Purpose: To evaluate if computed tomography (CT) coronary calcium scoring is needed after detection of coronary calcifications on conventional chest radiographs.
Materials and methods: One hundred and five patients (67 men; 57.2+/-12.8 years) with suspected coronary artery disease underwent conventional chest radiography and non-enhanced, retrospectively ECG-gated multislice spiral CT (MSCT) of the heart (4 mm x 2.5 mm, 120 kV, 133 mAs(eff.)). Chest radiographs were assessed independently by two radiologists. Detection of coronary calcifications was compared between both methods. Sensitivity, specificity, negative and positive predictive values, median, 25% and 75% percentiles for the detection of coronary calcifications were calculated. Receiver operating characteristics (ROC) analyses were computed.
Results: In 90 patients, MSCT revealed coronary calcifications. The mean coronary calcium score was 526.2 (0-4784.5). On chest radiographs, coronary calcifications were correctly detected in 46 (61) patients by observer 1 (observer 2). The corresponding sensitivity was 51.1% in observer 1 and 67.8% in observer 2. Median of detected coronary calcifications was 361.9 (426.4) for observer 1 (observer 2). Corresponding 25% und 75% percentiles were 109.6 (109.6) and 798.5 (898.5). The area under the ROC curve was 0.636 for observer 1 and 0.715 for observer 2. There was no correlation between image quality and the detection of coronary calcifications on plain film radiographs.
Conclusion: As coronary calcifications of various extents are inconsistently detected on plain chest radiographs, CT calcium scoring may not be omitted even if coronary artery calcifications were detected on conventional chest radiographs.