A collaborative team of Cancer Center at Illinois (CCIL) researchers is developing methods to improve magnetic resonance spectroscopic imaging (MRSI) for mapping important metabolites in the tumor microenvironment in vivo.
Magnetic resonance imaging (MRI or MR imaging) is an invaluable clinical tool for safe and non-invasive disease diagnosis and monitoring. MRI enables clinicians to obtain detailed images of the structures in the body, including the shape and size of tumors. MRSI uses the same hardware, but instead allows clinicians and researchers to visualize metabolites inside the tissue. The abundance of these metabolites, molecules that play important physiological functions, are often different in healthy and cancerous tissue and can change over time.
MRI produces anatomical images (top left). MRSI can provide a map of metabolites in the tumor region (top right) and spectra of various metabolites (bottom row).
“MRSI can measure and quantify spatially-resolved metabolite profiles which can offer insights into the metabolic and biochemical states of the tumors. This may be used to noninvasively characterize the stages of the tumor during progression and monitor therapeutic responses over time by measuring corresponding metabolic changes in the tumor before and after treatment,” said CCIL member Fan Lam, an associate professor of bioengineering who led the collaborative project.
Traditional MR images are generated by electromagnetic signals from water molecules. These signals are then reconstructed into an image with computer algorithms. The amount of water in a tissue and the water molecules’ interactions with the environment around it determine how bright or dim structures will appear in the MR image.
In comparison, MRSI uses the signals originating from metabolites beyond water. While the molecular insights gained from MRSI can be incredibly valuable, data collection takes longer and is much more complex than MRI, so it is difficult to obtain high quality images, especially in small, early-stage tumors.
Yizun Wang is a bioengineering Ph.D. student in Lam’s Research Group
“One key challenge is that the spatial resolution for MRSI is not good enough to map small tumors or their boundary clearly, so we need to improve this. As an example, if a tumor is one millimeter across or less, we need the resolution to be submillimeter to actually get any information about the tumor microenvironment.” said Yizun Wang, a Ph.D. student in Lam’s research group and first author of the publication.
Spatial resolution is a value for the smallest distance in which two adjacent objects can be distinguished from each other. Drawing a comparison to photography, small details may be pixelated and hard to pinpoint in a low-resolution photo but can become clear and distinguishable in a high-resolution photo.
However, the metabolite signals are 10,000 times weaker than those of water molecules, which make up 50-75% of the human body, so their signals need to be enhanced for higher resolution imaging while minimizing interference from water. Wang led the charge to develop and implement two main innovations to strengthen the metabolite signals and overcome water interference. First, they used a specific excitation strategy during data acquisition to exclusively look at the metabolites without disturbing the water molecules. Second, they further enhanced the quality of their metabolite data by using a “learned-subspace-based” image reconstruction method–an advanced computational algorithm.
“As far as we know, we are the first to report a submillimeter scale resolution for MRSI under 15 minutes,” said Wang. With this increased resolution, they were able to successfully use MRSI to map unique tumor specific metabolic signatures in tumor-bearing mice longitudinally on a Bruker 9.4 T MRI system.
But the researchers don’t want to stop there. Looking ahead, Wang said, “We want to continue to improve these techniques and eventually be able to translate our methods for human patients. In the future, if we can obtain tumor-cell-specific metabolic signatures in vivo, this has the potential to be a more precise diagnostic tool and to help monitor tumor recurrence, which is a significant unmet clinical need.”
Developing and accessing new clinical and preclinical technology is not an easy task, and often requires expert, interdisciplinary knowledge ranging from engineering to biology. Lam said, “This work would not have been possible without collaborative efforts from CCIL members and associate members Andrew Smith, Ed Roy, and Jonathan Sweedler, as well as graduate student Urbi Saha. We are very grateful for the CCIL seed grants that provided support for the graduate students who carried out this work.” The team also acknowledges the Biomedical Imaging Center for their support on the 9.4 T system.
Editor’s notes:
The research article “High-resolution 1H-MRSI at 9.4 T by integrating relaxation enhancement and subspace imaging” is available online. https://doi.org/10.1002/nbm.5161
Fan Lam is an Associate Professor of Bioengineering and an affiliate of the Department of Electrical and Computer Engineering, the Beckman Institute for Advanced Science and Technology,
The Carl R. Woese Institute for Genomic Biology, and the Carle Illinois College of Medicine.
To contact Fan Lam, email fanlam1@illinois.edu
This story was written by Katie Brady, CCIL Communications Intern