UW Radiology


A fundamental tenet underlying the Metabolic Spectroscopy Laboratory’s approach is that current gaps in clinical tools are the result of deficiencies in our current understanding of the basic science.  Our research focuses on generating new insights from basic research in order to develop new spectroscopic methods and tools that may ultimately have clinical applications.

Magnetic resonance spectroscopy (MRS) and optical spectroscopy (OS) provide two of the most powerful tools for non-invasive investigation of tissue properties and metabolism.  The potential application in clinical studies is vast but presently limited by lack of sufficient understanding of the basic cellular and tissue properties relevant to disease and by inadequate components of clinical MR instruments and FDA approved optical devices.  Four primary avenues of investigation utilizing MRS have been established:

  1.  Identifying the presence of specific metabolites in a tissue.
  2. Quantifying the content, and concentration, of specific metabolites.
  3. Measuring metabolic flux at a physiologically relevant time scale, seconds.
  4. Chemical shift imaging and spectroscopic imaging (CSI or SI) for “molecular imaging”


MRS and OS share the first three of the features listed above.  However, MRS methods for CSA or SI implementation are established, but the problem of imaging reconstruction of light scattered from tissue using OS has not yet been solved well enough to attain useful spatial resolution.


  1. Identifying the presence of specific metabolites in a tissue

MR and optical spectroscopy are both exquisitely sensitive and specific for chemical and molecular studies.  MR methods yields information about the chemical and physical status of a region of tissue within a living person, a chemical fingerprint unique to tissues and pathology. The power of MRS is based on the fact that nuclei that can be detected by MR (see Table).  When present in different compounds the nuclear spins have distinguishable peaks in a MR spectrum.  Thus in 13C spectroscopy of glycogen specific carbons in the glucosyl backbone can be measured, and in 31P spectroscopy ATP is readily distinguished from its product, inorganic phosphate, and its main cellular buffer, phosphorylcreatine or PCr.



Freq at 10 T (MHz)

Natural abundance


Relative sensitivity

Biological sensitivity

















6.6 x 10-2

6.6 x 10-2





9.3 x 10-2

9.3 x 10-2





1.6 x 10-2

1.8 x 10-4





1 x 10-3

4 x 10-6


The table gives the isotope name.  The spin state refers to the quantum mechanical separation of the spins in a magnetic field.  Spin 1/2 means there are only two possible spin states, either aligned with the main magnetic field (low Boltzman energy) or opposite the main magnetic field (high energy).  Natural abundance refers to the fraction of the MR sensitive nucleus that is present in the terrestrial environment. Essentially all nuclei of hydrogen, fluorine, phosphorus, and sodium in the environment have NMR detectable spin states.  In contrast only 1.1% of carbon isotopes are 31C; the most abundant isotope is 12C which is NMR silent.  Similarly only 0.37% of nitrogen is the MR sensitive isotope, 15N, whereas the most abundant isotope is 14N.  

Optical spectroscopy requires that light penetrates the tissue of interest and the scattered light is sufficiently intense to be detected by the spectrograph.  This difference is critical. Photons are highly scattered in living tissue so that light in the near infra-red region of the spectrum penetrates on the order of several to ten centimeters, whereas MR has no such limitation (except at extremely high magnetic fields).  However ultrasensitive optical detectors are available that render investigation of muscle and brain chromophores (hemoglobin, myoglobin, cytochrome).


  1. Quantifying the content and concentration of specific metabolites 

Having identified the presence of a specific chemical compound or metabolite in a tissue of a human or animal, MR spectra can be quantified because the area within a given peak is proportional to the concentration under appropriate situations.  To enable accurate quantification of concentration within a tissue, a number of strict experimental requirements must be achieved.  The details are beyond the scope of this introduction, but they include:

  • parameters of the MR signal acquisition,
  • how well known is the volume of tissue from which the spectrum was obtained,      
  • stability and reproducibility of the entire MR apparatus (from amplifiers to coils),
  • how well do the properties of the solution used to calibrate the spectrum match the properties of the tissue.


