Robust, Quantitative Methods Applied to Clinical Hyperpolarized C-13 MR of Prostate Cancer Patients
Peder Eric Zufall Larson1, Jeremy Gordon1, John Maidens2, Murat Arcak2, Hsin-Yu Chen1, Galen Reed1, Ilwoo Park1, Rahul Aggarwal3, Robert Bok1, Sarah J Nelson1, John Kurhanewicz1, and Daniel B Vigneron1

1Radiology and Biomedical Imaging, University of California - San Francisco, San Francisco, CA, United States, 2Electrical Engineering & Computer Sciences, University of California - Berkeley, Berkeley, CA, United States, 3Medicine, University of California - San Francisco, San Francisco, CA, United States

Synopsis

Clinical evaluation of metabolic MRI using hyperpolarized C-13 agents has begun in earnest at multiple sites with the availability of the SpinLab commercial polarizer. For this technology to succeed, robust imaging and analysis methods for quantification of metabolic activity are required. We have developed and are applying efficient dynamic imaging methods, robust kinetic models, and specialized calibration schemes to enable accurate and reproducible quantification in clinical hyperpolarized MR studies.

Purpose

Clinical evaluation of metabolic MRI using hyperpolarized (HP) C-13 agents has begun in earnest at multiple sites with the availability of the SpinLab commercial polarizer. For this technology to succeed, robust imaging and analysis methods for quantification of metabolic activity are required. The goal of this project was to develop and apply specialized calibration schemes, efficient dynamic imaging methods, and robust kinetic models for quantitative clinical HP MR exams.

Methods

Dynamic Imaging: Single time-point imaging with HP C-13 agents, as applied in the first clinical trial [1], can provide ratiometric quantification (e.g. lactate:pyruvate), but this approach is highly susceptible to variations in HP agent delivery times (Fig. 1).

In our clinical HP 13C-pyruvate studies of brain and prostate tumors, we are using dynamic imaging methods with kinetic modeling to provide more robust measurements of metabolic conversion.

In a Phase I trial of prostate cancer [1], we imaged 4 patients with a 2D dynamic MRSI acquisition with the following key parameters: multiband RF excitation (10º pyruvate, 20º lactate), symmetric EPSI, 10mm in-plane resolution, 1.2-3.5 cm slice thickness, temporal resolution = 5 s, imaging started 5 sec after end of saline flush.

Currently we are performing patient exams with a 3D dynamic MRSI acquisition using compressed sensing, based on [2] with variable flip angle RF pulse schemes aimed to maximize total lactate signal [3]. Other key parameters include: 8mm isotropic resolution, 12x12x16 matrix size, temporal resolution = 2 s, imaging started 5 sec after end of saline flush.

We have also developed a multi-slice metabolite-specific acquisition for clinical applications based on spectral-spatial excitation of a single metabolite (lactate and pyruvate) followed by an EPI readout similar to [4] with optimizations to reduce TE, improve SNR efficiency, and eliminate EPI artifacts [5].

Studies were performed with ≈250 mM HP [1-13C]-pyruvate at a dose of 0.43 mL/kg polarized using a 5T GE SpinLab instrument (additional details can be found in [1]).

Metabolism Quantification: We quantified our data using a 2-site kinetic model, in which pyruvate was taken to be the input and kPL values were fit [6], and accounting for variable flip angles [7]. We also used an area under curve ratio (AUCratio), AUClac:AUCpyr, which has been shown to correlate with kPL for both constant [8] and variable flips [7].

C-13 Calibration Methods: Achieving high SNR efficiency and robust quantification requires accurate hardware calibration. Custom 13C calibration procedures were used: First, the frequency of a 13C-urea reference embedded in the endo-rectal coil was measured. 13C RF calibration was performed using the urea reference, and adjusted based on phantom B1+ maps (shown below). B1+ was measured using double angle measurements. To improve the B0 homogeneity, we used the same shimming procedure as for our 1H MRSI clinical prostate studies.

Results & Discussion

Bolus dynamics: In the Phase I trial, peak pyruvate in the prostate was observed at 13 ± 6 s following injection due to individual physiologic and experimental variations (Fig. 1, Table 1), which is much larger than what we observe in preclinical experiments. Furthermore, this was in a relatively healthy population of patients with localized prostate cancer, no significant history of cardiac disease and adequate baseline organ function. The bolus delivery variability will increase in a broader range of patients with more significant disease burdens and larger variations in cardiac function.

Quantification of metabolic conversion: Both kPL and AUCratio corresponded well with regions of biopsy-proven prostate cancer with 2D and 3D dynamic MRSI acquisitions (Figs. 2 & 3). These two metrics of pyruvate to lactate conversion were similar, as we observed previously in a prostate tumor mouse model [7].

C-13 Calibration: Figure 4 shows 13C transmit RF (B1+) phantom results, which are used for RF calibration. These maps could also be used to improve the accuracy of kinetic modeling. We also measured B1+ in 5 prostate cancer patients following the imaging study, and found the flip angle to be 9.3 ± 4.2% less than expected, although some underestimation will result from slice profile effects [9] and biases towards the residual magnetization. In patient studies, we measured the peak B1+ at the 13C-urea reference to be 0.84 ± 0.30 G. We chose to limit the peak RF amplitude to 0.60 G, which has been achieved in all human prostate cancer studies.

