The impact of training on diagnostic accuracy with computed tomography coronary angiography

Abstract

The aim of this study is to assess the image quality and diagnostic accuracy of computed tomography (CT) coronary angiography (CTCA) in different hospital settings with the same trained team. Four hundred patients were consecutively enrolled for CTCA in a large academic hospital (Group 1; Sensation 64 Cardiac, Siemens - Iomeprol 400, Bracco; 200 patients) and in a small local hospital (Group 2; VCT, GE Healthcare - Iodixanol 320, GE Healthcare; 200 patients). All patients were enrolled for suspected coronary artery disease (CAD) and patients with stents or who had previously undergone coronary bypass were excluded. Scan protocols (retrospectively ECG-gated; no dose reduction modulation applied) were performed in accordance with standards reported in the international literature with the best solution available on site. Image quality was assessed in each coronary segment with a 4-point Likert scale: 0, not assessable; 1, low; 2, average; 3, good. Diagnostic accuracy was calculated against conventional coronary angiography with a threshold of at least 50% for significant stenosis. There was no significant difference between demographics, BMI, prevalence of obstructive CAD, calcium score and heart rate between the two populations. The average image quality was 2.83 ± 0.37 for Group 1 and 2.86 ± 0.31 for Group 2 (P > 0.05). Per-segment sensitivity, specificity, positive and negative predictive values were 92.6% (87-95), 97.9% (97-98), 75.9% (69-81) and 99.5% (99-99), respectively, for Group 1, and 90.4% (85-93), 98.6% (98-99), 84.2% (78-88) and 99.2% (98-99), respectively, for Group 2 (P > 0.05). There is no significant difference in image quality and diagnostic accuracy of CTCA when the investigation is performed by the same properly trained team. CTCA is a robust imaging modality for the detection of coronary artery stenosis.