UCSF

Background

Efficiency of coronary CT angiography (CCTA) in clinical practice depends on precise reporting and accurate result interpretation.

Objective

We sought to assess referring clinicians' understanding of patients' coronary artery disease (CAD) severity and to compare satisfactions of the free-form impression (FFI) with satisfactions of the structured impression (SI) section of CCTA reports.

Materials and methods

Fifty clinical CCTA reports from May 2011 to April 2012 were retrospectively selected (25 FFI and 25 SI), to include cases with the entire spectrum of CAD (6 categories encompassing normal, minimal, mild, moderate, severe stenosis, and occlusion). A survey containing only randomized blinded impressions was distributed to 4 cardiologists and 2 cardiac imaging specialists. Clinician interpretation was examined regarding (Q1) worst stenosis severity, (Q2) number of vessels with significant stenosis, and (Q3) the presence of nonevaluable segments. Agreement proportions and Cohen's kappa were evaluated between FFI versus SI. Satisfactions were measured with respect to content, clarity, and clinical effectiveness.

Results

Q1 agreement was excellent for both FFI and SI (by 6 categories: 80% versus 85%; P > .05; kappa: 0.87 versus 0.89; by no CAD versus nonsignificant versus significant CAD: 99% versus 97%; P > .05; kappa: 0.99 versus 0.94). Q2 agreement improved from fair to moderate (53% versus 68%; P = .04; kappa 0.31 versus 0.52). Q3 agreement was moderate (90% versus 87%; P > .05; kappa 0.57 versus 0.58). Satisfactions with impressions were high and similar for FFI and SI for clinicians.

Conclusion

Structured impressions were shown to improve result interpretation agreement from fair to moderate with regard to the number of vessels with significant stenosis.