Automatic multilead VCG based approach for QT interval measurement

The precision of QT interval measures depends on the correct location of QRS and T boundaries and also on the analysing lead. The usual automatic strategy to deal with multiple leads is to apply post-processing decision rules for selecting one of the single-lead (SL) measures. We proposed a multilead VCG strategy that locates the onset and the end of the QT interval, attending to an optimal transformed lead according to the spatial characteristics of the VCG. The performance was evaluated over the CSE database considering different VCG systems: corrected Frank system (lead set F), pseudo-orthogonal leads V5, aVF and V2 (M) and the X, Y and Z leads derived by Dower transformation (D). Multilead delineation over F achieved a better results than any single lead by itself or any other lead set, with an error dispersion similar to SL over 12 leads plus decision rules. The multilead approach is validated in the PTB database, in the Physionet/CinC Challenge 2006, with final scores of 27.04, 27.81 and 28.96, over F, M and D, respectively.