Bag of forests for modelling of tissue energy interaction in optical coherence tomography for atherosclerotic plaque susceptibility assessment

Atherosclerosis assessment using the high resolution of intravascular optical coherence tomography (IV-OCT) enables the visualisation of tissue pools which are susceptible to rupture. However, the stochastic nature of OCT speckles leads to subjectivity in interpretation. In this paper, we present a framework for to OCT tissue energy interaction driven marker for identifying susceptible pool of tissues. The local statistics of OCT speckles is coupled with estimate of optical attenuation and signal confidence measures in a multiscale approach. Further, a bag of random forests is trained on a bank of 22 OCT cross-sections where susceptible tissues have been identified by correlating with the histology. Evaluating the performance on 22 other diverse cross-sections substantiates robustness of the framework. It is envisaged that this framework would help in objective assessment of plaque susceptibility and aid in more confident decision making during interventions.