Comparison of two automated methods for QT interval measurement

In this paper we compared two methods of automated QT interval measurement on standard ECG databases: the Root-Mean-Square (RMS) lead combining method aimed at QT monitoring and the method of median of lead-by-lead QT interval measurements. We used the PhysioNet PTB (N=548) and CSE measurement (N=125) standard databases. Both have reference QT interval measurements from a group of annotators. The last 10 seconds of each PTB record was downsampled from 1000 sample per second (sps) and an amplitude resolution of 1 muV to 500 sps and 5 muV in order to match the CSE set. PTB records #205 and #557 were excluded due to ventricular paced rhythm and artifact, respectively. Twenty five cases were excluded from the CSE set to match the selection of cases for IEC algorithm testing (IEC 60601-2-51). We processed all records using the Philips resting 12- lead ECG algorithm to generate representative beats for QT interval measurement. The RMS method measures QRS onset and end of Ton an RMS waveform constructed from 9 leads I, II, III and V1-V6. The lead-by-lead method takes the median QT interval across leads. The automated QT intervals by the RMS and lead-by-lead methods were compared to the reference manual QT measurements. The mean difference between the lead-by-lead QT and the reference QT was 1.7plusmn9.7 ms and 12.4plusmn23.0 ms (mean plusmnstandard deviation (SD)) for the CSE and PTB sets respectively. For the RMS method, the mean difference was -2.8plusmn11.1 ms and 10.3plusmn20.9 ms. F-tests indicate that the standard deviation between methods is not significantly different for the CSE set (P=0.18) or the PTB set (P=0.77). The lead-by-lead and RMS methods perform similarly, leading to the conclusion that the choice between them should be based on considerations such as the number of leads available or computational efficiency.