Purpose: To compare contrast-enhanced CT angiography (CTA) and gadolinium-enhanced MR angiography (MRA) for the detection of subsegmental-sized pulmonary emboli in a pig model.
Material and methods: In 5 anesthetized pigs, 3-mm diameter embolic materials made of Konjac, a semisolid food, were introduced through the internal jugular vein into pulmonary arteries. After embolization, CTA and MRA images were obtained. Respiration was suspended during CTA and MRA image acquisition. Two readers reviewed the CTA and MRA images to detect emboli. The pigs were sacrificed, and sliced specimens of inflated lung served as the gold standard.
Results: Thirty-six emboli were detected within peripheral arteries. The sensitivity (and 95% confidence intervals) of CTA for the two readers were 57% (39-74%) and 66% (48-81%), and 88% (69-98%) and 92% (74-94%) for MRA. The specificity of CTA was 95% (91-97%) and 98% (96-99%), and that of MRA was 85% (74-93%) and 90% (80-96%). Interobserver agreement was higher for MRA (kappa 0.898) than CTA (kappa 0.574).
Conclusion: For the detection of subsegmental pulmonary emboli, MRA was superior to CTA, with a higher sensitivity and interobserver agreement by demonstrating perfusion deficits.