Initial in vivo experience with EIT as a thermal estimator during hyperthermia.

Thermal imaging experiments using electrical impedance tomography (EIT) have been conducted during hyperthermia treatments delivered to two human patients and one animal subject. Coplanar and circumferential arrays of 16 and 32 tin-plated copper electrodes etched on a 0.005" polyimide sheet were used to inject 12.5 KHz current patterns of increasing sinusoidal spatial frequencies and subsequent potential distributions were recorded at each electrode site. Image reconstruction was achieved with a finite element method and difference images of conductivity changes during the course of treatment were formed. An assumed linear relationship (2%/degree C increase) between tissue impedance change and temperature change was used to produce thermal images of the treatment field in patients whereas an empirically measured nonlinear relationship obtained from excised tissue samples was applied retrospectively in the animal subject case. Reconstructed conductivity changes are shown to be possible given electrical data measured in vivo during hyperthermia delivery with conventional equipment (spiral microstrip applicator at 433 MHz). These correlated well with direct temperature measurements and demonstrated quantitative levels of agreement to the extent that estimated temperature accuracies were approximately 1.5 degrees C; although large errors (> 5 degrees C) did exist. This work suggests that EIT is a potentially useful tool for hyperthermia treatment monitoring and assessment. The relationship between tissue impedance and temperature is complex and confounds the ability to make simple correlations between conductivity and temperature changes. Further, study is required to discern whether this will ultimately limit EIT as a thermal estimator or whether it will lead to more fundamental uses of impedance as an indicator of thermal effect.