Figure 1
Fast Quantitative 1 Constantinides ,
1 Maguire ,
19F
2 Swider ,
MRI: Optimized Imaging Strategies 3 Malandraki-Miller ,
4 Srinivas ,
3 Carr ,
1 Schneider
C. M. L. E. S. M. C. A. J. E. 1 3 Departments of Cardiovascular Medicine, and Physiology, Anatomy, Genetics, University of Oxford, Oxford, UK and the 4Department of Immunology, Radboud University, The Netherlands
Purpose-Introduction Implantation of stem cells (SCs) aims to provide tissue regeneration and cardiac functional improvements following ischemic injury. While the feasibility of SC therapy has been proven, its efficacy has yet to be demonstrated. Correspondingly, the visualization of implanted SCs to help define optimal therapy strategies has become a subject of intense research Over the past decade, 19F MRI has been used to image and track exogenously labeled nanoparticles (NP labels) in vivo [1, 2] While prior works attempted to optimize fluorine acquisitions, these had either focused on spectroscopy [3, 4] or used dedicated imaging sequences [5] We present an optimized, 19F MRI approach that enables a) ultrafast acquisitions with optimal signal-to-noise (SNR) and b) image-based quantification.
Aim: to optimize the MRI approach enabling a) ultrafast acquisitions with optimal signal-to-noise (SNR), b) image-based quantification, and c) postmortem murine applications
Methods Simulations: Spoiled gradient echo (SPGR), rapid acquisition with relaxation enhancement (RARE), fid/echo/balanced steady state free precession (SSFP) sequences were simulated [6], and optimized parametric SNR maps generated in MATLAB (Mathwords, USA) using typical 19F relaxation times and scan parameters. MRI/MRS Studies: Phantoms: Aqueous phantoms (1-100 mM), containing trifluoroacetic acid (TFA), perfluorocarbon (PFC) NP labels, and labeled cardiosphere-derived cells (CDC) were used to optimize radiofrequency (RF) coil responses at 9.4 T. A post-mortem study (C57BL/6) was conducted using the butterfly coil (19F: TR/TE=5.71-6.32/2.87-3.16 ms, flip angles=30-50°, NEX=1048-2096, FOV=50-60×50-60 mm2, slice thickness=35 mm, axial, BW=6 kHz, matrix=32×32 in a total of 3.11-7.4 min) where labelled cells were injected in the femoral area of the mouse hindlimb.
Figure 1: A. Parametric SNR plots using TFA T1=2.68 s and T2=2.37 s. B. 2D phantom SNR optimization using the birdcage coil (100 mM TFA), and (C) validation of quantification.
Figure 2: (A) 19F MRS in NPs and (B) labelled CDCs. (C) 1H and (D-E) 19F MRI of NP phantoms using (ii) SPGR and (iii) SSFP. (F) 19F MRI of NP labeled CDCs. Figure 3: Post-mortem application of labelled CT cell administration in the femoral skeletal muscle area of the C57BL/6 mouse. (A) 19F, and (B, C) merged 1H-19F sagittal and coronal views of skeletal muscle MRI.
RF Coils: A quadrature birdcage coil (diameter=34 mm), an 80×40 mm2 butterfly, and a 5×8 mm2 solenoid coil, were constructed for 19F-MRI 375.8 MHz. Pulse Sequences: SPGR, RARE, and SSFP sequences were optimized based on simulations and phantom experiments. Optimal sequences/settings were applied to phantoms, and PFC NP-labeled CDC cells. Quantification: 19F-MRI coil sensitivity and concentration detection limits were determined in TFA solutions and PFC NPs emulsions mixed in IMDM (Thermo Fisher Scientific, UK) using the birdcage coil (TR=8.5-34 ms, TE=4.1-5.7 ms, flip angles=20-30°, NEX=256-1098, FOV=40×40 or 60×60 mm2, slice thickness=2-10 mm, BW=4-6 kHz, matrix=32×32, and in RARE, TR=1100 ms, ETL=16, segments=2). Image Processing: Images were processed in MATLAB/ImageJ, and MRS spectra (TR=2 or 20 s, 512 points, BW=20 kHz, NEX=4 or 16, flip angle=90°) in CSX (Johns Hopkins, USA) and IDL (Harris Geospatial Solutions, USA). To minimize B1 effects, all samples were placed as close as possible on the coil’s surface at the same location.
Results Figure 1 shows parametric SNR maps for SPGR, RARE, and SSFP imaging Experimental results show that SSFP achieved the highest SNR (2, and and 7-fold higher than SPGR and RARE, respectively) based on phantom imaging (Fig. 1B) Image-based quantification confirmed linear signal-concentration dependence with a 10 mM detection limit (birdcage, TFA/NP labels) [Figs. 1C/2C] The minimum detectable NP dose was 2.5 mg/ml (birdcage), and the CDC-NP concentration was ~0.38 mM (solenoid) (Fig. 2) Feasibility of the fast imaging protocol is confirmed with 1H and 19F MRI in the post-mortem mouse (Fig. 3)
Acknowledgements The project leading to this article has received funding (CC) from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No. 652986 and BHF grant FS/11/50/29038.
