Digitizing ECG image: A new method and open-source software code

• Published in: Computer methods and programs in biomedicine Vol. 221; p. 106890
• Main Authors: Fortune, Julian D., Coppa, Natalie E., Haq, Kazi T., Patel, Hetal, Tereshchenko, Larisa G.
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier B.V 01.06.2022
• Subjects: Adolescent; Adult; Digitization; ECG; ECG paper digital conversion; Electrocardiography - methods; Female; Heart Ventricles; Humans; Male; Middle Aged; Paper ECG digitizing; Paper-to-digital conversion; Signal Processing, Computer-Assisted; Software; Young Adult; ECG; Digitization; Paper ECG digitizing; ECG paper digital conversion; Paper-to-digital conversion
• ISSN: 0169-2607; 1872-7565; 1872-7565
• DOI: 10.1016/j.cmpb.2022.106890

•We presented a new paper-ECG digitization algorithm, converting an ECG image to signal.•The digitized ECG was in substantial agreement with the digitally recorded ECG.•The disagreement was due to differences in simultaneous vs. asynchronous ECG leads.•We provided open-source Python code that will facilitate further development of the tool. We aimed to develop and validate an open-source code ECG-digitizing tool and assess agreements of ECG measurements across three types of median beats, comprised of digitally recorded simultaneous and asynchronous ECG leads and digitized asynchronous ECG leads. We used the data of clinical studies participants (n = 230; mean age 30±15 y; 25% female; 52% had the cardiovascular disease) with available both digitally recorded and printed on paper and then scanned ECGs, split into development (n = 150) and validation (n = 80) datasets. The agreement between ECG and VCG measurements on the digitally recorded time-coherent median beat, representative asynchronous digitized, and digitally recorded beats was assessed by Bland-Altman analysis. The sample-per-sample comparison of digitally recorded and digitized signals showed a very high correlation (0.977), a small mean difference (9.3 µV), and root mean squared error (25.9 µV). Agreement between digitally recorded and digitized representative beat was high [area spatial ventricular gradient (SVG) elevation bias 2.5(95% limits of agreement [LOA] -7.9–13.0)°; precision 96.8%; inter-class correlation [ICC] 0.988; Lin's concordance coefficient ρc 0.97(95% confidence interval [CI] 0.95–0.98)]. Agreement between digitally recorded asynchronous and time-coherent median beats was moderate for area-based VCG metrics (spatial QRS-T angle bias 1.4(95%LOA -33.2–30.3)°; precision 94.8%; ICC 0.95; Lin's concordance coefficient ρc 0.90(95%CI 0.82–0.95)]. We developed and validated an open-source software tool for paper-ECG digitization. Asynchronous ECG leads are the primary source of disagreement in measurements on digitally recorded and digitized ECGs.