Abstract:  Thermal  tissue damage predicted by MRI-derived  thermal dosimetry  acquired during  focused ultrasound  surgery  in uterine  fibroids was compared to that observed in contrast-enhanced T1W images. A method to estimate the effects of thermal build-up is described. Preliminary tests in 14 fibroids indicate that the thermal dose under-predicted the treatment response observed in contrast-enhanced T1-weighted imaging on average by 43±20%, even when the effects of thermal buildup were estimated.

Introduction:  Animal studies have demonstrated the ability of the thermal dose calculated from MR thermometry in predicting the onset and extent of tissue damage during thermal therapies (1,2). Recently, in a clinical trial of MRI-guided focused ultrasound surgery in uterine fibroids, it was reported that the thermal dose prediction for tissue damage was generally smaller than the observed treatment response (3). The purpose of this study was to quantify the under-prediction by the thermal dose of the treatment response observed in contrast-enhanced MRI, and to estimate the effects of thermal build-up in the focal plane on this prediction in a sample of patients treated at our institution.

Methods:  A large-scale focused ultrasound phased array system integrated with a clinical MRI scanner (Exablate2000, InSightec, Haifa Israel) was used to treat uterine fibroids. The treatments consisted of multiple sonications at different locations in the fibroids. Between each sonication, the tissue was allowed to cool for approximately 2-3 min. During each sonication, a fast spoiled gradient echo sequence (FSPGR) was used to generate a time series of temperature maps (4), which were used to estimate the thermal dose (5). A dose threshold of 240 equivalent min at 43°C was used. The images were acquired in a single plane. Over the course of each treatment, the imaging orientation was varied between planes perpendicular  to  the ultrasound beam at  the  focal plane and parallel  to  the direction of  the ultrasound beam.  Immediately after the treatments, dynamic contrast enhanced T1-weighted FSPGR images were acquired. The extent of the thermal dose in the focal plane perpendicular to the ultrasound beam (coronal) was compared to the non-enhancing region in a corresponding T1W image. To estimate the thermal dose in the coronal plane from axial or sagittal images, it was assumed that the temperature distribution was radially symmetric. The temperature distribution at each time point was fit to a Gaussian distribution. The temperature decay at the focus and the time dependence of the Gaussian width were then extrapolated to later times. This extrapolated temperature distribution could be then added to that of later sonications. Treatments of 14 fibroids in 10 patients were tested.

Results:  In most  cases,  the  shapes of  the non-enhancing  regions  in  the  fibroid  and  the dose prediction were  similar, but  the dose prediction was smaller than the non-enhancing  region.  In two cases,  the  fibroid was almost completely non-enhancing even  though only a portion of the fibroid was treated, indicating an obvious vascular effect. Figure 1 shows the agreement for all cases tested.

Estimation of the thermal build-up in the focal plane increased the dose prediction slightly (circles), but not enough to agree with the nonenhancing region. The effects of thermal build-up increased the dose-predicted area by 7±2% on average (range: 4-11%). Even taking the build-up into account, the non-enhancing area was 43±20% (range: 8-73%) greater than that predicted by the dose.

Discussion:  These preliminary results indicate that other effects may be largely responsible for the disagreement between the dose and the treatment response. Further investigations are necessary to confirm this result since the thermal build-up was only investigated in the focal plane, and it is expected that thermal build-up effects should be greater along the direction of the ultrasound beam. Development of a method to estimate the effects along this direction is a topic of current investigation. The technique described here for estimating the effects might be useful as a tool to use during focused ultrasound treatments to take the thermal build-up into account.