Percutaneous interstitial microwave thermoablation of locally recurrent prostate carcinoma was continually guided with magnetic resonance (MR) imaging. Phase images and data were obtained with a rapid gradient-echo technique and were used to derive tissue temperature change on the basis of proton-resonance shift. Thermally devitalized regions correlated well with the phase image findings. MR imaging-derived temperatures were linearly related to the fluoroptic tissue temperatures. MR imaging can be used to guide thermoablation.

An analytical model of tissue relaxation and restricted diffusion in human blood is presented. The blood tissue model is composed of three different compartments: red blood cells, plasma, and macromolecular protons. The relaxation rate constants and free diffusion coefficients of intracellular and extracellular water may differ. Analytical formulas for signal loss due to relaxation and diffusion in the Carr-Purcell Meiboom-Gill and pulsed-field-gradient multispin echo experiments for this tissue model are derived. The model is fitted to the experimental data for human blood with various concentrations of Gadolinium contrast agent. The obtained model parameters are realistic. The validity and sensitivity of the model are also discussed.

Gated SPECT (GSPECT) perfusion imaging has been increasing in popularity both with 99Tc(m) agents and 201Tl. However, both higher activities than administered in the UK and multi-headed cameras are often used. The aim of this study was to assess GSPECT imaging using lower activities of 201Tl with a single-headed camera. Seventy patients underwent stress and redistribution GSPECT imaging after a mean injected activity of 62 +/- 7 MBq 201Tl. These patients also underwent radionuclide ventriculography (RNVG) imaging. The Cedars Sinai Quantitative Gated SPECT (QGS) package was used to calculate left ventricular ejection fraction (LVEF) from the GSPECT studies. Comparison of ejection fractions calculated using GSPECT with those calculated using RNVG yielded a correlation coefficient of 0.70 for the stress studies and 0.71 for the redistribution studies. The width of the mean 95% prediction interval ranged from 22 to 74 percentage points for the stress studies and 22 to 86 percentage points for the redistribution studies. Ejection fractions calculated from stress and redistribution GSPECT studies showed a correlation of 0.80 with a mean 95% prediction interval of 42.6 +/- 0.4 percentage points. In conclusion, left ventricular ejection fractions calculated using the QGS algorithm from 201Tl GSPECT studies are inadequate for use in clinical practice.