Couderc JP, Zareba W, Moss AJ.
We designed a new technique for the quantification of repolarization changes using the wavelet transformations (WT) of the median beats of the 12-lead ECGs. WT based on the second Gaussian derivative functions were used. A 700 ms window starting from the R peak was analyzed using WT in 11 non-carriers and 13 carriers of SCN5A mutation (LQT3). Ten frequency bandwidths were analyzed: 5-9, 6-11, 7-14, 9-18, 12-24, 15-30, 20-39, 25-49, 32-63 and 42-80 Hz. The most significant frequency bandwidths and time intervals for detecting T wave abnormalities in each lead were defined.
Results: Lead V5 showed the most significant differentiation in time-frequency (TF) components of repolarization between carriers and non-carriers using the 130-340 ms segment after R peak and in 5-11 Hz. A WT quantitative parameter of T wave abnormalities (WQTa) was created. The median WQTa value of the 12 leads was compared to the median QT value (QTc: corrected by Bazett's formula). WQTa improved sensitivity and specificity of automatic ECG diagnosis of LQT3 carriers by 18% and 15%, respectively, in comparison with the QTc measurement (see table).
Conclusions: 1) LQTS carriers have different TF components of repolarization from their unaffected relatives, which can be identified using WT. 2) WT improves discrimination of LQT3 carriers without the need for T wave endpoint localizations.
|
|
Non-Car. |
Car. |
Sensitivity |
Specificity |
|
N |
13 |
11 |
11 |
13 |
|
Mdn QTc (s) |
0.44 ±0.11 |
0.62 ±0.19* |
72.7 % |
76.9 % |
|
Mdn WQTa (mV) |
3.8 ±1.5 |
0.7 ±0.9** |
90.9 % |
92.3 % |
*
: p<0.01, **: p<0.0001, Car.: Carriers, Mdn: Median