Comparison of computer simulations and clinical treatment results of magnetic resonance‐guided focused ultrasound surgery (MRgFUS) of uterine fibroids

PURPOSE: Magnetic resonance‐guided focused ultrasound surgery (MRgFUS) can be used to noninvasively treat symptomatic uterine fibroids by heating with focused ultrasound sonications while monitoring the temperature with magnetic resonance (MR) thermometry. While prior studies have compared focused ultrasound simulations to clinical results, studies involving uterine fibroids remain scarce. In our study, we perform such a comparison to assess the suitability of simulations for treatment planning. METHODS: Sonications (N = 67) were simulated retrospectively using acoustic and thermal models based on the Rayleigh integral and Pennes bioheat equation followed by MR‐thermometry simulation in seven patients who underwent MRgFUS treatment for uterine fibroids. The spatial accuracy of simulated focus location was assessed by evaluating displacements of the centers of mass of the thermal dose distributions between simulated and treatment MR thermometry slices. Temperature–time curves and sizes of 240 equivalent minutes at 43°C (240EM(43)) volumes between treatment and simulation were compared. RESULTS: The simulated focus location showed errors of 2.7 ± 4.1, −0.7 ± 2.0, and 1.3 ± 1.2 mm (mean ± SD) in the anterior–posterior, foot–head, and right–left directions for a fibroid absorption coefficient of 4.9 Np m(–1) MHz(–1) and perfusion parameter of 1.89 kg m(–3) s(–1). Linear regression of 240EM(43) volumes of 67 sonications of patient treatments and simulations utilizing these parameters yielded a slope of 1.04 and a correlation coefficient of 0.54. The temperature rise ratio of simulation to treatment near the end of sonication was 0.47 ± 0.22, 1.28 ± 0.60, and 1.49 ± 0.71 for 66 sonications simulated utilizing fibroid absorption coefficient of 1.2, 4.9, and 8.6 Np m(–1) MHz(–1), respectively, and the aforementioned perfusion value. The impact of perfusion on peak temperature rise is minimal between 1.89 and 10 kg m(–3) s(–1), but became more substantial when utilizing a value of 100 kg m(–3) s(–1). CONCLUSIONS: The results of this study suggest that perfusion, while in some cases having a substantial impact on thermal dose volumes, has less impact than ultrasound absorption for predicting peak temperature elevation at least when using perfusion parameter values up to 10 kg m(–3) s(–1) for this particular array geometry, frequencies, and tissue target which is good for clinicians to be aware of. The results suggest that simulations show promise in treatment planning, particularly in terms of spatial accuracy. However, in order to use simulations to predict temperature rise due to a sonication, knowledge of the patient‐specific tissue parameters, in particular the absorption coefficient is important. Currently, spatially varying patient‐specific tissue parameter values are not available during treatment, so simulations can only be used for planning purposes to estimate sonication performance on average.