Daniel Verdini1, Ashley M Lee1, Anand M Prabhakar1, Suhny Abbara1, Brian Ghoshhajra2. 1. Harvard Medical School, Boston, Massachusetts; Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts. 2. Harvard Medical School, Boston, Massachusetts; Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts. Electronic address: bghoshhajra@mgh.harvard.edu.
Abstract
PURPOSE: Routine chest CT and cardiac CT angiography (CTA) both image the heart, albeit with different precision and intent. The aim of this study was to evaluate the diagnostic ability of radiologists with different levels of cardiac training to identify cardiac findings on chest CT without electrocardiographic gating compared with a reference standard of electrocardiographically gated cardiac CTA. METHODS: Electrocardiographically gated cardiac CT angiographic studies performed between January 2005 to January 2010 in patients with routine chest CT within six months were retrospectively identified. Fourteen radiologists at four stages of training (stage 1, residents with no cardiac training [n = 4]; stage 2, residents who had completed at least one dedicated rotation of cardiac imaging [n = 3]; stage 3, radiologists without cardiac training [n = 3]; and stage 4, radiologists with formal cardiac fellowship training [n = 4]) performed blinded, anonymized cardiac readings of chest CT images. Findings were categorized (coronary arterial, noncoronary vessel, cardiac chamber, myocardial, pericardial, and valve findings) with cardiac CTA as a reference standard. RESULTS: Overall, 140 cardiac CT angiographic findings were reported in 63 of 77 patients. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of nongated CTA were 43.6%, 96.8%, 83.0%, 81.6%, and 81.8%, respectively, for all readers. Increasing training was associated with higher sensitivity (30.3%, 35.7%, 45.7%, and 61.2% from stages 1 to 4) but similar specificity (96.4%, 96.7%, 96.3%, and 97.6% from stages 1 to 4). Frequently missed findings categories were coronary arterial, myocardial, and cardiac chamber findings. CONCLUSIONS: Increasing cardiac imaging training correlates with increased sensitivity and stable specificity to detect cardiac findings on routine chest CT without electrocardiographic gating. Cardiac findings should be noted on chest CT when observed, and cardiac training should be encouraged.
PURPOSE: Routine chest CT and cardiac CT angiography (CTA) both image the heart, albeit with different precision and intent. The aim of this study was to evaluate the diagnostic ability of radiologists with different levels of cardiac training to identify cardiac findings on chest CT without electrocardiographic gating compared with a reference standard of electrocardiographically gated cardiac CTA. METHODS: Electrocardiographically gated cardiac CT angiographic studies performed between January 2005 to January 2010 in patients with routine chest CT within six months were retrospectively identified. Fourteen radiologists at four stages of training (stage 1, residents with no cardiac training [n = 4]; stage 2, residents who had completed at least one dedicated rotation of cardiac imaging [n = 3]; stage 3, radiologists without cardiac training [n = 3]; and stage 4, radiologists with formal cardiac fellowship training [n = 4]) performed blinded, anonymized cardiac readings of chest CT images. Findings were categorized (coronary arterial, noncoronary vessel, cardiac chamber, myocardial, pericardial, and valve findings) with cardiac CTA as a reference standard. RESULTS: Overall, 140 cardiac CT angiographic findings were reported in 63 of 77 patients. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of nongated CTA were 43.6%, 96.8%, 83.0%, 81.6%, and 81.8%, respectively, for all readers. Increasing training was associated with higher sensitivity (30.3%, 35.7%, 45.7%, and 61.2% from stages 1 to 4) but similar specificity (96.4%, 96.7%, 96.3%, and 97.6% from stages 1 to 4). Frequently missed findings categories were coronary arterial, myocardial, and cardiac chamber findings. CONCLUSIONS: Increasing cardiac imaging training correlates with increased sensitivity and stable specificity to detect cardiac findings on routine chest CT without electrocardiographic gating. Cardiac findings should be noted on chest CT when observed, and cardiac training should be encouraged.