An​ ​ Unbiased​ ​ Classification​ ​ Algorithm​ ​ for​ ​ Transverse​ ​ Tubule​ ​ Remodeling​ ​ within​ ​ Murine​ ​ Heart​ ​ Failure Models

Transverse tubules (TTs) are the main method of delivery of extracellular calcium into the myocyte, responsible for the initiation of calcium induced calcium release (CICR) necessary to carry out excitation contraction coupling (EC coupling) within ventricular myocytes. Disruptions in the regularly striated TT network are correlated with various degrees of calcium mishandling and are generally observed within heart failure models. Currently, TT remodeling is often judged qualitatively, subject to bias of the experimenter. To address the need to eliminate this bias, we propose a technique that utilizes: 1. Computer vision libraries. 2. Signal processing techniques such as matched filtering that provide signal-to-noise ratios, a quantitative measure of observed remodeling. Combining these methods, an unbiased classification algorithm is presented that is able to quantify both regions of TT loss and longitudinal remodeling as well as characterize whole myocytes by the magnitude of their deviation from the prototypical healthy myocyte. This allows for the expedient, and more importantly, unbiased and reliable classification of TT remodeling observed within murine ventricular myocytes.