Rofo 2020; 192(S 01): S24-S25
DOI: 10.1055/s-0040-1703178
Vortrag (Wissenschaft)
Herzdiagnostik/Gefäßdiagnostik
© Georg Thieme Verlag KG Stuttgart · New York

A prototype self-navigated MRI 3D ‘whole heart’ sequence for non-enhanced aortic root measurement in transcatheter aortic valve intervention: comparison to cadiac CT

M Pamminger
1  Medizinische Universität Innsbruck, Universitätsklinik für Radiologie, Innsbruck
,
C Kranewitter
1  Medizinische Universität Innsbruck, Universitätsklinik für Radiologie, Innsbruck
,
C Kremser
1  Medizinische Universität Innsbruck, Universitätsklinik für Radiologie, Innsbruck
,
B Henninger
1  Medizinische Universität Innsbruck, Universitätsklinik für Radiologie, Innsbruck
,
G Reiter
2  Siemens Healthcare Diagnostics GmbH, Forschung und Entwicklung, Graz
,
D Piccini
3  Siemens Healthcare Switzerland, Forschung und Entwicklung, Renens
,
G Klug
4  Medizinische Universität Innsbruck, Universitätsklinik für Innere Medizin III (Kardiologie), Innsbruck
,
A Mayr
1  Medizinische Universität Innsbruck, Universitätsklinik für Radiologie, Innsbruck
› Author Affiliations
Further Information

Publication History

Publication Date:
21 April 2020 (online)

 

Zielsetzung To evaluate image quality, inter-observer reliability and diagnostic accuracy of self-navigated noncontrast 3D whole-heart magnetic resonance angiography (MRA) for transcatheter aortic valve intervention (TAVI) evaluation in comparison to standardized contrast-enhanced computed tomography angiography (CTA).

Material und Methoden Whole-heart 1.5 T MRA was performed in 33 patients (aged 84 years [IQR 79-86], 48% male) for aortic root sizing and measurements of coronary ostia heights. A subgroup of 18 (55%) patients underwent additional CTA as gold standard for TAVI measurements. Image quality was assessed by a 4-point Likert scale, continuous MRA and CTA measurements were compared with regression and Bland-Altman analysis, valve sizing by kappa statistics.

Ergebnisse Median image quality of MRA as rated by two observers according was 1.5 [IQR 1.5-2.5]. In 4 patients (12%) one coronary ostium each (right coronary artery 3, left main artery 1) was not clearly definable on MRA. Inter-observer correlation was substantial to excellent (r=0.61 to 0.92) with a bias of 19 mm2 for annulus area (lower limit of agreement -59 mm2, upper limit of agreement 98 mm2; p=0.009). Aortic root and ostia height measurements by MRA and CTA showed substantial to excellent correlation (r=0.65 to 0.90) with no significant bias (all p ≥0.333). Mean annulus area for MRA was 414±71 mm2 and for CTA 422±80 mm2 (r=0.9) with a bias of -8 mm2 (lower limit of agreement -79 mm2, upper limit of agreement -62 mm2; p=0.333). Regarding prosthetic valve sizing there was complete consistency between MRA and CTA-based decisions (?=1).

Schlußfolgerungen Self-navigated noncontrast 3D whole-heart MRA enables reliable aortic root TAVI measurements without significant difference to standardized CTA. Prosthesis sizing by MRA measurements would completely match to CTA-based choice. However, in some cases coronary ostia may be difficult to define.