RESUMO
INTRODUCTION: Incorporating coronary computed tomography angiography (CTA) in the hospital workup for suspected acute coronary syndrome requires appropriate skills for interpreting this imaging test. Radiologists' skills can affect the interobserver agreement in evaluating these studies. OBJECTIVE: To determine the interobserver agreement according to radiologists' experience in the interpretation of coronary CTA studies done in patients who present at the emergency department with acute chest pain and low-to-intermediate probability of acute coronary syndrome. MATERIALS AND METHODS: We studied the interobserver agreement in the urgent evaluation of coronary CTA studies in which CAD-RADS was used to register the findings. We created pairs of observers among a total of 8 assessors (4 attending radiologists and 4 radiology residents). We used the kappa coefficient to estimate the overall concordance and the concordance between subgroups according to their experience. RESULTS: The agreement was substantial between experienced radiologists and residents (k=0.627; 95%CI: 0.436-0.826) as well as between all the pairs of observers (k=0.661; 95%CI: 0.506-0.823) for all the CAD-RADS together. The degree of agreement within the group of experienced radiologists was greater than that within the group of residents in all the analyses. The agreement was excellent for the overall CAD-RADS (k=0.950; 95% CI: 0.896-1) and for CAD-RADS ≥ 4 (k=1); the agreement was lower for CAD-RADS ≥ 3 (k=0.754; 95% CI: 0.246-1.255). The agreement for the residents for these categories was k=0.623, k=0.596, and k=0.473, respectively. CONCLUSION: The agreement among attending radiologists regarding the assessment of urgent coronary CTA studies is excellent. The agreement is lower when residents are paired with attending radiologists. These findings should be taken into consideration when implementing coronary CTA in emergency departments and in the organisation of radiological staff for interpreting and reporting this imaging test.
