Developing individualized treatment plans for breast cancer patients necessitates improved characterization of tumor microenvironment to predict tumor progression and response to treatment. Factors implicated in tumor progression include perfusion, oxygenation, and metabolism. Due to the poorly structured vasculature associated with oncologic angiogenesis, it is hypothesized that dynamic contrast-enhanced (DCE) and blood-oxygen-level dependent (BOLD) MRI can be used to characterize associations between regional vascular supply and oxygen availability in breast cancer. DCE parameters have been successfully correlated with exogenous hypoxia markers in cancerous tumors¹ and it is suggested that BOLD parameters provide additional tissue oxygenation information not solely controlled by perfusion². Additionally, aerobic glycolysis (Warburg Effect) is postulated to be upregulated in cancerous tumors³ and has the potential for in vivo interrogation using hyperpolarized (HP) ¹³C magnetic resonance spectroscopic imaging (MRSI) of [1-¹³C]pyruvate and its downstream metabolites⁴. In this study, murine mammary adenocarcinomas were imaged using DCE, BOLD, and HP ¹³C MRI to assess their viability to spatially and quantitatively evaluate tumor vasculature, oxygenation, and metabolism.A pilot study complying with institutional animal care and use committee regulations was conducted using a murine breast cancer model. Cells from ethylnitrosourea-induced mammary adenocarcinomas developed in FVB.B6-Apc^Min/+ mice were injected into axillary fat pads of syngeneic FVB/Tac mice. Mice with both moderately aggressive (fast-growing) and markedly aggressive (very fast-growing) tumor lines were imaged. Imaging was performed on a 4.7T small animal scanner (Agilent, Palo Alto, CA) with a ¹H/¹³C dual-tuned volume coil and ¹³C surface coil (Doty Scientific, Columbia, SC). A high-resolution (0.25×0.25×2mm³) T₂-weighted FSE sequence was acquired for anatomical reference (TR/TE­_eff= 3500/66ms) followed by a multi-echo SPGR T₂*-weighted BOLD sequence (TR/TE/ΔTE= 350/2/2.8ms, 32 echoes) with identical resolution. [1-¹³C]pyruvate was polarized (10-20%) via dynamic nuclear polarization (HyperSense, Oxford Instruments, UK) and 10μL/g was injected into the tail vein for imaging. A single-shot spiral acquisition (3×3×5mm³, TR/TE/ΔTE=90/1.00/1.19ms, echoes=5, FA=10°) acquired dynamic ¹³C images interleaved with slice-selective spectra (FA=5°) at ~5s temporal resolution. Spiral image reconstruction used an iterative, least-squares estimation technique⁵. Actual flip-angle imaging (AFI) and T₁ maps were generated from 3D SPGR sequences (TR/TE= 6.1/1.2ms and 5.9/1.7ms, respectively) with 0.5×0.5×0.5mm³ resolution. Finally, a T₁-weighted DCE SPGR sequence was acquired (TR/TE=20.5/2.9ms, 0.25×0.25×2mm³) with ~5s temporal resolution. 10 frames were acquired prior to IV injection of 0.15mmol/kg of gadodiamide followed by a 10min acquisition. R₂* maps were generated by fitting BOLD data to a linearized exponential decay model. Ratio maps of the volume transfer constant (K^trans) to extravascular-extracellular volume fraction (v_e) were developed by converting DCE signal to gadodiamide concentration using AFI-corrected T₁-maps⁶, then fitting to a linearized reference region model^7,8. Voxel-wise ratio maps of the area-under-the-curve (AUC) of HP lactate-to-pyruvate signal were also generated. R₂* and K^trans/v_e were corrected for major outliers with removal of values greater than three times the inter-quartile range. Mean R₂* and K^trans/v_e values were calculated for the total volume of tumor present in each slice. Three mice each hosting markedly and moderately aggressive tumor lines underwent HP ¹³C, DCE, and BOLD MRI. K^trans/v_e and R₂* maps displayed heterogeneity in both tumor strains, with a tendency of K^trans/v_e to be larger towards tumor peripheries and R₂* to be largest towards the tumor core or air-tissue interfaces secondary to susceptibility (Figure 1). Mean R₂* and K^trans/v_e values were similar for both tumor lines (Figure 2). The ¹³C AUC ratio map (Figure 3) displayed heterogeneous lactate/pyruvate signal within the tumors.R₂* indicates tissue deoxyhemoglobin concentration, with higher R₂* values indicating relative hypoxia, while K^trans/v_e indicates perfusion/permeability of the vasculature. Larger K^trans/v_e values were observed around the tumor periphery, suggesting more leaky, perfused vasculature in this region compared with the core. High R₂* values near the tumor core may indicate the local oxygenated hemoglobin content is low, likely due to a combination of rapid proliferation inhibiting oxygen diffusion and necrosis. Follow-up histology will reveal the degree of necrosis and predominance of tissue oxygen stress markers. ¹³C AUC maps demonstrated higher pyruvate and lactate values along the medial tumor, possibly due to more HP substrate delivery via native vasculature from the abdominal region. The ¹³C AUC ratio map demonstrates heterogeneous distribution of lactate-to-pyruvate. Ongoing studies will allow comparisons of patterns of angiogenesis, oxygen availability, and metabolism with tissue histology to better understand the role of these processes in tumor models of breast cancer.This study demonstrates the technical feasibility of using BOLD, DCE, and HP ¹³C MRI to characterize mammary carcinomas. Future studies are planned to investigate regional associations of elevated R₂*, perfusion/permeability, and glycolytic flux with tissue histology for different breast cancer models.