Determining Whole Lung Perfusion Using Hyperpolarized Xe129 MRI

Determining Whole Lung Perfusion Using Hyperpolarized Xe-129 MRI Ian Gerard Department of Medical Biophysics University Of Western Ontario Six Week Project

MR Imaging Basics B 1 Sample RF Coil Gradient Coil Set Magnet z z m x y

Hyperpolarized Media Isotope Spin * 107 rad s/T P (ppm) 1 H 1/2 1/2 -1/2 26. 75 20. 37 6. 73 7. 40 5. 11 3. 89 1. 28 1. 41 3 He 13 C 129 Xe (For T = 300 K and B 0 = 3 T ) In General After Hyper-polarization Method

Gas Exchange Model D ¶ 2 C ( r , t ) Xe Xe Xe ¶r 2 - ¶C ( r , t ) ¶t =0 Xe Xe Xe Fig 1: Alveolar-Capillary model of Mansson [2]. Va is the alveolar volume measured at standard temperature and pressure. ra is the radius of the alveolus. Lt and Lc are the compartment lengths for tissue and capillary. C(r, t) is the xenon concentration in each respective compartment as a function of distance and time. F is the rate of pulmonary flow. Fin=Fout. -t æ TTR ö S (t ) = S 0 ç 1 - e ÷ + S 1 t è ø S 1 = αλpl (1 – H)F

Methods Medical Air & O 2 Ventilator 3 He Or 129 Xe 3 T MRI Scanner Heater Rb Reservoir Polarization Cell Gas Cylinder 89% 4 He 10% N 2 1% Xe 794. 8 nm Circular Polarizer Magnet Polarized 129 Xe Out Polarized 129 Xe and 3 He Valve Assembly Acid Port Ventilator

Pilot Results: Xenon

Pulse Sequence Chemically Selective Saturation Recovery (CSSR) 90 o RF Pulse 90 o ~193 ppm f Loop N. Abdeen et al (2006) t Dissolved Gas Phase Maximized Minimized

Spectroscopy After Applying Pulse Sequence Mansson Model: Valid for all time.

Tissue Spectroscopy and Model Fit Mansson Model:

Results: Representative Figure Linear fit applied to long time values Slope and intercept values are S 1 and S 0 respectively and used to calculate lung perfusion

Calculated Lung Perfusion Typical Perfusion value from control rat 1. 2 ± 0. 2 Mansson et al. Rat Number Rat 1 Rat 2 Rat 3 PIP (cm. H 20) Mean Perfusion (ml s-1/ml) ± Std 7 ± 0. 1 0. 74 ± 0. 05 11 ± 0. 1 1. 01 ± 0. 01 12 ± 0. 1 1. 17 ± 0. 03 17 ± 0. 1 1. 29 ± 0. 04 7 ± 0. 1 0. 84 ± 0. 05 12 ± 0. 1 1. 07 ± 0. 02 17 ± 0. 1 1. 36 ± 0. 05 11 ± 0. 1 1. 14 ± 0. 15 13 ± 0. 1 1. 34 ± 0. 06 15 ± 0. 1 1. 40 ± 0. 05

Perfusion and PIP • Appears to be some relationship between PIP value and Perfusion

Discussion • Values agree with control perfusion from Mansson et al. • Something interesting may be happening at low pressure. • There may be a relationship between Perfusion and PIP • Hematocrit and typical perfusion values used were for Wistar rats while Sprague. Dawley rats were actually used but assumed equal for the project

Conclusions and Future Work • Successfully able to estimate whole lung perfusion using Mansson model and other minor assumptions • Investigate the effects of low PIP on perfusion • Next Step: Create a perfusion map determining lung perfusion on a pixel by pixel basis

Acknowledgements Heather Cadieux Eugene Wong Jake Van Dyk Funding: Giles Santyr Matthew Fox Alexei Ouriadov Ryan Kraayvanger William Dominguez Wilfred Lam Peggy Xu Adam Farag Marcus Couch Lynda Mc. Caig Ruud Veldhuizen Jim Lewis