Correlation between cardiac cavity size on non-gated chest computed tomography and cardiac ejection fraction on cardiac magnetic resonance imaging

Masaru Shiigai, Jonathan D. Dodd, Suhny Abbara, Conrad Wittram, Tsunehiro Yasuda

Research output: Contribution to journalArticle

Abstract

Background: We evaluated the correlation between linear measurements of the short-axis diameter (SA) of the left ventricle (LV) and right ventricle (RV), as determined by non-gated chest computed tomography (CT), and the ejection fraction (EF) measured by cardiac magnetic resonance imaging system (MRI). Methods: In this retrospective study, we examined the medical records of 41 patients (age [± SD], 58.7 ± 15.8 years) who underwent both cardiac MRI and non-gated chest CT. Patients with cardiac events between the tests were excluded. MRI-EF was calculated using the modified Simpson's method. The CT-LV short axis (CT-LSA) and CT-RV short axis (CT-RSA) were measured at the longest diameter at the midventricular level on axial images. These measurements were then examined for a correlation with MRI-EF (Spearman test). For more detailed analysis, patients were divided into 3 groups (MRI-EF% ≥55%, 40-55%, and <40%). Comparison of CT-short axis among the MRI-EF groups was performed by ANOVA followed by the Scheffe test. Results: The mean time interval (± SD) between the MRI and CT examinations was 176 ± 263 days. The overall correlation between MRI-LVEF and CT-LSA was good (r= -0.59, p<0.05). Sub-group analysis revealed that the strongest correlation was present for patients with an MRI-LVEF <40% (r = -0.83, p<0.05). According to the ANOVA test CT-LSA was significantly longer in patients with an LVEF <40% than in those with an LVEF of 40-55% or >55% (p<0.05). Receiver operating curve (ROC) analysis showed that when the CT-LSA was 4.8 cm, the specificity and sensitivity for detecting abnormal EF were 682% and 73.7%, respectively, which suggests that in patients with a CT-LSA longer than 4.8 cm, poor LV function should be suspected. No significant correlations between MRI-RVEF and CT-RSA were observed. Conclusions: Non-gated chest CT-LSA measurements are strongly correlated with LVEF on MRI. Patients with abnormal CT-LSA on non-gated chest CT would therefore provide clinically unexpected information in patients with LV heart failure.

Original languageEnglish (US)
Pages (from-to)283-290
Number of pages8
JournalJournal of the Medical Society of Toho University
Volume56
Issue number4
Publication statusPublished - Jul 2009

    Fingerprint

Keywords

  • Cardiac cavity size
  • Cardiac function
  • Computed tomography
  • Ejection fraction
  • Magnetic resonance imaging

ASJC Scopus subject areas

  • Medicine(all)

Cite this