That leaves patients with a tricky choice. Do they wait and watch? Let doctors take a brain biopsy? Or, in some cases, endure another brutal round of treatment just in case the tumor has returned?
But a new University of Michigan study shows that a relatively new kind of brain scan may give these patients the reassurance -- or early warning -- that they can't get from the usual scans. U-M radiologists will present the evidence today at the annual meeting of the American Roentgen Ray Society, a major radiology organization.
The approach is called 2D CSI MRS, short for two-dimensional chemical shift imaging magnetic resonance spectroscopy. It allows doctors to non-invasively detect the levels of certain chemicals in brain tissue. Using the relative quantities of these chemicals, doctors can tell what's really going on near a tumor's original location.
The U-M Health System neuroradiology team will show that they have successfully developed a way to use the technique so that, in the vast majority of cases, they can tell the difference between recurring tumor, normal tissue, and tissue that's inflamed or dying because of successful treatment.
"Using 2D CSI is like making a chemical thumbprint of the tumor and the surrounding tissue, and we can use the unique readings from various areas to determine what's cancerous, and what's treatment-related change," says Patrick Weybright, M.D., the U-M radiology resident who will present the results. "This allows us to give the patient earlier and more accurate information on what's happening in their brain at a molecular level."
Weybright will show results from 29 patients ages four years to 54 years, who were split almost evenly between those whose cancer had returned, and those with no recurrence. In addition to the biopsy or surgery that verified their status, nearly all of them had the 2D CSI type of scan, also called a multi-voxel scan. One had a less sophisticated single-voxel chemical scan. All the patients also had conventional diagnostic imaging with MRI scans to show brain tissue and tumor structure.
"In all, we were able to show a significant difference in chemical ratios for those who had treatment-related changes and those who had recurrent tumor," says senior study author Pia Maly Sundgren, M.D., Ph.D., associate professor of radiology.
"This, combined with conventional MRI and clinical indicators, should help increase the accuracy of diagnosis, make diagnosis more timely, and improve the quality of life for patients," Sundgren continues. "And helping them have a better life is the ultimate goal."
The U-M researchers are looking forward to the arrival of advanced three-Tesla MRI scanners that will increase their ability to look for more chemicals on the 2D CSI scans, and allow them to do three-dimensional scans that will increase accuracy even more.
But at the same time, they hope their results will help physicians and insurance companies see the true clinical value of 2D CSI MRS, which is still not universally covered by health plans in an amount sufficient to cover the time needed to read the scans.
Single-voxel CSI MRS -- more common, but less exhaustive, than the 2D multi-voxel approach -- is routinely used to evaluate patients with strokes, oxygen deprivation, epilepsy, multiple sclerosis, and other brain disorders. It is also used to determine the grade of brain tumors.
Both single-voxel and multi-voxel spectroscopy techniques add far more information to the diagnostic process than a regular MRI, which uses an injected dye and a strong magnetic field to make a grayscale picture of brain tissue structure based on the chemical signal of water molecules.
The MRI dye, also called a contrast agent, helps radiologists see finer details of the tissue structure more clearly. But because regular MRI doesn't reveal anything about what's going on inside those tissues at a molecular level, the "contrast-enhanced" areas that look suspicious on an MRI of a brain tumor survivor's brain might be either dying tissue or growing tumor.
But the multi-voxel 2D CSI MRS scans, made using the same MRI equipment, produce line graphs with peaks that correspond to the levels of certain chemicals in various areas around the tumor.
The spikes on the graphs reveal the relative concentrations of chemicals that are produced or used up by growing cancers or dying cells. Depending on which area of the brain is scanned, those concentrations might be high or low. And, the U-M team shows, by making a ratio between the levels of two chemicals, it's possible to show clear differences between different kinds of tissue.
Some of the chemicals studied in the new U-M research include N-acetylaspartate, or NAA, a chemical that's produced by normal brain cells; lactate, produced by cells that have been starved of oxygen; choline, which marks an area of rapidly dividing or growing cells; and creatine, which serves as an internal control molecule.
The U-M team showed that the ratios of choline to creatine, of NAA to creatine, and of choline to NAA, were significantly different in normal tissue, versus recurrent tumor or treatment-altered tissue. For example, the choline/creatine ratio for recurrent tumor tissue was 2.30, while normal tissue had a ratio of 1.02 and tissue with treatment-related changes was 1.56.
They also showed they were able to avoid one of the biggest pitfalls of 2D CSI MRS, an effect known as "susceptibility artifact" that occurs when bone, sinuses or other areas of low or high density are near the region being scanned.
Multi-voxel, 2D CSI MRS is more accurate than the single-voxel form because the chemical scan is made from two different angles, allowing multiple small areas of tissue to be analyzed individually. Single-voxel scans are created by averaging the chemical signatures from a broader area, which means that tumor tissue and nearby dying tissue might be considered together. Multi-voxel scanning also allows radiologists to look at the tissue in a variety of locations inside and outside the original tumor area -- possibly revealing new tumor growth that isn't yet visible on a regular MRI.
"All in all, the likelihood of picking up a recurrence is greater, and the chance that we'll make an accurate, specific diagnosis is much higher, perhaps around 90 percent," says Weybright. "That means a patient can start a new round of treatment much earlier if the cancer is coming back, or avoid biopsies and unnecessary treatment, if the tissue is just inflamed or dying."