were 85.32% (95% CI: 80.15-89.54%) and 84.63% (95% CI: 80.30-88.34%). Analysis of image quality/gradability was not feasible due to the heterogeneity of reporting in included studies.
We found that hand-held cameras are capable of achieving acceptable sensitivity and specificity values as screening tools. While these values may not be acceptable for all clinical programs, they offer a foundation to build upon for future development of handheld fundus cameras.
This abstract was presented at the 2019 ARVO Imaging in the Eye Conference, held in Vancouver, Canada, April 26-27, 2019.
Figure 1: Flow diagram of the review and meta-analysis
Table 1: Summary of included studies
This PDF is available to Subscribers Only