Detecting predisposition to torsade de points using a PCA-based method

Torsade de points (TDP) is a form of polymorphic ventricular tachycardia. It is associated with alternation ofT wave and prolongation of the QT interval. The primary objective of this work is to find characteristics of the T waves before and after TDP using principal component analysis (PCA). PCA was applied on T wave of 60 normal 24-hour tapes and 10 TDP 24-hour tapes from different studies recorded during ´Dofetilide´ clinical trials (Pfizer, Inc.). All signals were first conditionedby eliminating baseline wander, detecting their significant points and extracting T waves of each channel into a data matrix. Afterwards, for every zero-centred data matrix, a covariance matrix and its corresponding eigenvalues and eigenvectors were calculated. Then, every beat is explained in terms of the eigenvectors delivering scores that characterise individual T wave. Results showed that Standard deviation (SD) of PCA scores for TDP patients before TDP syndrome are clearly higher than in case of healthy subjects.