DOI: 10.1161/circ.148.suppl_1.18102 ISSN: 0009-7322

Abstract 18102: Impact of Regional Cardiac Wall Motion Abnormalities on Ejection Fraction and Mortality: Large Data Enabled by Natural Language Processing

Albert J Rogers, Neal K Bhatia, James Tooley, Vyom Thakkar, JUSTIN XU, Rayan Ansari, Jessica Torres, Jagteshwar Tung, Mahmood Alhusseini, Paul L Clopton, Reza Sameni, Gari Clifford, Euan A Ashley, Marco V Perez, Matei Zaharia, Sanjiv M Narayan
  • Physiology (medical)
  • Cardiology and Cardiovascular Medicine

Background: Cardiac wall motion abnormalities (WMA) independently predict mortality and other adverse events beyond ejection fraction. The impact on mortality and ejection fraction has not been studied in large cohorts with long term follow up.

Aim: We aimed to describe the relationship of specific regional cardiac WMAs with ejection fraction and mortality in a large referral academic center.

Methods: We collected echocardiograms, demographics, and follow-up data from 35,210 patients from Stanford University. Specific regional abnormalities were determined by natural language processing (NLP) of clinical reports. WMAs were stratified across 5 anatomic regions and compared. Ejection fraction effect were compared by the Mann-Whitney U test and visualized with violin plots. Mortality between regions was assessed with Cox proportional hazards model and visualized with Kaplan-Meier plot.

Results: In the Stanford University cohort, (62.9±16.9 years, 47% female, 58% white), regional WMAs were associated variable reductions in EF with apical, anterior, and septal WMAs resulting in an EF 41.6-42.8 +/- 12.8%, which was greater than lateral (45.4 +/- 10.2%) and inferior/posterior (46.3 +/- 12.1) (p< .01 for all comparisons). Distributions for inferior/posterior and lateral were more left-skewed and apex/septal were more bimodal ( Panel A ). In follow-up, having at least one WMA increased the risk of overall mortality with a hazard ratio of 1.8 (CI: 1.61-2.00) over a > 4 year period ( Panel B ).

Conclusion: In a very large population of echocardiograms performed in an academic referral center, wall motion abnormalities in different areas were associated with variable effects on ejection fraction and long-term mortality. Natural language processing enabled analysis of a large database of echocardiogram reports.

More from our Archive