Abstract
Purpose:
The organization of stromal collagen and keratocyte distribution may be altered in corneal disease (e.g., keratoconus) and following crosslink therapy. Because of such alterations, changes in optical and ultrasound backscatter within the stroma might be expected. We report here a novel approach to characterization and display of stromal microstructural changes via quantitative ultrasound analysis.
Methods:
We used a 60 MHz transducer with a 2-mm aperture and 6-mm focal length. We initially scanned rabbits (8 eyes) pre-and post-crosslinking, acquiring 3D rectilinear sets of 200 scans at 10-mm intervals. We then adapted the system to scan human subjects using an immersion setup with arc-shaped scan geometry matched approximately to the radius of the human cornea. Phase-resolved data were acquired in 6 radial planes at 30-degree intervals over a 5.5 mm diameter region. We generated spectral parameter images of the stroma representing spectral slope (dB/MHz), 0-MHz intercept (dB) and midband fit (dB) by using a sliding window approach such that the analysis window was approximately 5 wavelengths axially by 5 vectors laterally (approximately one beamwidth). We then performed radial interpolation to produce maps of spectral parameters representing mean spectral parameters over the full stromal depth, or over selectable depth ranges.
Results:
Cross-linked rabbit corneas showed variable change in backscatter and frequency-dependence; however, changes in midband backscatter and spectral slope (dB/MHz) were correlated with corneal thickness change. Initial results in human eyes demonstrated higher backscatter in the anterior stroma in the normal cornea, even after compensation for attenuation. The acquisition of 3D data in the rabbit eye and in humans allowed generation of en face images, mapping backscatter parameters as a function of position and depth-range.
Conclusions:
Alteration of stromal organization by disease or therapy may result in optical backscatter changes detectable by Scheimpflug, OCT or high-frequency ultrasound. The 60 MHz probe used in this study is more sensitive to small tissue inhomogeneities than UBM probes typically operating in the 35-40 MHz range. Also, ultrasound is sensitive to acoustic impedance rather than optical refractive index, and thus is complementary to optical methods. En face depth-wise display of stromal acoustic backscatter offers a new means for assessment of corneal disease and therapy.