Graphics processor unit acceleration enables realtime endovascular Doppler optical coherence tomography imaging

摘要

Endovascular Optical Coherence Tomography (OCT) has previously been used in both bench-top and clinical environments to produce vascular images, and can be helpful in characterizing, among other pathologies, plaque build-up and impedances to normal blood flow. The raw data produced can also be processed to yield high-resolution blood velocity information, but this computation is expensive and has previously only been available a posteriori using post-processing software. Real-time Doppler OCT (DOCT) imaging has been demonstrated before in the skin and eye, but this capability has not been available to vascular surgeons. Graphics Processing Units (GPUs) can be used to dramatically accelerate this type of distributed computation. In this paper we present a software package capable of real-time DOCT processing and circular image display using GPU acceleration designed to operate with catheter-based clinical OCT systems. This image data is overlayed onto structural images providing clinicians with live, high-resolution blood velocity information to complement anatomical data.