MRI of the lung using the PROPELLER technique: Artifact reduction, better image quality and improved nodule detection. Issue 4 (April 2016)

• Record Type: Journal Article
• Authors: Meier-Schroers, Michael; Kukuk, Guido; Homsi, Rami; Skowasch, Dirk; Schild, Hans Heinz; Thomas, Daniel
• Is Part Of: European journal of radiology. Volume 85:Issue 4(2016)
• Page Start: 707
• Page End: 713
• Publication Date: 2016-04
• Subjects: Lung -- MRI -- PROPELLER -- MultiVane -- Artifacts; Medical radiology -- Periodicals; Radiology -- Periodicals; Radiologie médicale -- Périodiques; Medical radiology; Periodicals; 616.075705
• DOI: 10.1016/j.ejrad.2015.12.016
• Languages: English
• ISSNs: 0720-048X

Abstract

Highlights: T2w MultiVane images can be aqcuired during free breathing. The T2w MultiVane sequence shows great artifact reduction and better image quality than FSE. The MultiVane technique could be the new standard in T2w pulmonary MR imaging. Abstract: Purpose: To evaluate the benefit of the PROPELLER technique (Periodically Rotated Overlapping ParallEL Lines with Enhanced Reconstruction, MultiVane, MV) for MR imaging of the lung. Materials and Methods: 30 Participants of a lung cancer screening program were recruited for the comparison of T2-MV and T2-Fast Spin Echo (FSE) sequences at 1.5T. Two readers evaluated artifacts, image quality, and pulmonary lesions. Artifacts and image quality were rated using a four-point scale. Lesion detection was correlated to low-dose computed tomography (CT). Wilcoxon rank-test for ratings of artifacts and image quality, sensitivity and specificity values for lesion detection, and Cohen's kappa for inter-rater agreement were used. Results: The MV sequence showed less pulsation and motion artifacts, and higher image quality ( p = 0.001 for R1, p = 0.002 for R2) than FSE ( p < 0.001 for both readers, R1 and R2). Inter-rater agreement was excellent for lesion detection (0.84–0.95) and good to excellent for artifacts and image quality (0.66–0.84). 17 patients had lesions <8 mm, and 7 had lesions >8 mm as seen on CT. For R1 and R2, the MV sequence allowed for higher detection rates of pulmonary lesions <8 mm with a sensitivity of 56% (R1) and 59% (R2); the FSE sequence achieved 50% (R1) and 53% (R2). Specificity was also higher for MV with 94% (R1) and 83% (R2) compared to 78% (R1) and 76% (R2). Lesions >8 mm were detected with a sensitivity of 100% by both readers on both MV and FSE images. For both readers, specificity for larger lesions was higher on MV images with 100% compared to 96%. Conclusion: The superior image quality and the very robust artifact reduction make MV a promising technique for MRI of the lung compared to FSE, especially since it is not requiring breathholds. Moreover, MV allows for improved lesion detection.