Page 294 - Computational Retinal Image Analysis

292 CHAPTER 14 OCT fluid detection and quantification

Performance in exudative disease diagnosis and fluid detection has been shown to be
very high in a number of papers reviewed in this chapter. Interrater variability was also
demonstrated to be small and gold standards for training OCT classification methods
are generally available. This is important, as only a very high sensitivity and specificity
are clinically accepted for fluid detection. False negatives and false positives bring a
high risk of vision loss and an unnecessary increased treatment burden, respectively. By
contrast, the segmentation performance level required and most appropriate quantifica-
tion metric are currently unclear and likely depend on the particular clinical use case.
Fluid segmentation methods have shown lower performance, although approaching the
interrater variability. The large interrater variability reflects the difficulty of the annota-
tion task and produces a lower quality of the pixel-wise reference standard for training.
Thus, weakly supervised and unsupervised efforts are of special importance to support
the segmentation methods going beyond the limit of human annotator capabilities.
OCT is a relatively young imaging modality and is still rapidly evolving. In par-
ticular, the speed of acquisition is increasing and devices already operate at over
100,000 A-scans per second [81]. Thus, the next generation of OCT is expected to
have fewer artifacts and a higher SNR, facilitating the tasks of automated fluid de-
tection and segmentation. Clinically applicable automated fluid analysis is expected
to increase the set of readily available quantitative OCT biomarkers, which together
will enable personalized and predictive medicine in macular disease.

Acknowledgments

This work was supported by the Christian Doppler Research Association, the Austrian
Federal Ministry for Digital and Economic Affairs, and the National Foundation for
Research, Technology and Development.

References
[1] U. Schmidt-Erfurth, S.M. Waldstein, A paradigm shift in imaging biomarkers in neovas-
cular age-related macular degeneration, Prog. Retin. Eye. Res. 50 (2016) 1–24.
[2] M. Adhi, J.S. Duker, Optical coherence tomography—current and future applications,
Curr. Opin. Ophthalmol. 24 (3) (2013) 213–221.
[3] E.A. Swanson, J.G. Fujimoto, The ecosystem that powered the translation of OCT from
fundamental research to clinical and commercial impact, Biomed. Opt. Express 8 (3)
(2017) 1638, https://doi.org/10.1364/BOE.8.001638.
[4] U. Schmidt-Erfurth, S. Klimscha, S.M. Waldstein, H. Bogunović, A view of the current
and future role of optical coherence tomography in the management of age-related macu-
lar degeneration, Eye 31 (1) (2017) 26–44, https://doi.org/10.1038/eye.2016.227.
[5] B. Gerendas, C. Simader, G.G. Deak, S.G. Prager, J. Lammer, S.M. Waldstein, M. Kundi,
U. Schmidt-Erfurth, Morphological parameters relevant for visual and anatomic out-
comes during anti-VEGF therapy of diabetic macular edema in the RESTORE trial,
ARVO, 2014.