Reduced blood flow by laser speckle flowgraphy after 125I-plaque brachytherapy for uveal melanoma

Abstract

Background: To determine whether reductions in retinal and choroidal blood flow measured by laser speckle flowgraphy are detected after ¹²⁵I-plaque brachytherapy for uveal melanoma.

Methods: In a cross-sectional study, retinal and choroidal blood flow were measured using laser speckle flowgraphy in 25 patients after treatment with ¹²⁵I-plaque brachytherapy for uveal melanoma. Flow was analyzed in the peripapillary region by mean blur rate as well as in the entire image area with a novel superpixel-based method. Relationships between measures were determined by Spearman correlation.

Results: Significant decreases in laser speckle blood flow were observed in both the retinal and choroidal vascular beds of irradiated, but not fellow, eyes. Overall, 24 of 25 patients had decreased blood flow compared to their fellow eye, including 5 of the 6 patients imaged within the first 6 months following brachytherapy. A significant negative correlation between blood flow and time from therapy was present.

Conclusions: Decreases in retinal and choroidal blood flow by laser speckle flowgraphy were detected within the first 6 months following brachytherapy. Reduced retinal and choroidal blood flow may be an early indicator of microangiographic response to radiation therapy.

Keywords: Laser speckle flowgraphy; Radiation retinopathy; Retinal blood flow; Uveal melanoma.

Fig. 1

Imaging modalities to assess relative flow changes after.¹²⁵I-plaque brachytherapy. A, C Laser speckle flowgraphy (LSFG) scans of the optic nerve head and surrounding retina of fellow (top) and irradiated (bottom) eyes. The color scale represents the highest mean blur rate as red and the lowest as blue. B, D Dosimetry maps constructed to calculate radiation dose to the optic nerve head (indicated by green arrow). Representative images are shown from treated and fellow eyes of a patient without clinical signs of radiation retinopathy (seen 18 months after treatment, top) and with radiation retinopathy (seen 68 months after treatment, bottom) as diagnosed by color fundus photography. Note that in the irradiated eyes the branch retinal arterioles and venules show reduced flow. The large blue (low flow) areas correspond to reduced choroidal blood flow. Dosimetry models constructed using Plaque Simulator (v.6.6, EyePhysics, LLC)

Fig. 2

Peripapillary relative blood flow velocity after.¹²⁵I-plaque brachytherapy by standard LSFG analysis. A Representative LSFG image with overlaid measurement annuli for analysis. B Flow in the annulus is separated into areas primarily representing retinal (white pixels) and choroidal (black pixels) vascular supply following automated histogram thresholding. C Pairwise comparison of retinal vessel flow in irradiated (purple) and fellow (black) eyes. D Differences in retinal vessel flow between irradiated and fellow eye for each patient. E Pairwise comparison of choroidal flow in irradiated (green) and fellow (black) eyes. F Differences in choroidal flow between irradiated and fellow eye for each patient. G Retinal vessel flow in irradiated (purple) and fellow (black) eyes as a function of time from treatment. H Choroidal flow in irradiated (green) and fellow (black) eyes as a function of time from treatment. * p < 0.05 by two-tailed paired t-test (C, F). Slope (β) and its p value calculated by simple linear regression (G, H)

Fig. 3

Categories of relative blood flow velocity by superpixel segmentation of the entire scan area. (A) Composite maps describing the average mean blur rate (MBR) in superpixels from an irradiated (left) and fellow (right) eye of a patient seen at 54 months post-brachytherapy. The color scale represents the highest mean blur rate as red and the lowest as blue. White superpixels contain undefined values over the 4 s period of data acquisition. (B) Histogram of the percent of superpixels with five different ranges of MBR values, following the same color scale as in (A). (C-G) Composite images highlighting superpixels with a mean MBR < 5 (C), a mean MBR of 5 to 10 (D), a mean MBR of 10 to 15 (E), a mean MBR of 15 to 20 (F), and a mean MBR ≥ 20 (G). The color scales for (C-G) match the colors in (A) and (B). Note that in this example there is a shift in the percent of superpixels in the irradiated eye away from the highest (red) to the lowest (blue) blood flow range, compared to the fellow non-irradiated eye in both the histogram and the blood flow pixel maps

Fig. 4

Relative blood flow velocity after.¹²⁵I-plaque brachytherapy by superpixel analysis. (A) Representative composite image of MBR by superpixel analysis from the irradiated eye of a patient seen 18 months post-brachytherapy. (B) Histogram distribution of percent superpixels for each MBR range in (A). Colors match the color scale shown in (A). (C, D) Superpixel segmented composite image of MBR ≥ 20 (C) and < 5 (D). (E, G) Pairwise comparison of retinal flow area (MBR ≥ 20, E) and choroidal flow area (MBR < 5, G) in irradiated and fellow eyes. (F, H) Differences in retinal (F) and choroidal (H) flow area between irradiated and fellow eye for each patient. (I, J) Percent superpixels with MBR ≥ 20 (I) and MBR < 5 (J) in the irradiated and fellow eyes as a function of time from treatment. ** p < 0.01, **** p < 0.0001 by two-tailed paired t-test (E, G). Slope (β) and its p value calculated by simple linear regression (I, J)

Fig. 5

Spearman correlation matrix of readouts. Interrelationships between outcome measures are reported for each statistically significant measure and color coded for degree of significance. Mean choroidal blur rate and mean retinal vessel blur rate describe flow rate within the peripapillary annulus (Fig. 2). The % superpx MBR < 5 and % superpx ≥ 20 describe the percent area of choroidal and retinal flow calculated using a superpixel method (Fig. 4). Distance from the tumor margin to the optic disc, as well as dose to the optic disc, were obtained from clinical treatment planning records