Background: Measurements of parameters from electrocardiograms (ECGs) are still largely performed from paper ECG records. Recent guidelines from regulatory agencies and, in particular, the requirement of the Food and Drug Administration to enforce the digital submission of annotated ECGs have triggered significant efforts in the pharmaceutical industry, which, to comply with the new guidelines, is adopting digital ECG technology. At the same time, the new requirements justify the need for tools to convert existing paper ECG records into digital format, particularly for retrospective studies.
Methods: This article presents ECGScan, a computer application developed for the conversion of paper ECG records to digital ECG files. An image processing engine is used to first detect the underlying grid and, subsequently, to extrapolate the ECG waveforms using a technique based on active contour modeling.
Results: ECGScan was validated using a set of 60 ECGs for which both the original digital waveform and paper printouts were available. Sample-by-sample comparisons provided evidence of a robust wave reconstruction (root mean square value from 169 PQRST complexes was 16.8+/-11.8 microV). Semiautomatic measurements of QT intervals performed on 144 complexes also indicated a strong agreement between original and derived ECGs (DeltaQT=0.577+/-5.41 milliseconds).
Conclusions: ECGScan provides a robust reconstruction of a digital ECG, both in waveform reconstruction and in QT measurements performed on original (digital) ECGs and on digitized ECGs from paper printouts.