Correcting common OCT artifacts enhances plaque classification and identification of higher-risk plaque features

Benn Jessney,Xu Chen,Sophie Gu,Adam Brown,Daniel Obaid,Charis Costopoulos,Martin Goddard,Nikunj Shah,Hector Garcia-Garcia,Yoshinobu Onuma,Patrick Serruys,Stephen P Hoole,Michael Mahmoudi,Michael Roberts,Martin Bennett

DOI: https://doi.org/10.1016/j.carrev.2024.06.023

2024-07-01

Abstract:Background: Optical coherence tomography (OCT) is used widely to guide stent placement, identify higher-risk plaques, and assess mechanisms of drug efficacy. However, a range of common artifacts can prevent accurate plaque classification and measurements, and limit usable frames in research studies. We determined whether pre-processing OCT images corrects artifacts and improves plaque classification. Methods: We examined both ex-vivo and clinical trial OCT pullbacks for artifacts that prevented accurate tissue identification and/or plaque measurements. We developed Fourier transform-based software that reconstructed images free of common OCT artifacts, and compared corrected and uncorrected images. Results: 48 % of OCT frames contained image artifacts, with 62 % of artifacts over or within lesions, preventing accurate measurement in 12 % frames. Pre-processing corrected >70 % of all artifacts, including thrombus, macrophage shadows, inadequate flushing, and gas bubbles. True tissue reconstruction was achieved in 63 % frames that would otherwise prevent accurate clinical measurements. Artifact correction was non-destructive and retained anatomical lumen and plaque parameters. Correction improved accuracy of plaque classification compared against histology and retained accurate assessment of higher-risk features. Correction also changed plaque classification and prevented artifact-related measurement errors in a clinical study, and reduced unmeasurable frames to <5 % ex-vivo and ~1 % in-vivo. Conclusions: Fourier transform-based pre-processing corrects a wide range of common OCT artifacts, improving identification of higher-risk features and plaque classification, and allowing more of the whole dataset to be used for clinical decision-making and in research. Pre-processing can augment OCT image analysis systems both for stent optimization and in natural history or drug studies.