The original name nuclear magnetic resonance conceals a bit of history. The first users of this technology were chemists who employed the resonance of chemical nuclei in magnetic fields to create tracings of peaks and valleys — spectra — from which they could determine the composition of materials. Spectroscopy (MRS) was the original application of magnetic resonance technology.

However, spectroscopy did not remain the focus once it was demonstrated that the data could be converted into pictures of human tissues. Equipment manufacturers multiplied, and for many years, they ignored spectroscopy in favor of imaging.

Interest in MRS revived because it is an analytical technique with some very desirable and unusual properties. As Robert E. Lenkinski, PhD, of the University of Pennsylvania, Philadelphia, has said, “Clinical spectroscopy is a sampling method that will tell you the chemical composition of a tissue noninvasively and nondestructively.” In a sense, he pointed out, MRS is a more refined and elaborate version of something radiologists do all the time: use the chemical difference between fat and water to make diagnoses. Today, most new MRI scanners are equipped to carry out the most basic MRS technique, namely single-voxel spectroscopy.

This article focuses on the most popular clinical use of MRS today: studies of the brain. Decisions in Axis Imaging News spoke with two leaders in the field: Walter A. Hall, MD, associate professor of neurosurgery at the University of Minnesota in Minneapolis, which has installed a high-field interventional MRI scanner in one operating room; and Robert Zimmerman, MD, director of MRI and neuroradiology at Children’s Hospital in Philadelphia, a premier site for MRI studies in pediatric patients. Decisions asked them how they use MRS.


“The main indication for MRS at present is distinguishing actively growing tumor from radiation injury,” Hall explains. “Many patients with a history of brain tumor present with new enhancing masses or enlargement of a known mass. Most of them have already had radiation treatment, and the question is whether the tumor is progressing despite therapy or the changes are a radiation effect and indicate that the patient’s steroid dose needs to be increased.

” At present, I do not think that MRS can make this diagnosis definitively,” Hall continues. “We have performed biopsies in about 30 of these patients. We use both single-voxel spectroscopy, which is conventionally available, and chemical shift imaging, which I think will be more useful. We biopsy areas with elevated choline concentrations and then have the pathologist tell us, ‘Yes, it is tumor,’ or ‘No, it is not.’ So far, MRS has been correct in more than 90% of the cases. There were only a couple of patients in whom MRS said tumor and the tissue proved to be necrotic.”

Not all patients seen at the University of Minnesota for brain tumor surgery undergo MRS, Hall explains. “Whether we do an intraoperative study depends somewhat on the time available. There are some patients for whom we do not want to spend 1 or 2 hours for imaging, but whenever we have the opportunity to obtain MRS images, we do,” Hall says.

What to Look for When You Visit the Brain

Although 31 phosphorus spectra are sometimes obtained, most MRS studies of the brain are performed by proton spectroscopy with water suppression. In the simplest study, which can be done with most MRI scanners now being sold, a single small voxel is selected for study. Two-dimensional chemical shift imaging depicts a cross-sectional slice, and three-dimensional chemical shift imaging provides data from a stack of slices.

The data captured by the scanner may be presented in one of three forms. Quantitative data are obtained by calculating the area under the compound’s peak in the spectrum; the area is proportional to the concentration. The results are expressed as parts per million. Alternatively, one may use semiquantitative data or determine the ratio between the areas under the peaks of two compounds.

There are two types of MRS findings. First, the spectrum may reveal a metabolite that is not normally present, at least not in quantities that can be detected. Second, and more commonly, one sees an abnormal quantity of a metabolite that is usually present.

In studying the brain, some of the metabolites of greatest interest are N-acetyl aspartate (NAA), choline, and myo-inositol. The normal concentrations of some of these metabolites differ according to the region of the brain being examined and the age of the patient. Lactate can be detected if the concentration is abnormal. Occasionally, MRS of lipids is performed.

