Abstract
Purpose:
To develop an objective method of quantifying choroidal neovascularization (CNV) areas based on fluorescein angiography (FA) images obtained from a confocal scanning laser ophthalmoscope (SLO) and to determine the method’s accuracy at various stages of CNV growth.
Methods:
Five CNV lesions per eye were induced in Long-Evans rats using an Ar-green laser (400 mW, 50 μm, 0.1 sec). To test different stages of CNV, two groups received 5μl of 40mg/ml triamcinolone acetonide (TAC) or 40 mg/ml dexamethasone sodium phosphate (DSP) solution immediately after the laser treatment. A third group received no drug and served as control. FA was performed weekly for 5 weeks using SLO. To quantify CNV areas, late-phase FA images of each lesion were processed using the “multi-Otsu threshold” plug-in for ImageJ software. Three regions were defined: background, diffuse leakage, and CNV. Areas were calculated using a pixel-scaling factor based on the dimensions of the eye and field of view of the image. Each lesion was compared to its corresponding histological area.
Results:
The “background” threshold output contained only the background of the picture (deep choroidal vessels). The “diffuse leakage” output included the hyperfluorescence around the CNV lesion and often major and small vessels adjacent to the lesion. The “CNV” output included the bright center of the CNV lesion and occasionally major blood vessels. CNV lesions from control animals increased in size until week 3 and did not significantly change thereafter (p=0.02); at week 5, the corresponding CNV area was 0.046 mm2. Histology confirmed significant CNV in these animals. The multi-level Otsu thresholding technique consistently yielded CNV areas 68% the size determined via histology (0.068 mm2). TAC-treated lesions were initially well-defined and comparable in size to the original laser burn; by week 5, lesions resolved and lacked CNV. The method was able to track weekly changes in CNV areas. DSP-treated CNV lesion areas were 3.5 times smaller at week 4 compared to week 2 (p=0.05). Histology confirmed less CNV growth in DSP-treated eyes than control.
Conclusions:
A quantitative area measurement technique has been developed to monitor the progression of CNV in vivo. This approach provides the ability to objectively quantify CNV lesion growth and can be applied to determine the efficacy of various therapeutics on treating CNV.
Keywords: 453 choroid: neovascularization •
551 imaging/image analysis: non-clinical •
552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)