Quantification of Prostate Cancer Metabolism Using <scp>3D Multiecho bSSFP</scp> and Hyperpolarized [<scp>1‐<sup>13</sup>C</scp>] Pyruvate: Metabolism Differs Between Tumors of the Same Gleason Grade
Rafat Chowdhury, Christoph A. Mueller, Lorna Smith, Fiona Gong, Marianthi‐Vasiliki Papoutsaki, Harriet Rogers, Tom Syer, Saurabh Singh, Giorgio Brembilla, Adam Retter, Max Bullock, Lucy Caselton, Manju Mathew, Eoin Dineen, Thomas Parry, Jürgen Hennig, Dominik von Elverfeldt, Andreas B. Schmidt, Jan‐Bernd Hövener, Mark Emberton, David Atkinson, Alan Bainbridge, David G. Gadian, Shonit Punwani
Abstract
Background Three‐dimensional (3D) multiecho balanced steady‐state free precession (ME‐bSSFP) has previously been demonstrated in preclinical hyperpolarized (HP) 13 C‐MRI in vivo experiments, and it may be suitable for clinical metabolic imaging of prostate cancer (PCa). Purpose To validate a signal simulation framework for the use of sequence parameter optimization. To demonstrate the feasibility of ME‐bSSFP for HP 13 C‐MRI in patients. To evaluate the metabolism in PCa measured by ME‐bSSFP. Study Type Retrospective single‐center cohort study. Phantoms/Population Phantoms containing aqueous solutions of [1‐ 13 C] lactate (2.3 M) and [ 13 C] urea (8 M). Eight patients (mean age 67 ± 6 years) with biopsy‐confirmed Gleason 3 + 4 ( n = 7) and 4 + 3 ( n = 1) PCa. Field Strength/Sequences 1 H MRI at 3 T with T 2 ‐weighted turbo spin‐echo sequence used for spatial localization and spoiled dual gradient‐echo sequence used for B 0 ‐field measurement. ME‐bSSFP sequence for 13 C MR spectroscopic imaging with retrospective multipoint IDEAL metabolite separation. Assessment The primary endpoint was the analysis of pyruvate‐to‐lactate conversion in PCa and healthy prostate regions of interest (ROIs) using model‐free area under the curve (AUC) ratios and a one‐directional kinetic model ( k P ). The secondary objectives were to investigate the correlation between simulated and experimental ME‐bSSFP metabolite signals for HP 13 C‐MRI parameter optimization. Statistical Tests Pearson correlation coefficients with 95% confidence intervals and paired t ‐tests. The level of statistical significance was set at P < 0.05. Results Strong correlations between simulated and empirical ME‐bSSFP signals were found ( r > 0.96). Therefore, the simulation framework was used for sequence optimization. Whole prostate metabolic HP 13 C‐MRI, observing the conversion of pyruvate into lactate, with a temporal resolution of 6 seconds was demonstrated using ME‐bSSFP. Both assessed metrics resulted in significant differences between PCa (mean ± SD) (AUC = 0.33 ± 012, k P = 0.038 ± 0.014) and healthy (AUC = 0.15 ± 0.10, k P = 0.011 ± 0.007) ROIs. Data Conclusion Metabolic HP 13 C‐MRI in the prostate using ME‐bSSFP allows for differentiation between aggressive PCa and healthy tissue. Evidence Level 2 Technical Efficacy Stage 1