N-acetyl aspartate is a component of neurons and thus is abundant in the normal brain. A significant decline indicates neuronal death. Choline is important in cell membranes and in the myelinization of nerves. Abnormally high concentrations are found in tumors. Myo-inositol is made predominantly by glial cells and therefore is present in abnormal amounts in areas of gliosis. Lactate is a product of anaerobic energy metabolism. As the body prefers aerobic metabolism, the presence of large amounts of lactate is abnormal. Genetic diseases of the mitochondria, such as Leigh’s subacute necrotizing encephalomyelopathy, cause abnormally high brain lactate concentrations even when the amounts in the serum and cerebrospinal fluid are normal.

At Children’s Hospital in Philadelphia, MRS is a method of distinguishing among various pediatric brain lesions.

“We started out looking at posterior fossa tumors,” Zimmerman explains. “We recently showed that you can differentiate medulloblastoma, cerebellar astrocytoma, and ependymoma with reasonable success. With the help of a computer-savvy resident fellow, we wrote a program that uses basic clinical and imaging information and MRS data to determine tumor type. The algorithm has an accuracy of close to 95%. We also have used MRS for suprasellar and supratentorial lesions. Sometimes the surgical approach is determined by the findings: for example, whether a tumor is a pituitary adenoma or craniopharyngioma.”

Surgeons do not insist necessarily on an MRS study before they take the patient to the operating room. “If the patient is about to die, it does not matter what histologic type the tumor is,” Zimmerman notes. “Also, many of our patients have already had imaging studies elsewhere, and nobody wants to pay us to do MRS. So whether we do that study depends on how important it is to know the type of tumor preoperatively.

“Sometimes it is important. In medulloblastoma, gross total resection can increase the chance of therapeutic success from about 30% to 75%-80%. If you knew the tumor was a medulloblastoma before you began the operation, you probably would be willing to go further.”

At Children’s Hospital, a number of patients are seen in whom there is a question whether a detected mass is threatening.

“A child came in here about a month ago for a spine study,” Zimmerman reports. “One of the radiologists noted something abnormal near the top of the film that proved to be an asymptomatic tumor in the fourth ventricle.

“The surgeon now has to talk to the parents. This lesion could be low grade or benign, in which case they may want to wait and see what happens, because there are risks from surgery. On the other hand, the lesion could be something that will be all over the place in 6 weeks, when it will be too late for surgery. Many times, MRS studies resolve the issue.

Another area where MRS has contributed is one in which people often do not want the information. “The surgeon often believes that resection has been complete, but when we perform MRS, we sometimes see that even though the tissue that was removed was benign and there is no visible tumor remaining on conventional imaging, the white matter contains malignant tumor,” Zimmerman reports. “Such MRS findings virtually guarantee that with time, obvious tumor will appear in that area.”

Children’s Hospital is part of a new 10-member consortium organized by the National Institutes of Health to use MRS as a tool for judging the response of brain tumors to new chemotherapeutic agents. Presumably, the technique could also be used for early determination of the effectiveness of standard drugs, “but there has not been any significant interest in that application,” Zimmerman notes. “I think that is an application that ought to be investigated, but in the meantime, everyone is excited about the new trials.”

Children’s Hospital also has been using MRS to study metabolic diseases. “As you identify metabolites related to a disease, you can quantitate them in the brain,” Zimmerman explains. “The metabolite may be there because the patient is not adhering to the prescribed special diet, as we see, for example, with galactitol in patients with galactosemia. Alternatively, it may be a by-product of the genetic defect, such as branched-chain amino acids in maple syrup urine disease. So we can monitor treatment with MRS. The greatest problem is that we are dealing with the brain and, quite frankly, nobody understands all the metabolic pathways and chemicals that are being produced there.”


