The chemical composition of vascular lesions, an important determinant of plaque progression and rupture, can not presently be determined in vivo. Prior studies have shown that Raman spectroscopy can accurately quantify the amounts of major lipid classes and calcium salts in homogenized coronary artery tissue. This study determines how the relative cholesterol content, which is calculated from Raman spectra collected at the luminal surface of an artery, is related to its depth in an intact arterial wall. Raman spectra of human atherosclerotic plaques were measured after thin tissue layers were successively placed on them. From these spectra, relative cholesterol contents were calculated and used to determine how cholesterol signal strength is attenuated by overlaying tissue. Then, intact artery samples (n=13) were examined spectroscopically, sectioned and stained specifically for cholesterol. Images of these sections were digitized, and image intensities were related to cholesterol content. These cholesterol amounts were weighed appropriately for depth into the tissue and area-integrated for comparison with spectroscopy results. A decaying exponential curve was fit to the layer study data (r2=0.97) and showed that ∼300 μm of tissue attenuates cholesterol signals by 50%. In intact plaques, the spectroscopically-determined cholesterol amounts correlated strongly and linearly with those determined by digital microscopy (r2=0.94). With Raman spectroscopy techniques, the cholesterol content of a lesion can be determined by properly accounting for its depth into an arterial wall. Our results suggest that chemical concentrations in an artery wall could be mapped throughout its thickness, possibly by combining Raman spectroscopy methods with other techniques.

We present a novel multi-resolution variational framework for vascular optical coherence elastography (OCE). This method exploits prior information about arterial wall biomechanics to produce robust estimates of tissue velocity and strain, reducing the sensitivity of conventional tracking methods to both noise- and strain-induced signal decorrelation. The velocity and strain estimation performance of this new estimator is demonstrated in simulated OCT image sequences and in benchtop OCT scanning of a vascular tissue sample.