Medizinische Universität Graz Austria/Österreich - Forschungsportal - Medical University of Graz
Gewählte Publikation:
Erhart, M.
Comparison of different strategies for cardiac magneticresonance imaging short-axis based evaluation of right ventricular function
Humanmedizin; [ Diplomarbeit ] Graz Medical University; 2017. pp. 64
[OPEN ACCESS]
FullText
- Autor*innen der Med Uni Graz:
- Betreuer*innen:
- Fuchsjäger Michael
- Altmetrics:
- Abstract:
- Purpose: Cardiac magnetic resonance (CMR) imaging enables the quantification of right ventricular (RV) function from segmentation of stacks of cine short-axis (SA) images. The asymmetric RV shape and the skew motion of the tricuspid valve complicate the definition of RV base plane. The purpose of the present study was to derive and compare reference ranges for RV function parameters employing different base plane definition algorithms and to investigate the consistency of the derived RV stroke volumes (RVSV) with left ventricular stroke volumes (LVSV). Material and Methods: ECG-gated 1.5T CMR cine imaging was performed on forty healthy volunteers. LVSV was evaluated automatically, RV volumes (enddiastolic volume and endsystolic volume), RVSV and the RV ejection fraction where derived from manual segmentation of end-diastolic and end-systolic SA images employing standard software. RV base plane was defined from SA images only (merged SA and SA methods) and from defining junction points of the tricuspid valve in 4-chamber view (4CH) and RV 2-chamber view (RV 2CH), respectively (4CH, 4CH & RV 2CH and optimized 4CH & RV 2CH methods). Results were compared pair-wise employing correlation and Bland-Altman anal-ysis as well as t-test. Results: RV function parameters demonstrated significant biases and large 95% limits of agreement for the comparisons between the merged SA and the SA, the SA and the 4CH, the 4CH and the 4CH & RV 2CH, and the 4CH & RV 2CH and the optimized 4CH & RV 2CH evaluation method. LVSV correlated similarly with RVSV of all evaluation methods (r = 0.82 – 0.88), but only the evaluation methods employing two long-axis views did not significantly underestimate RVSV (LVSV = 98 ± 23 ml; 4CH & RV 2CH method: RVSV = 100 ± 30 ml, p = 0.43; optimized 4CH & RV 2CH method: RVSV = 102 ± 28 ml, p = 0.06). Conclusion: All RV function parameters derived from stacks of cine SA images crucially depend on the type of RV base plane definition, whereby tricuspid valve plane modelling by two long-axis views provides the most consistent results to LV volumetric evaluation. Established RV reference ranges might serve as normal values for RV volumetric function indices in future studies.