gms | German Medical Science

33. Internationale Konferenz für Elektrokardiographie

Internationale Konferenz für Elektrokardiographie

Automated ECG Classification of RVH and LVH in Neonatal Population

Meeting Abstract

  • corresponding author presenting/speaker S. Zhou - Advanced Algorithm Research Center, Philips Medical, Thousand Oaks, USA
  • J. Liebman - Case Western Reserve Univ., Cleveland, USA
  • A. Dubin - Stanford Univ., Palo Alto, USA
  • K. Sun - Shanghai Jiaotong Univ., Shanghai, China
  • F. Li - Shanghai Jiaotong Univ., Shanghai, China
  • Y. Zhou - Shanghai Jiaotong Univ., Shanghai, China
  • R. Gregg - Advanced Algorithm Research Center, Philips Medical, Thousand Oaks, USA
  • E. Helfenbein - Advanced Algorithm Research Center, Philips Medical, Thousand Oaks, USA
  • J. Lindauer - Advanced Algorithm Research Center, Philips Medical, Thousand Oaks, USA

33rd International Congress on Electrocardiology. Cologne, 28.06.-01.07.2006. Düsseldorf, Köln: German Medical Science; 2007. Doc06ice044

Die elektronische Version dieses Artikels ist vollständig und ist verfügbar unter: http://www.egms.de/de/meetings/ice2006/06ice044.shtml

Veröffentlicht: 8. Februar 2007

© 2007 Zhou et al.
Dieser Artikel ist ein Open Access-Artikel und steht unter den Creative Commons Lizenzbedingungen (http://creativecommons.org/licenses/by-nc-nd/3.0/deed.de). Er darf vervielf&aauml;ltigt, verbreitet und &oauml;ffentlich zug&aauml;nglich gemacht werden, vorausgesetzt dass Autor und Quelle genannt werden.


Gliederung

Text

Interpretation of the neonatal electrocardiogram (ECG) is very complex for many reasons including the rapid ECG changes, proximity effect and the large heart in relation to the size of the chest. Interpretation of both RVH and LVH in the first few days of life remains a challenge to pediatric cardiologists. An appropriate computer ECG analysis program with automatic interpretation in the first days of life has yet to be developed although there are obvious advantages. This study is to report our experience in development of an automatic computer interpretation program for newborn infants. The study group includes 501 ECGs recorded from consecutive neonatal patients with age range from 1 to 72 hours at the Stanford Packard Children’s Hospital, Palo Alto, California. Pediatric cardiologist readings served as the gold standard in validation testing. The 242 ECGs classified as RVH or LVH by pediatric cardiologists were divided randomly as training and testing sets. The 188 ECGs in the control group were recorded from healthy newborn babies with age range from 1 to 72 hours in Shanghai Children’s Medical Center, Shanghai, China. All ECG's were recorded using Philips PWXLi with 500Hz sampling rate using the standard 12 leads plus additional leads V3R, V4R and V7. ECGs were measured using the latest Philips measurement program (PH10) with high measurement accuracy. In addition to commonly used voltage criteria and normal limits, measurements from the scalar ECG synthesized vector were added for automated RVH and LVH classification. Tested against the existing newborn infant ECG database, we obtained a sensitivity of 55% and a specificity of 94% in RVH classification, a sensitivity of 50% and a specificity of 95% in LVH classification. We observed the ECGs in newborn infants presented unusually high measurement variation. A single voltage criterion is not reliable. The vector measurements help to improve the interpretation accuracy.