Discussion-Conclusions
We have established fast, quantitative 19F-MRI, with optimized SNR performance (SSFP) The 19F detection threshold was approximately 10 mM (birdcage), and 0.5-1 mM (butterfly) in ultrafast acquisitions (
1 Maguire ,
19F
2 Swider ,
MRI: Optimized Imaging Strategies 3 Malandraki-Miller ,
4 Srinivas ,
3 Carr ,
1 Schneider
C. M. L. E. S. M. C. A. J. E. 1 3 Departments of Cardiovascular Medicine, and Physiology, Anatomy, Genetics, University of Oxford, Oxford, UK and the 4Department of Immunology, Radboud University, The Netherlands
Purpose-Introduction Implantation of stem cells (SCs) aims to provide tissue regeneration and cardiac functional improvements following ischemic injury. While the feasibility of SC therapy has been proven, its efficacy has yet to be demonstrated. Correspondingly, the visualization of implanted SCs to help define optimal therapy strategies has become a subject of intense research Over the past decade, 19F MRI has been used to image and track exogenously labeled nanoparticles (NP labels) in vivo [1, 2] While prior works attempted to optimize fluorine acquisitions, these had either focused on spectroscopy [3, 4] or used dedicated imaging sequences [5] We present an optimized, 19F MRI approach that enables a) ultrafast acquisitions with optimal signal-to-noise (SNR) and b) image-based quantification.
Aim: to optimize the MRI approach enabling a) ultrafast acquisitions with optimal signal-to-noise (SNR), b) image-based quantification, and c) postmortem murine applications
Methods Simulations: Spoiled gradient echo (SPGR), rapid acquisition with relaxation enhancement (RARE), fid/echo/balanced steady state free precession (SSFP) sequences were simulated [6], and optimized parametric SNR maps generated in MATLAB (Mathwords, USA) using typical 19F relaxation times and scan parameters. MRI/MRS Studies: Phantoms: Aqueous phantoms (1-100 mM), containing trifluoroacetic acid (TFA), perfluorocarbon (PFC) NP labels, and labeled cardiosphere-derived cells (CDC) were used to optimize radiofrequency (RF) coil responses at 9.4 T. A post-mortem study (C57BL/6) was conducted using the butterfly coil (19F: TR/TE=5.71-6.32/2.87-3.16 ms, flip angles=30-50°, NEX=1048-2096, FOV=50-60×50-60 mm2, slice thickness=35 mm, axial, BW=6 kHz, matrix=32×32 in a total of 3.11-7.4 min) where labelled cells were injected in the femoral area of the mouse hindlimb.
Figure 1: A. Parametric SNR plots using TFA T1=2.68 s and T2=2.37 s. B. 2D phantom SNR optimization using the birdcage coil (100 mM TFA), and (C) validation of quantification.
Figure 2: (A) 19F MRS in NPs and (B) labelled CDCs. (C) 1H and (D-E) 19F MRI of NP phantoms using (ii) SPGR and (iii) SSFP. (F) 19F MRI of NP labeled CDCs. Figure 3: Post-mortem application of labelled CT cell administration in the femoral skeletal muscle area of the C57BL/6 mouse. (A) 19F, and (B, C) merged 1H-19F sagittal and coronal views of skeletal muscle MRI.
RF Coils: A quadrature birdcage coil (diameter=34 mm), an 80×40 mm2 butterfly, and a 5×8 mm2 solenoid coil, were constructed for 19F-MRI 375.8 MHz. Pulse Sequences: SPGR, RARE, and SSFP sequences were optimized based on simulations and phantom experiments. Optimal sequences/settings were applied to phantoms, and PFC NP-labeled CDC cells. Quantification: 19F-MRI coil sensitivity and concentration detection limits were determined in TFA solutions and PFC NPs emulsions mixed in IMDM (Thermo Fisher Scientific, UK) using the birdcage coil (TR=8.5-34 ms, TE=4.1-5.7 ms, flip angles=20-30°, NEX=256-1098, FOV=40×40 or 60×60 mm2, slice thickness=2-10 mm, BW=4-6 kHz, matrix=32×32, and in RARE, TR=1100 ms, ETL=16, segments=2). Image Processing: Images were processed in MATLAB/ImageJ, and MRS spectra (TR=2 or 20 s, 512 points, BW=20 kHz, NEX=4 or 16, flip angle=90°) in CSX (Johns Hopkins, USA) and IDL (Harris Geospatial Solutions, USA). To minimize B1 effects, all samples were placed as close as possible on the coil’s surface at the same location.
Results Figure 1 shows parametric SNR maps for SPGR, RARE, and SSFP imaging Experimental results show that SSFP achieved the highest SNR (2, and and 7-fold higher than SPGR and RARE, respectively) based on phantom imaging (Fig. 1B) Image-based quantification confirmed linear signal-concentration dependence with a 10 mM detection limit (birdcage, TFA/NP labels) [Figs. 1C/2C] The minimum detectable NP dose was 2.5 mg/ml (birdcage), and the CDC-NP concentration was ~0.38 mM (solenoid) (Fig. 2) Feasibility of the fast imaging protocol is confirmed with 1H and 19F MRI in the post-mortem mouse (Fig. 3)
Acknowledgements The project leading to this article has received funding (CC) from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No. 652986 and BHF grant FS/11/50/29038.
Discussion-Conclusions
We have established fast, quantitative 19F-MRI, with optimized SNR performance (SSFP) The 19F detection threshold was approximately 10 mM (birdcage), and 0.5-1 mM (butterfly) in ultrafast acquisitions (
