Dear Editors

We read with interest the recent study by Krebs et al. demonstrating the limitations of StratusOCT mapping software in the context of age-related macular degeneration (AMD).[1] We wish to applaud the authors for their study - while the limitations of StratusOCT automated analysis have previously been reported,[2] the fact remains that many ophthalmologists may not be well acquainted with these errors. We agree with the authors and wish to underline the importance of these errors in this era of optical coherence tomography (OCT)-guided retreatment regimens for anti-angiogenic agents and increasing adoption of quantitative OCT measurements as secondary outcome parameters in clinical trials.[3]

In their study, the authors present evidence that automated Stratus software provides correct results in only 57.1% of eyes with AMD.[1] These results are consistent with previous reports and in line with clinical experience.[2] As mentioned in the discussion, manual placement of boundaries on OCT scans may represent a solution to this problem. We wish to thank the authors for mentioning our custom OCT grading software termed “OCTOR” as one possible software tool to perform this kind of analysis. The authors comment that "this software is not commercially available or supported by the StratusOCT". However, OCTOR was designed specifically with StratusOCT in mind and readily handles raw exported StratusOCT images. OCTOR is publicly available and can be downloaded from www.diesel.la. This software facilitates manual segmentation of OCT images and allows quantitative analysis of any area of interest in these images e.g. retina, subretinal fluid, subretinal tissue, or pigment epithelium detachment.[4] Furthermore, the most recent version of the Stratus OCT software (version 5.0) enables users to manually correct errant boundaries.

In their discussion, the authors comment that "manually set boundary lines would add subjectivity to the retinal thickness measurements". Recent research demonstrating both the accuracy and reproducibility of manual grading with OCTOR software in the setting of neovascular AMD suggests that manual grading does not necessarily add subjectivity to retinal thickness measurements.[5] Instead, manually corrected boundaries allow for more reliable data than the frequently erroneous automated boundary detection of the StratusOCT software. We recently used the manual grading software OCTOR to quantify the volumes of the neurosensory retina, subretinal fluid, subretinal tissue, and pigment epithelial detachments. Intergrader comparisons showed a high level of agreement and a strong correlation between measurements for all spaces (weighted Kappa= 0.72-0.97; ICC = 0.92-0.99). Although these values were obtained by graders who had undergone a formal certification program in our reading center, in our experience OCT grading can be learned both more quickly, and more easily, than grading of fluorescein angiography. This hypothesis is supported by the level of agreement between measurements in our study which was appreciably better than that which has been reported elsewhere for fluorescein angiography.[6]

Although significant progress is being made, the complex morphology of neovascular AMD presents a challenge to the development of automated OCT interpretation software. As automated segmentation algorithms improve, it will be necessary to assess their efficacy in a quantitative manner against a 'gold standard'. We believe that the accuracy of manual segmentation by trained human graders is such that it could serve as a 'gold standard' against which to compare the results of automated analysis.

In conclusion, we wish to commend the authors for highlighting this area, which is likely to be of critical importance as the next generation of OCT technology is incorporated into the diagnosis and management of neovascular AMD.

LICENCE FOR PUBLICATION:
The Corresponding Author has the right to grant on behalf of all authors, and does grant on behalf of all authors, an exclusive licence on a worldwide basis to the BMJ Publishing Group Ltd to permit this article (if accepted) to be published in BJO and any other BMJPGL products and sublicences such use and exploit all subsidiary rights, as set out in our licence (http://bjo.bmj.com/ifora/licence.pdf).

COMPETING INTERESTS:
Drs. Walsh and Sadda are co-inventors of Doheny intellectual property related to spectral domain optical coherence tomography that has been licensed by Topcon Medical Systems.

FUNDING:
Supported in part by NIH Grant EY03040 and NEI Grant R01 EY014375

References

1. Krebs I, Haas P, Zeiler F, Binder S. Optical coherence tomography: limits of the retinal-mapping program in age-related macular degeneration. Br J Ophthalmol 2008;92:933-5.

2. Sadda SR, Wu Z, Walsh AC, et al. Errors in retinal thickness measurements obtained by optical coherence tomography. Ophthalmology 2006;113:285-93.

3. Fung AE, Lalwani GA, Rosenfeld PJ, et al. An optical coherence tomography-guided, variable dosing regimen with intravitreal ranibizumab (Lucentis) for neovascular age-related macular degeneration. Am J Ophthalmol 2007;143:566-83.

4. Sadda SR, Joeres S, Wu Z, et al. Error correction and quantitative subanalysis of optical coherence tomography data using computer-assisted grading. Invest Ophthalmol Vis Sci 2007;48:839-48.

5. Joeres S, Tsong JW, Updike PG, et al. Reproducibility of quantitative optical coherence tomography subanalysis in neovascular age-related macular degeneration. Invest Ophthalmol Vis Sci 2007;48:4300-7.

6. Holz FG, Jorzik J, Schutt F, et al. Agreement among ophthalmologists in evaluating fluorescein angiograms in patients with neovascular age-related macular degeneration for photodynamic therapy eligibility (FLAP-study). Ophthalmology 2003;110:400-5.