OCTA: Non-perfused areas for diagnosis

 

Optical Coherence Tomography Angiography (OCTA) can display blood vessels by visualizing signal changes caused by moving particles, in this case erythrocytes in the blood. It is able to visualize areas of reduced capillary perfusion which occur in various diseases, including glaucoma.

Studies have usually quantified the decreased perfusion by using vessel density (VD), which is the percentage of vessel pixels in the image. However, this method may miss subtle changes which occur early during the course of the disease. Newer methods quantify the non-perfused or inter-capillary areas between the vessels. This depends on accuracy of vessel segmentation, which can be a limiting factor for reliability in certain cases.

Intercapillary areas computed from perfusion-distance measures are less sensitive to errors in the vessel segmentation since the distance to the next vessel is only slightly changed if gaps are present in the segmentation. Therefore, inter-capillary areas computed from perfusion-distance measures are less sensitive to errors.

Schottenhamml and colleagues have recently reported a new method based on features computed from the probability density function of these perfusion-distance areas. [1]

Deep learning methods (Artificial Intelligence, AI) can detect areas of reduced capillary perfusion. Therefore, these regions can act as a very good biomarker for detecting glaucoma. Unfortunately, vessel density, the commonly used parameter, is not sensitive enough to detect small changes occurring in these areas. Therefore, the authors utilized the non-perfusion or inter-capillary areas to quantify the areas between the vessels. This approach is probably more sensitive since even if only a few vessels are not visible on OCTA and VD is not much affected, the region of non-perfused areas can change appreciably.

In order to quantify the inter-capillary areas, a very accurate segmentation of the vascular network is needed, since small errors and gaps in the segmentation will affect the results significantly. The authors have found that a more robust alternative to using inter-capillary areas is to use perfusion distance. This approach computes the distance from any pixel to its next vessel pixel and is consequently not as sensitive to small errors in vessel segmentation.

REFERENCE:

[1] Schottenhamml J, Würfl T, Ploner S, Husvogt L, Lämmer R, Hohberger B, Maier A, Mardin C. Glaucoma detection using non-perfused areas in OCTA. Sci Rep. 2024 May 5;14(1):10306. doi: 10.1038/s41598-024-60839-4. PMID: 38705883; PMCID: PMC11070420.