Published online 17 October 2008 | Nature | doi:10.1038/news.2008.1178


Non-invasive thermometer checks tissue hot spots

Magnetic resonance imaging technique could help to target heat therapy for cancer.

MRIMagnetic resonance imaging could help to monitor internal body temperatures.Punchstock

Doctors could soon get precise temperature readings from inside the body, without using probes or needles, thanks to a magnetic resonance imaging (MRI) technique. Its inventors believe that it could improve the viability of medical therapies that involve heating tissue.

Heat has been touted as an important weapon in the war against cancer1. Heating cancer cells may make them more susceptible to radiation treatment, and help certain cancer drugs to work better. However, these thermal therapies will need to be monitored carefully to check that the heat 'dose' has reached its intended target, and to prevent over-cooking. Ideally, the temperature check should be non-invasive and cover a broad area, not just a few points.

Existing MRI techniques can measure only changes in temperature in vivo, and do not give an absolute value. The accuracy of measurements also suffers when the magnetic field varies over the sample being examined — as is typically the case for human bodies, with a mixture of tissue types.

MRI works by using radio-frequency pulses to nudge the nuclei of hydrogen atoms, found in water and fat, out of alignment with the scanner's strong magnetic field. As the protons return to their original positions they emit a radio-frequency signal of their own, which is detected and used to construct an image of the tissue.

A team of scientists at Duke University in Durham, North Carolina, has now modified these radio-frequency pulses to produce an MRI thermometer that is five to ten times more accurate than the best possible alternative. They outline their method, termed HOT thermometry, in this week's Science2.

Follow the fat

Established methods of monitoring temperature with MRI rely on the fact that as tissue temperature changes, water molecules vibrate at different frequencies, which affects their radio-frequency signal. But this only gives a relative measure of temperature.

This latest work uses a particular sequence of radio-frequency pulses to look at fat molecules as well as water molecules, and compares their resonant frequencies. The difference in the frequencies corresponds directly to the absolute tissue temperature, explains chemist Warren Warren at Duke University. "For example, if we find that this difference in a certain scanner magnet is 950 Hz, that would tell us that the temperature in that region is 40.2 °C." Comparing signals from different molecules also allows them to clean up uncertainties caused by non-uniform fields.

The team tested their technique on live, obese mice, selected because the levels of fat and water in their tissue are similar to those in the normal human breast. The mice were heated by contact with a warm water tube and scanned repeatedly as their internal temperature rose from 28.6 °C to 39 °C.

Chrit Moonen, director of the Laboratory for Molecular and Functional Imaging at the University Victor Segalen in Bordeaux, France, says that the method presents a solution to one of the main problems of temperature mapping.

"The main benefit is that it allows measurements of absolute temperature independent of magnetic field inhomogeneity," he says. "But the low spatial and temporal resolution may make it difficult to use this technique to guide therapy." He suggests that further analysis is needed of the technique's sensitivity to motion, and of its precision when measuring temperature in tissues with a very low concentration of fat molecules.

MRI scanners used for clinical work at Duke University are now being reprogrammed with the radio-frequency sequences needed for HOT thermometry. The technique will then be incorporated into an ongoing clinical trial that is investigating targeted hyperthermia as a treatment for breast cancer. "My expectation is that we are six months, at most, away from human trials," Warren says. 

  • References

    1. van der Zee, J. Ann. Oncol. 13, 1173–1184 (2002).
    2. Galiana, G., Branca, R. T., Jenista, E. R. & Warren, W. S. Science 322, 421-424 (2008).
Commenting is now closed.