Hyperpolarized ¹³C Pyruvate MRI: An Important Window into Tumor Metabolism

See also the article by Dou et al in this issue.

Michael A. Ohliger, MD, PhD, is an associate professor of radiology in the abdominal imaging section of the Department of Radiology and Biomedical Imaging at the University of California San Francisco. His research interest is in developing novel methods for molecular imaging in the abdomen, with a focus on hyperpolarized ¹³C MRI for liver disease and cancer, as well as targeted PET agents for imaging infection.

Molecular characterization of tumors, particularly tumor metabolism, has become increasingly important for detection, grading, and determining treatment response. In recent years, hyperpolarized (HP) carbon 13 (¹³C) MRI has emerged as a promising new approach for assessing tumor metabolism in vivo (1). Using dynamic nuclear polarization, the nuclear magnetization (and therefore the MRI signal) of carbon-containing molecules is increased by a factor of 10 000, permitting the observation of metabolism in real time. The most used hyperpolarized compound for human studies, 1-¹³C-pyruvate, sits at a crossroads in the pathway of glycolysis and can be observed to change into hyperpolarized lactate or, through mitochondrial oxidation, hyperpolarized bicarbonate (2). Increased HP ¹³C lactate production is observed in the majority of tumors and is reflective of the Warburg effect (3), where tumors produce excess lactate even in the presence of sufficient oxygen. HP ¹³C pyruvate MRI has successfully been used in numerous preclinical cancer models and recently, under investigational new drug (IND) approval by the U.S. Food and Drug Administration, in humans with cancers of the breast, brain, liver, kidney, and prostate (1).

In this issue of Radiology: Imaging Cancer, Dou et al (4) examined the behavior of HP ¹³C pyruvate metabolism in two different rodent models of hepatocellular carcinoma (HCC): N1S1, which generally produces single lesions, and McA-RH7777, which has high metastatic potential. The tumors were examined at baseline and following stimulation by applying radiofrequency ablation (RFA) to the non–tumor-bearing liver. This method was previously shown to worsen metastatic disease in animal models of liver cancer (5). Finally, metabolomic profiling of the tumors was performed and correlated with the HP ¹³C pyruvate MRI results. The authors found that while one model, N1S1, generated the expected increase in HP ¹³C lactate signal (approximately 80% higher than normal liver), the HP ¹³C lactate signal in McA-RH7777 could not be distinguished from liver background. This behavior persisted even when the tumors were stimulated using RFA, in which case the HP ¹³C lactate signal increased further in N1S1 tumors but not in McA-RH7777 tumors. The authors hypothesized that this reflected underlying metabolic differences in the two tumor types. Indeed, metabolomic and genetic analysis showed elevation in proteins and genes related to glycolysis in N1S1 but not in McA-RH7777.

This study is highly impactful because it establishes that HP ¹³C pyruvate MRI can be used to distinguish between two broad liver tumor types, depending on their metabolic programming. As a practical matter, the study demonstrates that when HP ¹³C pyruvate MRI is used to follow HCC, not all HCC may demonstrate increased HP ¹³C lactate signal. Rather, it will be important to first determine whether a patient’s tumor is one that relies on upregulated glycolysis. A similar consideration is encountered when using fluorine 18 fluorodeoxyglucose (FDG) PET, as it is well-known that HCC frequently does not show increased FDG accumulation compared with background (6). It will be interesting to determine in future studies if there is overlap between FDG-negative tumors and those that do not show elevated HP ¹³C lactate signals.

Although 1-[¹³C]pyruvate is the most frequently used hyperpolarized compound in humans, a large number of other ¹³C-containing molecules can be hyperpolarized, including those that are sensitive to pH, oxidative stress, perfusion, or other metabolic processes (7). IND-enabling studies for many of these molecules by different research groups are ongoing, and it will be interesting to see if one of these newer compounds may be sensitive to metabolism in tumors that do not rely principally on glycolysis. The authors suggest that because McA-RH7777 shows upregulation in branched-chain amino acid metabolism, hyperpolarization of ¹³C- or nitrogen 15–labeled amino acids may potentially be useful for this purpose (4). One of the unique aspects of HP ¹³C MRI compared with other molecular imaging technologies, such as FDG PET, is that more than one molecule can be polarized at one time, allowing several aspects of metabolism to be imaged at once (8).

Although the results of this study are encouraging, more development is required to determine its ultimate impact on human disease. The study focuses on rodent models of HCC, and it is uncertain whether the results will hold when applied to humans. Additionally, it is uncertain whether the method used for tumor stimulation (off-target RFA of the liver) is representative of tumors encountered in human disease. On the other hand, the techniques presented are readily translatable. HP ¹³C pyruvate MRI can be performed in human liver and human liver tumors (9). The tools are available to test these ideas in humans, and this study has provided motivation for why such tests should be performed. One final question is whether distinguishing tumor types in this way might be used to determine the most optimal therapy for patients. HCC is often treated initially without first obtaining a biopsy, making noninvasive molecular characterization of tumor metabolism even more valuable. Furthermore, many systemic treatments for HCC target glycolysis (10), and this method may potentially help target tumors in which these treatments are likely to be effective.

In summary, HP ¹³C pyruvate MRI has great promise for following and assessing tumors of the liver. The scan is fast, taking only 1–2 minutes, and no ionizing radiation is required. Finally, because it is MRI-based, it can be integrated into regular imaging for HCC, which often relies on MRI. The metabolic information provided by HP ¹³C pyruvate MRI combined with the anatomic and perfusion information of standard MRI is likely to be a powerful combination. In this study, Dou et al show that by capturing the metabolic heterogeneity of HCC, HP ¹³C pyruvate MRI can play an even more important role in caring for these patients.