Rofo 2012; 184 - TNE11
DOI: 10.1055/s-0031-1300912

Respiratory dependent stroke volume changes assessed by real time MR velocity mapping at 3 Tesla – A validation study

H Körperich 1, P Barth 1, J Gieseke 2, H Esdorn 1, A Peterschröder 1, G Uges 3, D Kececioglu 3, W Burchert 1, KT Laser 3
  • 1Herz- und Diabeteszentrum NRW, Institut für Radiologie, Nuklearmedizin und Molekulare Bildgebung, Bad Oeynhausen
  • 2Philips Medical Systems, MR Clinical Science, DA Best
  • 3Herz- und Diabeteszentrum NRW, Klinik für Angeborene Herzfehler/Kinderkardiologie, Bad Oeynhausen

Background:

Respiration has a strong impact on stroke volumes (SV) in great thoracic vessels [1]. Conventional MR techniques for quantitative flow (QF) measurements are unsuitable to differentiate between different respiration phases due to its averaging character. Real time flow velocity mapping may provide unique hemodynamic information that is not available with other methods.

Methods:

QF measurements were performed on a 3T-TX MR scanner (Philips) using a non-triggered, free-breathing, real-time phase-contrast EPI sequence (RT-QF). In-plane resolution was 2.7×2.7mm2, slice thickness 6mm. To keep temporal resolution below 25ms a SENSE factor of 4 combined with half-Fourier was applied. Data were recorded during a 12s period resulting in 500 dynamic frames. Real time data were compared with those obtained by a validated conventional QF sequence (Ref-QF). Respiratory dependent QF measurements were performed in the ascending aorta (AAo) and vena cava superior (VCS) in 16 healthy kids (9 male, age=12.9±3.3y). Breathing curve was divided in 4 sections: expiration, breath in, inspiration and breath out.

Results:

Using Bland-Altman statistics, high agreement was observed comparing “mean” SV provided by RT-QF and Ref-QF in AAo (mean±SD; limits-of-agreement: 3.4±9.7%; -16.0 to 22.8), whereas a higher overestimation was found for RT-QF in VCS (11.0±8.9%; -6.8 to 28.9). Pearson correlation was >0.9.

Assigning relative RT-QF SVs (normalized to mean SV) to the different respiration phases revealed a stronger dependency of VCS compared to AAo. This may be related to the venous vessel wall composition which is less bulky in comparison to arterial walls. In detail (mean±SD; expiration, breath in, inspiration and breath out): VCS (-7.7±10.5[%]; -6.6±8.3[%]; 9.4±10.8[%]; 4.9±7.5 [%]) and AAo (-6.7±4.1[%]; 0.0±2.6[%]; 6.1±3.1[%]; 0.6±2.7 [%]).

ANOVA demonstrates that all volumes belonging to the various respiratory phases are statistically different of each other for VCS (P<0.05) except the relation “inspiration/breath out” (P=0.19). Addressing AAo all respiratory dependent volumes are highly statistically different of each other (P<0.05) except the relation “breath in/breath out” (P=0.57).

Discussion and Conclusion:

The impact of respiration on blood flow pattern in thoracic vessels must be considered. It is unknown whether and to what extent flow in the great thoracic vessels is attenuated by breath-holding if fast flow mapping methods are applied. As opposed to conventional techniques with real time MR velocity mapping, no respiratory gating or ECG triggering is needed. Furthermore, cardiac arrhythmia should no longer preclude safe and reliable flow measurements.

Reference:

[1] Körperich H et al. (2004). Flow volume and shunt quantification in pediatric congenital heart disease by real-time magnetic resonance velocity mapping: a validation study. Circulation; 109: 1987–93.