Purchase this article with an account.
K. A. Muldrew, W. P. Patton, T. Peto, M. R. Stevenson, B. Reeves, U. Chakravarthy, P. Lenfestey, S. P. Harding, VPDT Cohort Study Team; One Year Results of Quality Assurance in the UK VPDT Cohort Study Reading Centre Network. Invest. Ophthalmol. Vis. Sci. 2007;48(13):5105.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To report on quality assurance (QA) of baseline angiograms in the newly established UK Reading Centre Network (Belfast, Liverpool, London) after one year. This network is responsible for monitoring and evaluating angiographic outcomes in all patients undergoing photodynamic therapy within the UK VPDT Cohort Study provided by the national health service of the United Kingdom.
The Central Angiographic Resource Facility (CARF) receives angiograms from a network of 45 nominated treating centres and distributes them to one of the three Reading Centres (RC). The study involves the grading of around 5000 angiograms per annum. The QA process is integrated into the grading routine and involves the automatic selection of 1 in every 8 angiograms for re-grading in a masked fashion. Lesions are subtyped into 4 categories: predominantly classic without occult, predominantly classic with occult, minimally classic and occult no classic.
Of the 4980 completed gradings to date, 538 were entered into the QA grading route. Of these 238 were baseline visits with complete data sets.Discordance for the presence of active CNV occurred in 27 cases (unweighted Kappa = 0.43). With agreement on the presence of active CNV the weighted linear Kappa was moderate at 0.44 (95% CI = 0.32-0.57). 68% of comparisons had complete agreement, 6% disagreed by one subtype category, 19% by two categories and the remaining 17% by three categories. When the presence or absence of active CNV was discordant (27 cases), 17 of these (63%) had fibrosis or atrophy present. Image quality was found to be an important factor for discordance in determining CNV activity.
The present study has shown consistent concordance within the UK RC network. The results confirm that a multi-centre approach to angiographic grading is feasible and thus is suitable for grading a high volume of FAs. Discordance can be attributed mainly to the complex nature of the AMD lesions particularly when assessing fibrosis or atrophy as markers of CNV activity. The study also highlights the importance of image quality and its role in grading of angiograms.
This PDF is available to Subscribers Only