developed biopotential fiber sensors (BFSs) will reduce the noise artifacts and thus result in a better ST-segment measurement sensitivity. Methods: We recorded 12-lead ECGs simultaneously from wet gel electrodes and from low-mass, nonmetallic BFS assemblies in routine stress testing of the obese patients. The stress test ECG signals were recorded using the X-Scribe II system and a hi-resolution H-12 Holter, both from Mortara Instruments. In each ECG comprising preexercise, peakexercise, and postexercise phases, we have computed mean sinus rhythm cycles then synchronized them with the raw signal and subtracted. The resulting “residue” (noise) was quantified by computing its SD and rootmean-square of successive differences in each signal. We then computed ST levels and J-ST delays in both ECGs for individual beats and on all 12 mean complexes. We evaluated the results using graphic trends of the computer-generated ST measurements and histograms of the differences between the BFS data–generated ST values and the Ag-AgCl electrode data ST values. Two performance measures, the ST measurement sensitivity and the ST measurement positive predictivity, were used to quantify the performance results of the comparison study. This technique allows for accurate evaluation and quantification of the impact of noise on the performance of the ST measurements in both data sets. Results and conclusions: The root-mean-square of successive differences residuum values were significantly lower with the BFS sensors (16.62 ± 2.71 μV) than with the Ag-AgCl electrodes connected to patient cables (37.47 ± 11.69 μV, P b .001 and P b .002). The ST measurement sensitivity and ST measurement positive predictivity were significantly higher for the BFS signals: 75% vs 52% and 64% vs 46, respectively, for Ag-AgCl electrode signals. We conclude that ECGs from BFS were less noisy than from AgAgCl electrodes, resulting in higher ST-segment measurement sensitivity.
Read full abstract