A Quantitative and Standardized Method for the Evaluation of Choroidal Neovascularization Using MICRON III Fluorescein Angiograms in Rats

Abstract

Introduction: In-vivo imaging of choroidal neovascularization (CNV) has been increasingly recognized as a valuable tool in the investigation of age-related macular degeneration (AMD) in both clinical and basic research applications. Arguably the most widely utilised model replicating AMD is laser generated CNV by rupture of Bruch's membrane in rodents. Heretofore CNV evaluation via in-vivo imaging techniques has been hamstrung by a lack of appropriate rodent fundus camera and a non-standardised analysis method. The aim of this study was to establish a simple, quantifiable method of fluorescein fundus angiogram (FFA) image analysis for CNV lesions.

Methods: Laser was applied to 32 Brown Norway Rats; FFA images were taken using a rodent specific fundus camera (Micron III, Phoenix Laboratories) over 3 weeks and compared to conventional ex-vivo CNV assessment. FFA images acquired with fluorescein administered by intraperitoneal injection and intravenous injection were compared and shown to greatly influence lesion properties. Utilising commonly used software packages, FFA images were assessed for CNV and chorioretinal burns lesion area by manually outlining the maximum border of each lesion and normalising against the optic nerve head. Net fluorescence above background and derived value of area corrected lesion intensity were calculated.

Results: CNV lesions of rats treated with anti-VEGF antibody were significantly smaller in normalised lesion area (p < 0.001) and fluorescent intensity (p < 0.001) than the PBS treated control two weeks post laser. The calculated area corrected lesion intensity was significantly smaller (p < 0.001) in anti-VEGF treated animals at 2 and 3 weeks post laser. The results obtained using FFA correlated with, and were confirmed by conventional lesion area measurements from isolectin stained choroidal flatmounts, where lesions of anti-VEGF treated rats were significantly smaller at 2 weeks (p = 0.049) and 3 weeks (p < 0.001) post laser.

Conclusion: The presented method of in-vivo FFA quantification of CNV, including acquisition variable corrections, using the Micron III system and common use software establishes a reliable method for detecting and quantifying CNV enabling longitudinal studies and represents an important alternative to conventional CNV quantification methods.

Fig 1. Illustration of area and intensity assessment of CNV lesions in FFA and choroidal flatmount images including corresponding colour Ffundus and unmodified FFA.

(A-C): Colour fundus, FFA and representative hyperfluorescent CNV lesion analysis by freehand selection tool in ImageJ under digital magnification of the green channel from a FFA image produced by the Micron III system. Native background fluorescence intensity avoiding choroidal vessels was measured within a defined annulus (green) around the lesion. (D-F): Corresponding images of lesion analysis for chorio-retinal burn. (G): Representative CNV lesion analysis by freehand selection tool in ImageJ of Isolectin GS-IB4-AF488 stained choroidal flatmount. *Note scale bar applicable to Fig 1G only.

Fig 2. Fluorescein Administration Route CNV Intensity Comparison.

CNV intensity analysis of unprocessed FFA images with fluorescein administration by intravenous (IV) and intraperitoneal (IP) injection. Each time point represents the mean grey value ± SD (n = 4). FFA images with IP fluorescein were taken at 1 frame/second for 120, and 1 frame/5 seconds thereafter up to 10 minutes post injection. IV fluorescein angiograms were taken at 30 frames/second for 120 seconds. Intensity data from 130–470 seconds not shown.

Fig 3. Representative Micron III Images with Micrographs of conventional Histopathological preparations.

