Synopsis

Keywords: Prostate, Cancer

The first-in-human hyperpolarized [1-¹³C]pyruvate+[¹³C,¹⁵N₂]urea dual-agent MRI demonstrated the safety and feasibility of simultaneous characterization of prostate cancer metabolism and blood flow as a two-minute addition to the standard ¹H-multiparametric MRI. Metabolism-perfusion mismatch (i.e. elevated pyruvate-lactate conversion and decreased urea perfusion) in subregions of the high-grade prostate tumor was in agreement with the histopathological and immunochemical markers that reflected lethal phenotypes and their associated hypoxic tumor microenvironment. Correlative analysis of urea and Gadolinium-derived pharmacokinetic parameters found low correlation between the two, which indicates that HP ¹³C urea’s unique contrast mechanism offers independent information inaccessible through conventional ¹H DCE-MRI.

Purpose

Prostate cancer impacts 1 out of 6 US men¹, but many with low-risk disease suffer from excess morbidity associated with overdiagnosis and overtreatment, and there is an unmet clinical need to distinguish lethal from indolent cancer forms. Preclinical hyperpolarized (HP) ¹³C pyruvate+urea MRI revealed that significantly elevated pyruvate-lactate metabolism and decreased urea perfusion, reflective of hypoxia response signaling, are hallmarks of aggressive (high-grade) prostate cancer phenotypes in mice^2,3. This first-in-human research explored the safety and feasibility of simultaneous metabolism and perfusion/permeability characterization using HP [1-¹³C]pyruvate+[¹³C,¹⁵N₂]urea dual-agent MRI in a patient with histologically-confirmed high-grade prostate cancer. Correlative analysis was performed between HP ¹³C-¹H pharmacokinetic parameters and histopathological/immunochemical markers.

Methods

Patient Characteristics: A patient with biopsy-confirmed high-grade prostate cancer (Gleason 4+3, PSA = 7.7) was enrolled via an IRB-approved protocol (NCT02526368). Key eligibility criteria include biopsy-confirmed prostate cancer, planned radical prostatectomy, and ECOG status of 0 or 1.

Hyperpolarized-¹³C Exam: GMP [1-¹³C]pyruvic acid and [¹³C,¹⁵N₂]urea were co-polarized in a 5T SpinLab (GE Healthcare, Chicago IL) polarizer for 3 hours⁴. The dissolution yielded 125/25 mM pyruvate/urea solution with 36.5% polarization and 0.8μM residual trityl radical. Imaging was conducted on a clinical 3T scanner (MR750, GE Healthcare) using a clamshell ¹³C transmitter and a dual-element ¹H-¹³C endorectal receiver⁵. Pharmaceutical release followed QC verification of IND-approved safety criteria⁴. Post-release quality analysis used a high-field spectrometer (Varian INOVA 500).

Imaging Sequence: Time-resolved pyruvate-lactate imaging used a 2D multislice single-shot spiral sequence (TR =80ms) with 7x7x11.6mm spatial resolution⁶, whereas urea used a 3D stack-of-spiral balanced-SSFP sequence (TR = 12.3ms, α = 50°) at 9x9x11.6mm resolution^7,8. Scan time window was 52s with 2.6s temporal resolution.

Data Processing and Analysis: Quantitative ¹H-DCE parameters (k_TRANS, v_e) were calculated with 3D Slicer’s PkModeling toolbox⁹, using a femoral artery ROI in the prostate plane as arterial input. Semi-quantitative DCE parameters (peak height, slope, washout) were derived using an in-house pipeline¹⁰, and those of urea were calculated in MATLAB. Pairwise Pearson correlation used voxels within the prostate ROI, with DCE parametric maps downsampled to ¹³C resolution. Pyruvate-to-lactate conversion rate k_PL was calculated using an inputless two-site model¹¹. Spatial distributions of all ¹³C tracers were corrected for receiver coil sensitivity profile.