Optical spectra also are quantitative.  Most scientists are familiar with the Beer-Lambert Law wherein the intensity of the absorption spectrum is proportional to the concentration of the compound present.  Quantification of OS has related but distinct issues relative to MRS.  Here the relavant parameters of the OS acquisition are the tissue scattering properties.  Volume from which the spectra are obtained is also difficult to quantify.  Usually stability of the apparatus is not a problem with the exception of the light source which requires special care to maintain a stable output within a few parts in ~ 105


  1. Measuring metabolic flux at physiologically relevant time scale, seconds 

OS can be obtained repeatedly with time resolution <<1 sec, whereas MRS typically has time resolution of 1 sec under conditions of high signal to noise.  Because both MRS and OS are non-invasive, time resolutions of this order allow exploration of a great number of interesting physiological and clinical problems.  Usually (but see below for a unique feature of MRS) some sort of perturbation is needed to alter the physiological state from one steady state to another.  Activation of cellular metabolism is a well-used approach.  Then when the chemical reactions are well defined and separated from each other, the time course of a spectral intensity can be quantified in terms of the underlying chemical and metabolic flux.  For example in muscle exercise, the time course of ATP utilization can be much faster than its regeneration.  Thus the initial decrease in PCr or ATP (and increase in the product Pi) can be interpreted as an ATPase flux.

MRS has a unique capability; it can measure the unidirectional fluxes in an equilibrium reaction occurring within the tissue.  No other method can do this.  Indeed making this measurement is difficult even in a chemical analysis of pure solutions!  A description of this unique method is beyond the scope of this introduction, but essentially is is possible to place a temporary magnetic label on one of atoms of a metabolite of interest involved in a chemical exchange with another metabolite.  From the lifetime of that label, one can deduce that unidirectional flux.  The temporary label is similarly put on the product molecule and the experiment is repeated to obtain the other unidirectional flux.  The ratio of the fluxes also gives the equilibrium constant in the living cells.


  1. Chemical shift imaging and spectroscopic imaging (CSI or SI) for “molecular imaging”

Information on chemical composition or flux is often desired on a localized region of an organ or body part.  One simple way to attain a degree of localized information is by the use of small coils, that by definition, obtain MR information only from the region of influence of the coil.  For example a circular coil can be placed on a head or limb, and information can be obtained from approximately one diameter sphere from the surface of the coil within the leg muscle or brain parenchyma.  The use of these “surface coils” is common and gives anatomically relavant information.

An alternative to get anatomically relevant MRS is to combine the techniques of MRS with MR imaging.  Unfortunately the duration needed for acquiring this type of information is very long.  The reason is that MR spectroscopy is relatively insensitive and significant signal averaging is needed, even for time resolutions of a few seconds.  The requirements of MR imaging are determined by the resolution desired and by encoding spatial coordinates by magnetic gradients in each of three directions.  For crude anatomical resolution (8 for a given field of view), 83 replications of each MRS acquisition is needed.  Higher resolution, e.g. 64 resolved elements, obviously would require unacceptably long examinations. This example would require 512 times a few seconds or a study lasting more than 15 minutes.  Thus there is a trade-off between time resolution and anatomical resolution.  Longer acquisitions by chemical shift imaging and spectroscopic imaging are used to characterize a stable state of metabolism throughout an organ, but these are obviously not useful when high time resolution is needed.

Much research effort is spent on methods to enhance the speed of chemical shift imaging and spectroscopic imaging.  Likewise vast efforts are spent to analyze the light scattered from a region of the body to obtain spatially resolved optical images.  The “holy grail” can be described as a localized image of deoxy- and oxy-hemoglobin concentration so that oxygen partial pressure can be calculated in each volume element in brain or muscle by either MRS or OS.  To data the only approach to this goal that has found use is functional magnetic resonance imaging (fMRI).   It is a non-quantitative surrogate measure of tissue oxygenation.