• M J O'Connell and J G Murray.
• Department of Radiology, Mater Misericordiae Hospital, Dublin 7, Ireland. martinoconnell@hotmail.com
• Clin Radiol. 2001 Jul 1; 56 (7): 545-9.

AimTo assess routine image subtraction in 3D gadopentate dimeglumine (Gd-DTPA)-enhanced magnetic resonance (MR) angiography of the thoracic aorta.Materials And MethodsThis was a prospective study of 22 consecutive patients referred for magnetic resonance imaging (MRI) of the thoracic aorta. All patients had 3D MR aortography (TR/TE/FA; 5/2 ms/25 degrees ) performed before and after bolus intravenous injection of Gd-DTPA. The Gd-DTPA enhanced and unenhanced data sets were subtracted and maximum intensity projections (MIP) projections of the thoracic aorta were performed. The standard unsubtracted MIP images were initially evaluated. These were then reviewed together with the subtracted images to assess for additional diagnostic information. Signal to noise ratios (SNR) and contrast to noise ratios (CNR) were measured.ResultsIn four cases there was mild image degradation due to patient movement. In no case did subtraction alter the diagnosis. The mean SNR in the unsubtracted MIP images was 10.8 +/- 4.0 (median 11.1) and on the subtracted images was 21.2 +/- 9.9 (median 20.7;P < 0.0001). The mean aorta-to-mediastinal fat CNR was 3.9 +/- 2.8 (median 3.9) on the unsubtracted images and 15.0 +/- 10.6 (median 13) on the subtracted images (P < 0.0001). The mean aorta-to-vertebral body CNR was 5.2 +/- 3.1 (median 4.4) on the unsubtracted images and 15.1 +/- 9.3 on the subtracted images (P < 0.0001).ConclusionImage subtraction significantly improved both the SNR and CNR, but did not alter the final diagnosis, and does not appear warranted in routine practice.

22 keywords...

hide…