Abstract 1596: Optimization of Small Pulmonary Artery Visualization for Pulmonary Embolism Detection with Multidetector CT
AUTHORS:
S. Patel, MD, E.A. Kazerooni, MD, B.H. Gross, MD.
ABSTRACT:
PURPOSE: To compare the frequency of well-visualized pulmonary arteries by anatomic level on single and multidetector helical CT in patients studied for suspected pulmonary embolism.
METHODS AND MATERIALS: 60 patients undergoing CT for suspected pulmonary embolism were scanned with 3 different protocols, with 20 consective patients in each of three groups. The three protocols were 1) single detector CT at 3 mm collimation, table speed 3.9-4.8 mm/rotation; 2) multidetector helical CT at 2.5 mm collimation, table speed 15 mm/rotation; 3) 1.25 mm collimation, table speed 7.5 mm/rotation. Overlapping reconstructions were performed on all data sets. All studies were performed with 150 cc intravenous contrast administered at 4 cc/second and a 25 second scan delay. Two thoracic radiologists independently reviewed each CT scan on a computer workstation to determine if the main, lobar, segmental and subsegmental arteries were sufficiently visualized for pulmonary embolism detection.
RESULTS: All main, right and left pulmonary arteries were well-visualized by both readers using all 3 scan protocols. With protocols 1, 2 and 3, 93%, 90% and 95% of lobar arteries respectively were well-visualized by both readers. The differences were not statistically significant. Reader 1 identified 76%, 80% and 87% of segmental arteries (p = 0.01; x2 test) and 37%, 49% and 69% of subsegmental arteries (p < 0.0001); x2 test) as well-visualized with scan protocols 1, 2 and 3 respectively. Reader 2 identified 73%, 78% and 89% of segmental arteries (p = 0.0004; x2 test) and 35%, 44% and 68% of subsegmental arteries (p < 0.0001); x2 test) with scan protocols 1, 2 and 3. For both readers, 25% or more of the subsegmental arteries were not visualized in 90%, 90% and 40% of patients using protocols 1, 2 and 3 respectively. On average 26, 22 and 13 subsegmental arteries per patient (40 total subsegmental arteries per patient) and 5, 3 and 2 segmental arteries per patient (20 total subsegmental arteries per patient) were not visualized using scan protocols 1, 2 and 3 respectively. While the non-visualization of segmental and subsegmental arteries was most common with in-plane arteries of the lingula and right middle lobe with protocols 1 and 2, this was not a feature of protocol 3. The greatest factors in non-visualization for protocol 3 were patient-related (motion, bolus, lung abnormalities).
CONCLUSIONS: Multi-detector helical CT at 1.25 mm collimation with overlapping reconstructions significantly enhances the visualization of both segmental and subsegmental arteries for the evaluation of suspected pulmonary embolism. (EAK is the recipient of a research grant from GE Medical Systems.)
