MO-G-18C-09: Thermal Characterization and Initial Imaging Results From the Diffusive Quantitative Imaging Phantom (DQIP)




The goals of this project were a) to characterize thermal relaxation of the phantom and b) to demonstrate correlations between temperature, signal-to-noise ratio (SNR) and diffusion tensor imaging (DTI) metrics. A site-based clinical DTI protocol was used in this study. The hypothesis is that thermal stability of the phantom will be adequate such that temperature monitoring during a scan may be neglected.


The Diffusive Quantitative Imaging Phantom (DQIP) is a prototype phantom consisting of fifteen cylindrical compartments containing capillary arrays, encased in a larger compartment. Thermal stability of the phantom was established using a cork container and a Peltier incubator. After the phantom was cooled or heated, thermal relaxation was measured as the phantom was allowed to return to room temperature, with and without cork. A clinical-grade DTI protocol was used to collect scan and temperaturedependent data from the phantom over ten days. SNR and thermal dependences of fractional anisotropy (FA) and apparent diffusion coefficient (ADC) were characterized by scanning a heated or cooled phantom over several hours.


The phantom temperature was stable within ±0.08°C/hour per degree difference from the scan room using the cork enclosure, compared to ±0.3°C/hour per degree difference without the cork enclosure. During scanning, temperature variation was ∼ 0.1°C/hour. The dependences of (FA, ADC) on (SNR, temperature) were characterized for the heating and cooling experiments. However, for the range of temperature (1.2°C) and SNR (max: ±77%) variations in a particular compartment over all daily measurements, no significant improvement in total variation was achieved after regressing out temperature and SNR, although one compartment showed significant decline after regression.


Thermal stability of the DQIP phantom and incubator system is sufficient for neglecting temperature variations during scanning. For a standard clinical protocol, SNR dependence of FA and ADC are not sufficient to warrant correction.

Schott Glass North America and The Phantom Laboratory have donated materials and personnel time to this project.