Conclusion

Evaluation and widespread adoption of clinical HP MR requires robust imaging and analysis methods for quantification of metabolic activity. We have developed and are applying efficient dynamic imaging methods, robust kinetic models, and specialized calibration schemes to enable accurate and reproducible quantifications in clinical hyperpolarized MR studies.

Acknowledgements

This work was supported by NIH grants R01EB016741, R01EB017449 and P41EB013598, and GE Healthcare.

References

[1] Nelson, S. J., Kurhanewicz, J., Vigneron, D. B., Larson, P. E. Z., Harzstark, A. L., Ferrone, M., van Criekinge, M., Chang, J. W., Bok, R., Park, I., Reed, G., Carvajal, L., Small, E. J., Munster, P., Weinberg, V. K., Ardenkjaer-Larsen, J. H., Chen, A. P., Hurd, R. E., Odegardstuen, L.-I., Robb, F. J., Tropp, J., and Murray, J. A. Metabolic imaging of patients with prostate cancer using hyperpolarized [1-13c]pyruvate. Sci Transl Med 5, 198 (Aug 2013), 198ra108.

[2] Larson, P. E. Z., Hu, S., Lustig, M., Kerr, A. B., Nelson, S. J., Kurhanewicz, J., Pauly, J. M., and Vigneron, D. B. Fast dynamic 3d mr spectroscopic imaging with compressed sensing and multiband excitation pulses for hyperpolarized 13c studies. Magn Reson Med 65, 3 (Mar 2011), 610–9.

[3] Xing, Y., Reed, G. D., Pauly, J. M., Kerr, A. B., and Larson, P. E. Z. Optimal variable flip angle schemes for dynamic acquisition of exchanging hyperpolarized substrates. J Magn Reson 234 (Sep 2013), 75–81.

[4] Cunningham, C. H., Chen, A. P., Lustig, M., Hargreaves, B. A., Lupo, J., Xu, D., Kurhanewicz, J., Hurd, R. E., Pauly, J. M., Nelson, S. J., and Vigneron, D. B. Pulse sequence for dynamic volumetric imaging of hyperpolarized metabolic products. J Magn Reson 193, 1 (2008), 139–146.

[5] Gordon, Jeremy, Machingal, Sonam, Kurhanewicz, John, Vigneron, Daniel B., Larson, Peder. Ramp-Sampled, Symmetric EPI for Rapid Dynamic Metabolic Imaging of Hyperpolarized 13C Substrates on a Clinical MRI Scanner. Proc. 23rd ISMRM (2015), p. 4605.

[6] Khegai, O., Schulte, R. F., Janich, M. A., Menzel, M. I., Farrell, E., Otto, A. M., Ardenkjaer-Larsen, J. H., Glaser, S. J., Haase, A., Schwaiger, M., and Wiesinger, F. Apparent rate constant mapping using hyperpolarized [1-(13)C]pyruvate. NMR Biomed 27, 10 (Oct 2014), 1256–65.

[7] Bahrami, N., Swisher, C. L., Von Morze, C., Vigneron, D. B., and Larson, P. E. Z. Kinetic and perfusion modeling of hyperpolarized (13)c pyruvate and urea in cancer with arbitrary rf flip angles. Quant Imaging Med Surg 4, 1 (Feb 2014), 24–32.

[8] Hill, D. K., Orton, M. R., Mariotti, E., Boult, J. K. R., Panek, R., Jafar, M., Parkes, H. G., Jamin, Y., Miniotis, M. F., Al-Saffar, N. M. S., Beloueche-Babari, M., Robinson, S. P., Leach, M. O., Chung, Y.-L., and Eykyn, T. R. Model free approach to kinetic analysis of real-time hyperpolarized 13c magnetic resonance spectroscopy data. PLoS One 8, 9 (2013), e71996.

[9] Deppe, M. H., Teh, K., Parra-Robles, J., Lee, K. J., and Wild, J. M. Slice profile effects in 2d slice-selective mri of hyperpolarized nuclei. J Magn Reson 202, 2 (Feb 2010), 180–9.

Figures

Figure 1: (a) Simulation results demonstrating that a modest 5 s differences in delivery will result in lactate:pyruvate ratio differences of > 20% under typical conditions. (b) Dynamics in two prostate cancer patients from the Phase I trial [1], showing timing differences of 10 s in prostate tumors (red arrows).

Table 1: Pyruvate and lactate kinetic characteristics as measured in 3 prostate cancer patients with 2-D dynamic MRSI. *10º flip for both pyruvate and lactate, #lactate flip = 20º, pyruvate flip = 10º.

Figure 2: Quantification of metabolic conversion using kPL and AUCratio from 2D dynamic MRSI acquisitions (in Phase I trial [1]). The yellow arrows indicate the location of the prostate tumors, as identified in the 1H multi-parametric MR exam. Additional kinetic parameters are summarized in Table 1.

Figure 3: Quantification of metabolic conversion using kPL and AUCratio from a 3D dynamic MRSI acquisition (in ongoing clinical trial). The yellow arrows indicate locations of prostate tumors, as identified by biopsy and in a prior 1H multi-parametric MR exam.

Figure 4: 13C RF calibration results, including both 13C-urea standards placed within the RF receive coils as well as ethylene glycol phantoms used to mimic the tissues of interest.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
2347