AIMS: Optical coherence tomography (OCT) of a bioresorbable vascular scaffold (BVS) produces a highly reflective signal outlining struts. This signal interferes with the measurement of strut thickness, as the boundaries cannot be accurately identified, and with the assessment of coverage, because the neointimal backscattering convolutes that of the polymer, frequently making them indistinguishable from one another. We hypothesise that Gaussian line spread functions (LSFs) can facilitate identification of strut boundaries, improving the accuracy of strut thickness measurements and coverage assessment. METHODS AND RESULTS: Forty-eight randomly selected BVS struts from 12 patients in the ABSORB Cohort B clinical study and four Yucatan minipigs were analysed at baseline and follow-up (six months in humans, 28 days in pigs). Signal intensities from the raw OCT backscattering were fit to Gaussian LSFs for each interface, from which peak intensity and full-width-at-half-maximum (FWHM) were calculated. Neointimal coverage resulted in significantly different LSFs and higher FWHM values relative to uncovered struts at baseline (p<0.0001). Abluminal polymer-tissue interfaces were also significantly different between baseline and follow-up (p=0.0004 in humans, p<0.0001 in pigs). Using the location of the half-max of the LSF as the polymer-tissue boundary, the average strut thickness was 158±11 µm at baseline and 152±20 µm at six months (p=0.886), not significantly different from nominal strut thickness. CONCLUSIONS: Fitting the raw OCT backscattering signal to a Gaussian LSF facilitates identification of the interfaces between BVS polymer and lumen or tissue. Such analysis enables more precise measurement of the strut thickness and an objective assessment of coverage.
AIMS: Optical coherence tomography (OCT) of a bioresorbable vascular scaffold (BVS) produces a highly reflective signal outlining struts. This signal interferes with the measurement of strut thickness, as the boundaries cannot be accurately identified, and with the assessment of coverage, because the neointimal backscattering convolutes that of the polymer, frequently making them indistinguishable from one another. We hypothesise that Gaussian line spread functions (LSFs) can facilitate identification of strut boundaries, improving the accuracy of strut thickness measurements and coverage assessment. METHODS AND RESULTS: Forty-eight randomly selected BVS struts from 12 patients in the ABSORB Cohort B clinical study and four Yucatan minipigs were analysed at baseline and follow-up (six months in humans, 28 days in pigs). Signal intensities from the raw OCT backscattering were fit to Gaussian LSFs for each interface, from which peak intensity and full-width-at-half-maximum (FWHM) were calculated. Neointimal coverage resulted in significantly different LSFs and higher FWHM values relative to uncovered struts at baseline (p<0.0001). Abluminal polymer-tissue interfaces were also significantly different between baseline and follow-up (p=0.0004 in humans, p<0.0001 in pigs). Using the location of the half-max of the LSF as the polymer-tissue boundary, the average strut thickness was 158±11 µm at baseline and 152±20 µm at six months (p=0.886), not significantly different from nominal strut thickness. CONCLUSIONS: Fitting the raw OCT backscattering signal to a Gaussian LSF facilitates identification of the interfaces between BVS polymer and lumen or tissue. Such analysis enables more precise measurement of the strut thickness and an objective assessment of coverage.
Authors: Ancong Wang; Shimpei Nakatani; Jeroen Eggermont; Yoshi Onuma; Hector M Garcia-Garcia; Patrick W Serruys; Johan H C Reiber; Jouke Dijkstra Journal: Biomed Opt Express Date: 2014-09-12 Impact factor: 3.732