MRS also is finding diagnostic applications in metabolic disease. “We have children referred to us with nonspecific problems such as failure to thrive or developmental delay,” Zimmerman says. “The hope is that we can figure out what is wrong. We find abnormalities, and sometimes, we find diseases that have not previously been described. We published an article on one such disease last year, and I think there are other diseases out there waiting to be discovered.

“We also are learning more about previously known diseases. For example, we have found abnormal concentrations of myo-inositol in the brains of children with Down’s syndrome. This could be a consequence of the abnormal number of copies of some genes because of the trisomy, and it might be important in producing the effects of this condition.”

Children’s Hospital is installing an MRI system that will readily quantify brain metabolites. The MRS team has submitted a grant application for a study aimed at using the system to profile brain development in infants and young children. With such information, it will be easier to determine when a child is developmentally delayed and whether a delay is the result of metabolic disease or an in utero problem.

Another research use of MRS is in evaluating the progress of gene therapy. A trial in which Children’s Hospital is participating is attempting to correct the metabolic defect of Canavan’s disease, in which a deficiency of the enzyme aspartoacylase leads to an elevated concentration of N-acetyl aspartate (NAA), a low choline concentration, and failure of myelinization. A genetically engineered virus is being used in an attempt to persuade brain cells to produce the missing enzyme. MRS studies to measure NAA and choline show whether the virus has been taken up and expressed.


A billing code for MRS was introduced by the Health Care Financing Administration in January 1998, with reimbursement set at $540 per examination. “There is a billing code for MRS, but whether you collect is another issue,” Zimmerman notes wryly. “A couple of months after the code took effect, our billing people told me that they had billed for 20 spectroscopy studies, and we had been paid for one. Ironically, I received a call a few days ago from a large insurance company asking if we did spectroscopy studies in children. They said we were listed as the place for their clients to have MRS, but they did not know whether we did it.”

The situation reportedly is better in Minnesota. “We are reimbursed by HMOs and insurance companies if we use MRS on an outpatient basis to determine whether a lesion is likely to be a recurrent tumor,” Hall reports. “Some insurance companies do require prior authorization, but they just want to know that we plan to do the study to answer that specific question. A year and a half to 2 years ago, we had trouble getting reimbursed, but this is no longer true.”

Hall is not, however, billing for intraoperative MRS. “When we do MRS in the operating room, we bill only for the MRI scan that is done at the same time,” Hall explains. “We do not submit a separate spectroscopy bill.”


Several other uses of MRS in the brain are being used clinically or are in development. For example, the high acetate concentration in abscesses makes them easy to distinguish from tumors. Patients with AIDS, seizures, dementia, or ischemic stroke are among those who sometimes benefit from the findings of MRS studies. A new area is functional MRS, in which the technology is used to study the normal and abnormal chemical workings of the brain.

Indeed, enthusiasm for MRS studies may be outstripping the ability to make use of the data. “I often get faxes of spectra from people asking me, ‘What does this mean’?” Zimmerman reports.

Both Hall and Zimmerman believe that the use of MRS for studying brain tumors is only in its infancy. “Basically,” Hall says, “I think MRS will replace positron emission tomography for brain tumor studies.”

“MRS is one of many ways we can augment the information we obtain from routine imaging,” Zimmerman says. “I think this is the direction in which radiology is headed, and we need to be cognizant of what we can do with the new tools.” n


Further Reading

  1. Castillo M, ed. Proton MR spectroscopy of the brain. Neuroimaging Clin N Am 1998;8(4). [Includes 11 articles, one by Dr Zimmerman and an associate on the use of MRS for metabolic diseases.]
  2. Howe FA. Magnetic resonance spectroscopy in vivo. In: Markisz JA, Whalen JP, eds. Principles of MRI: Selected Topics. Stamford, Conn: Appleton & Lange; 1998: 17-107. [An overview of the technical aspects of MRS and of its uses in various sites in the body].

Judith Gunn Bronson, MS, is a contributing writer for Decisions in Axis Imaging News.