iZQC Thermometry

Most MR thermometry methods are easily distorted by variations in the magnetic field (which are unavoidable, and often fluctuating in vivo and in other interesting systems). We can correct for these local variations by applying our two-spin methods and using one of the spins as a physical reference and correction factor. The resulting signal is implicitly corrected for variations in the magnetic field, and can be an order of magnitude more accurate than conventional methods, both in vitro and in vivo.

It is well known to NMR spectroscopists that the chemical shift of water depends on temperature, and that this phenomenon can be the basis for an NMR thermometer. The 0.01ppm/°C shift is easily observed in a well shimmed sample under high field. In vivo, however, the water peak itself is often several ppm wide and unstable (from susceptibility gradients and motion), thus burying the small shift that reflects temperature. The inadequacy of conventional MR thermography limits its usefulness in many applications, and is the foremost obstacle impeding hyperthermic cancer therapy. These promising therapies selectively heat tumors to bring about enhanced drug delivery (up to 30-fold), inhibited cell repair, and even cell death. Unfortunately, effective treatment cannot be guided by conventional MR temperature maps because they provide only a relative measurement of temperature and are easily distorted. We present a physically different method based on the detection of water-fat iZQCs (intermolecular Zero Quantum Coherences) which produce a far more accurate map of absolute temperatures. With novel modifications to the HOMOGENIZED sequence, we isolate these coherences and generate accurate temperature images. Data acquired with this sequence demonstrates, both in vitro and in vivo, that iZQC temperature maps can be an order of magnitude more accurate than conventional methods.

As exploited previously in our MRS applications, iZQCs are insensitive to both static and drifting changes of the local magnetic field. Though the local field varies over time or space, the difference frequency (ω_H2O-ω_FAT) between two protons in nearly the same place (<100 µm apart) depends only on chemical shifts. To isolate these water-fat coherences, two novel filtering schemes were developed to dephase and then cancel all signals except those arising from inequivalent spin pairs. For dephasing, a selective 180 pulse on water is applied to cause a transition from zero to double quantum coherence, but only for iZQCs between water and an off resonance spin:
Coherences between two water spins (or any other like spins) remain iZQC, and can be dephased by a double quantum gradient filter. Furthermore, tailored excitation pulses constructively amplify signal from water-fat iZQCs, but cause iZQCs between like spin pairs to cancel. In addition, the method can separately acquire two different types of iZQCs in each scan for added resilience against motion artifacts. The modified HOMOGENIZED sequence only transfers coherences between water and off-resonance spins into observable signal, so we refer to it as Homogenized with Off-resonance Tranfer, or the HOT sequence. A water-fat phantom allowed to equilibrate for one hour at each temperature was mapped by conventional and iZQC thermography sequences. The conventional images wrongly report large temperature gradients across the sample, even though those temperatures are calculated with high precision. With iZQC thermography, the precision of the measurements, though somewhat less, falls into an acceptable range, and the temperature maps correctly reflect the uniform heating of the phantom. These measurements provide an accurate measurement of temperature on an absolute scale.

Figure 1 illustrates the advantages of HOT thermography with a water/fat phantom (heavy whipping cream) at three uniform temperatures. The iZQC images accurately reflect frequency shifts from temperature, but are not distorted by other factors. In contrast, the standard imaging method shows a wide range of proton frequency shifts which cannot reasonably be attributed to temperature variations. Since the phantom was maintained at uniform temperature using a thermoprobe controlled heater, it is clear that the wide range of frequencies detected in the conventional image reflect local field distortions and not temperature. The method has also been demonstrated in vivo, where it was shown that iZQC thermography can be an order of magnitude more accurate than conventional methods. (Figure 2) These results directly verify the potential of iZQC thermography in vivo.