Colour fundus photo (A) and fluorescein angiogram (B) of a Brown Norway rat exhibiting 4 choroidal neovascular lesions generated by rupture of the Bruch’s membrane by laser. Fluorescein angiogram (B) taken at 10.2 seconds post intravenous injection, corresponding to peak CNV fluorescence. Corresponding choroidal flatmount image (C) of the same eye taken at 2 weeks post laser stained with Isolectin-IB4 conjugated with Alexa Fluor 488. *Scale bar represents 500μm and is applicable to Fig 3C only. Representative micrograph of haematoxylin and eosin stained section of (D) Chorio-Retinal Burn at 3 weeks post laser (E) Retina without laser treatment (F) CNV lesion at 3 weeks post laser (D) Classical fusiform shaped sub retinal neovascular lesions are observed in both treatment groups confirming CNV formation by Bruch’s Membrane rupture by laser. *Scale bar represents 100μm and is applicable to Fig 3D, 3E and 3F only. *(Vitr = Vitreous, GCL = Ganglion Cell Layer, IPL = Inner Plexiform Layer, INL = Inner Nuclear Layer, OPL = Outer Plexiform Layer, ONL = Outer Nuclear Layer, IS = Inner Segment, OS = Outer Segment, RPE = Retinal Pigment Epithelium, Chor. = Choroid)

Fig 4. Fluorescein Angiogram CNV Area Analysis.

Calculated area of Laser Burn without CNV and CNV lesions receiving anti-VEGF treatment versus PBS treatment of FFA images using manual border delineation by trained specialists blinded to experimental treatment in ImageJ, normalised to optic nerve head area. Values represent mean lesion area in pixels ± SD (n = 16). * Denotes statistically significant (p<0.05) difference in the calculated normalised lesion area tested by two-tailed Student’s t-test. ** Denotes statistically significant (p<0.01) difference in the calculated normalised lesion area tested by two-tailed Student’s t-test. *** Denotes statistically significant (p<0.001) difference in the calculated normalised lesion area tested by two-tailed Student’s t-test.

Fig 5. Representative micrographs of CNV lesions using choroidal flatmount and corresponding area calculation.

(A) Calculated CNV area on choroidal flatmounts using free hand selection technique in ImageJ, adjusted from pixels to μm². Each column represents the mean area ± SD (n = 16). Representative fluorescent micrographs of neovascular lesions of PBS injected (B) and anti-VEGF treated eyes (C) at 2 weeks post laser, produced by routine choroidal flatmount and stained with Isolectin-IB4 conjugated with Alexa Fluor 488. Note reduced vascular budding in anti-VEGF treated eyes at 2 weeks. * Denotes statistically significant (p<0.05) difference in the measured CNV area tested by two-tailed Student’s t-test. *** Denotes statistically significant (p<0.001) difference in the measured CNV area tested by two-tailed Student’s t-test.

Fig 6. Fluorescein Angiogram CNV Net Fluorescence Analysis.

Calculated net fluorescence above background of FFA images at 10.2 seconds post intravenous fluorescein injection of laser burns without CNV generation versus CNV lesions receiving anti-VEGF treatment and lesions receiving PBS. Values represent net average grey value ±SD (n = 16). ** Denotes statistically significant (p<0.01) difference in the calculated net fluorescence between treatment groups tested by two-tailed Student’s t-test. *** Denotes statistically significant (p<0.001) difference in the calculated net fluorescence between treatment groups tested by two-tailed Student’s t-test.

Fig 7. Area corrected lesion fluorescent intensity in fluorescein angiogram.

Calculated average area corrected lesion fluorescent intensity of Laser Burn without CNV generation versus generated CNV with anti-VEGF treatment and PBS treatment. The ‘corrected lesion intensity’ of the hyperfluorescent region was calculated by multiplying the ‘net fluorescence intensity’ by the normalised calculated CNV area. Values represent net average grey value ±SD (n = 16). ** Denotes statistically significant (p<0.01) difference in the calculated total lesion grey value tested by two-tailed Student’s t-test. *** Denotes statistically significant (p<0.001) difference in the calculated total lesion grey value tested by two-tailed Student’s t-test. # p value determined by two-tailed Student’s t-test comparing respective experimental group with corrected lesion grey values from PBS treated rats.