Results and Discussions

The first-in-human co-hyperpolarized [¹³C,¹⁵N₂] urea and [1-¹³C]pyruvate as a two-minute addition to ¹H multiparametric MRI of prostate cancer was safe and feasible without adverse events. The novel workflow (Figure 1) features pharmaceutical compounding of ¹³C-labelled pyruvate+urea⁴, specialized MR coils⁵, pulse sequences^6-8, quantification of novel biomarkers, along with QC and release criteria for co-polarized HP agents⁴.

Surgical pathology from radical prostatectomy identified Gleason 4+5 tumor in the left posterior mid-apex peripheral zone (PZ), confirming the high-risk diagnosis (UCSF CAPRA score 5, Decipher 0.84)¹². The Gleason pattern 5 consisted of solid sheets of cells and comedonecrosis.

Composite biomarker k_PL/urea_AUC ratio (kUR) looks at increased pyruvate metabolism coupled with decreased urea perfusion. This metabolism-perfusion mismatch is a signature of the hypoxia-driven lethal, treatment-resistant prostate tumor phenotypes^2,3,13,14, where overdrive of hypoxia-induced factor Hif1α signaling pathway leads to high glycolysis (Warburg effect) and hyperproliferation^15,16. Heterogeneous kUR (Figure 2) was observed intratumorally with subregions (yellow arrow) as high as 10x the normal-appearing PZ. Indeed, immunohistochemical analysis identified presence of comedonecrosis (green arrow, Gleason 5) with significantly stronger lactate dehydrogenase (LDHA; a direct downstream target of Hif1α upregulated in hypoxic tumor microenvironment¹⁷) immunochemical staining surrounding a central area of necrosis (red arrow) and within intraductal carcinoma component.

The semi-quantitative parametric maps (Figure 3) showed focal urea hyperintensity at the histologically-confirmed left PZ tumor (green arrow), whereas the central gland BPH nodules had low urea (red arrow), as opposed to ¹H-DCE which had moderate enhancement. Time-resolved urea vs Gd-enhanced T₁-weighted signal in tumor and normal-appearing PZ voxels showed that urea distributed much more rapidly throughout the tumor than Gadolinium, but also decreased rapidly - possibly through combined washout and T₁-decay mechanisms. General correlation and pairwise regression diagrams (Figure 4) between urea and DCE pharmacokinetic biomarkers gave weak correlation (Pearson r²_mean = 0.04, range:0.001-0.17), indicating that HP ¹³C urea provided unique, independent perfusion information not accessible through standard ¹H contrast-enhanced MRI.

A possible explanation for HP ¹³C urea’s unique contrast mechanism is that urea is a highly polar, diffusive small-molecule agent that is particularly sensitive to permeability differences in the microvasculature^18,19. Urea extravasates more rapidly from the highly permeable, leaky angiogenic tumor neovasculature into the interstitial space, than it does from the highly vascularized but less permeable benign central gland nodules, analogous to von Morze’s findings of much higher urea signal levels in the highly permeable mouse tumor than the poorly permeable brain¹⁸. These preliminary findings warrant future investigations of composite metabolism-perfusion biomarkers in a larger cohort of patients for early detection of aggressive prostate tumors, and illustrated the need to develop more sophisticated quantitative modeling of urea perfusion/permeability to fully exploit the wealth of information it has to offer.

Conclusions

This first-in-human co-hyperpolarized [¹³C,¹⁵N₂] urea and [1-¹³C]pyruvate + ¹H mpMRI of prostate cancer demonstrated safety and feasibility of simultaneous characterization of perfusion and metabolism in human malignancies using hyperpolarized agents. Metabolism-perfusion mismatch detected hypoxia-driven aggressive cancer subtypes, consistent with pathological/immunochemical findings. Carbon-13 urea biomarkers provided unique perfusion information inaccessible through conventional ¹H DCE-MRI.

Acknowledgements

This work was supported by grants from the NIH (U01EB026412, R01CA214554, U01CA232320, P41EB013598). We would like to thank Priscilla Chan, Heather Daniel, Evelyn Escobar, Mary Frost, Jasmine Hu, Dr. Yaewon Kim, Dr. Philip Lee, Kimberly Okamoto, Dr. Andrew Riselli, and Dr. James Slater for their help with this research.