Page 138 - GTM-4-1

Global Translational Medicine MRgFUS sonication parameters prediction

A

B

Figure 3. Prediction accuracy indicators for linear and neural network models. (A) The linear model has a determination coefficient value of 0.71, while
the (B) neural network model has a determination coefficient value of 0.76. The observed temperature is on the x-axis, and the predicted temperature is
on the y-axis.

importance of skull bone ultrasonic conductivity coefficient
in MRgFUS treatment. Gagliardo et al. suggested that better
control of the procedure can be achieved by prospectively
determining ultrasound energy based on energy delivery
curves and skull bone ultrasonic conductivity coefficients
at each treatment step. However, they did not propose a
formula for calculating the heat energy.
Boutet et al. also investigated the role of SDR in the
42
effectiveness of MRgFUS for movement disorders in two
Figure 4. A diagram of the residual distribution density (difference cohorts. They noted that low SDR values can impede
between the predicted and achieved temperature) the passage of acoustic energy. In their study, 98 patients
who underwent MRgFUS thalamotomy (Cohort 1) were
side effects, overall tremor reduction was comparable. analyzed. While patients with lower SDRs needed higher
These findings suggest that MRgFUS is safe for patients energy levels for treatment, their overall clinical outcomes
with low SDR, although they may be at higher risk for were not significantly different. Furthermore, a study of
treatment failure and intraoperative discomfort. 163 emergency department patients (Cohort 2) showed
Gagliardo et al. described the relationships between that approximately one-third had low SDRs, and SDR was
23
sonication parameters, emphasizing their significant not associated with age or gender. This finding suggests
impact on MRgFUS treatment. They highlighted the that skull density does not vary predictably with these

Volume 4 Issue 1 (2025) 130 doi: 10.36922/gtm.5419