Purpose:
Retinal hyperspectral imaging offers the potential to capture in vivo metabolic and physiologic information by identifying unique patterns of fundus reflectance that classify tissues and quantify cellular metabolites. Distinctive hemoglobin spectral signatures can provide potentially both qualitative and quantitative oxygen saturation assessment. Current ophthalmic hyperspectral imaging systems typically required long exposure times and therefore are plagued with motion artifacts from saccades. Snap–shot spectroscopy avoids these potential pitfalls by capturing a full spatial–spectral image cube in a single camera flash.
Methods:
6 eyes of 3 normal subjects were imaged with a computed tomography ‘snap–shot’ imaging spectrometer (CTIS) that obtains full spectral information for each pixel in a 2–D scene with a single flash. Expectation maximization routines were used to reconstruct the spectral image cube from captured data. Optical density within the blood vessel was then calculated by taking the logarithm of the ratio of perivascular intensity to vascular intensity. These curves were compared to simulated oxy–deoxyhemoglobin reflectance curves with Pearson Product Moment Correlation Coefficients analysis.
Results:
Pseudocolor oxygen–saturation correlation maps representing tissue oxygenation across the image were generated by correlating captured spectra with simulated spectra (fig). The average correlation for subgroups of spectra (72 points each) was excellent (r=0.9, p<0.01). Areas of high correlation, however, occurred in patches with large differences existing between adjacent vessel segments. This may be due to lighting variation along the vessel length and will need to be investigated in future studies.
Conclusions:
Snap–shot retinal imaging spectroscopy may be a useful and practical tool for both qualitative and quantitative oxygen saturation measurements within retinal vessels. Expansion of this capability to include other clinically–relevant compounds such as lipofuscin and cytochrome aa3 holds promise as a functional imaging modality for use both in clinical trials and in clinical practice.
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • blood supply • diabetic retinopathy