A pressing need facing the clinical management of prostate cancer patients is an accurate method for distinguishing aggressive, potentially lethal, prostate cancer from indolent disease. Prostate cancer (PCa) is the second most prevalent cancer in American men, with 1 in 6 American men being diagnosed, but is fatal for only 12% of these cases [1]. This hyperpolarized ¹³C MR study of 19 prostate-cancer mice was designed to measure metabolic and perfusion parameters simultaneously by HP ¹³C-urea & ¹³C-pyruvate dynamic MRI in primary and metastatic tumors. Correlations with ex vivo measures of aggressiveness were performed with subsequent histochemical, gene expression and enzymatic activity assays.

Sequences: 3D dynamic compressed-sensing ¹³C-EPSI provided high spatiotemporal resolution for simultaneous imaging of metabolism and perfusion. The sequence employed variable flip angle and multiband excitation, followed by double spin echo refocusing and a compressed-sensing 3D EPSI readout. (Spatial Resolution: 3.3mm x 3.3mm x 5.4 mm, Temporal Resolution = 2s, FOV = 4cm x 4cm x 8.6cm) Anatomical reference images were acquired using T₂-FSE.

MRI experiments: 19 transgenic prostate cancer mice (TRAMP) were studied using a clinical 3T scanner with a dual-tuned mouse coil for ¹³C and proton imaging. The imaging studies were conducted for each animal prior to or on the day of sacrifice. [1-¹³C]pyruvate and ¹³C urea were co-polarized using a prototype clinical polarizer for 2 hours, reaching approximately 20-30% percentage polarization. ~350ul of the ¹³C-labelled compounds (concentration = 80mM for both pyruvate and urea) were rapidly injected into the mice via tail vein catheter over 15 seconds, and the ¹³C scan began at the end of injection.

Tissue analysis: Following the HP MR studies the animals were sacrificed and the excised tissue was sectioned and histochemical, gene expression and activity analyses were performed. Histochemical assays included H&E, Ki-67 and PIM immunostaining. Gene expression analysis consists of HIF-1α, LDHA, LDHB, VEGF, MCT1 and MCT4. LDH isoenzyme activity was also measured.

A total of 19 TRAMP mice were studied. 9 were histo-pathologically categorized as low-grade, while 10 were high-grade. Lymph node metastases were found in 5 of the high-grade mice. The mean tumor size in low-grade animal was 0.184±0.082(cm³), whereas that of high-grade was 4.03±3.60(cm³), and of metastasis was 0.11±0.09(cm³). Rate of pyruvate-to-lactate conversion (k_PL) was significantly (P<0.00001) higher in high-grade (0.0560±0.0047s^-1) and metastatic cancers (0.0593±0.0261s-1) versus low-grade tumor (0.0197±0.0012s^-1) as shown in Figures 1-3. No overlap in k_PL was observed between low- and high-grade tumors. Perfusion area-under-curve (AUC) measurements significantly reduced in high-grade disease (high-: 640.5±94.1, low-: 1407.4±221.9 A.U., High grade/Low grade=45.5%, P<0.004), while k_trans significantly increased (high-: 358.4±38.5, low-: 180.1±24.1 s^-1, P<0.002) as shown in Fig1.D). Fig1.B) shows significantly higher LDH activity in high grade tumors (High grade/Low grade=196.9%). As shown in Fig.2&5, both Ki-67 staining for proliferation and PIM staining for hypoxia increased in high-grade tumors. No significant difference in k_PL(P>0.8) or perfusion AUC(P>0.4) was found between high-grade and metastatic tissue (Figures 3-4). k_trans, as a combined measure of perfusion and permeability, indicated a loss of morphology/function and increased leakiness in the vasculature of high-grade cancers.Histologically high-grade tumor exhibited significantly higher k_PL, with no overlap from the low-grade cases, whereas perfusion AUC was significantly lower. Lymph-node metastases demonstrated similar k_PL and perfusion indices as high-grade tumor, indicating high aggressiveness in these metastatic sites as well. These data are highly consistent with previous findings but the 3D dynamic HP ¹³C imaging is much less susceptible than single-timepoint studies to timing variability of bolus infusion and arrival, thus offering a more robust way to quantitatively analyze metabolism and perfusion in vivo. In conclusion, this 3D dynamic CS-EPSI acquisition and modeling approach not only allows identification and assessment of aggressiveness in preclinical primary and metastatic tumors, but also shows great potential for clinical translation.