Nicholas Demartini, M.D.

MRI guided focused ultrasound (MRIFUS) is a developing method for non-invasive treatment of diverse pathologies ranging from uterine fibroids, to bone metastases and to brain tumors. While many of the proposed applications for FUS are made relatively straight forward by good acoustic windows, adaptation of MRIFUS to the brain has been challenging due to the difficulties presented by sonication through the skull.

The skull provides two potential difficulties. First of all bone has a high degree of ultrasound absorption.  This can lead to unintended heating and damage of the skull when attempting sonication of an intracranial lesion. This obstacle has been largely overcome using large phased array transducers which spread the ultrasound energy over a wide surface area, and water baths which help cool the skull during the procedure.

The second difficulty with sonication through the skull is achieving adequate focusing of the US beam despite the absorption, reflection and refraction which occurs as the sound waves pass through the skull. Precise focusing is essential to minimize collateral damage to non diseased tissues and to achieve temperatures capable of inducing coagulative necrosis at acceptable power deposition levels. While early attempts to overcome this problem resorted to invasive methods such as craniotomy and intracranial sonographic mirror placement, newer studies have demonstrated that effective focusing can be achieved through the skull using non-invasive back projection algorithms. These methods use MRI or CT to create a map of the skull. They then simulate an ultrasound source emanating from the target area outward and model the effect of the skull on the US waves as they pass outward through the skull. The algorithms then back-project this information to focus an external ultrasound source to the target area.

As mentioned, both MRI and CT have been used to create the skull maps necessary for focusing. While CT models have taken into account the complex heterogeneity of the calvarium, previous MRI models have assumed a homogenous calvarial structure. The more robust modeling based on CT data, not surprisingly, provides better focusing. Therefore to provide the most precise focusing for MRIFUS, a patient must have a CT performed prior to the MRIFUS with subsequent co registration of the CT and MRI data. This extra step exposes the patient to ionizing radiation and adds the time of the CT acquisition and co-registration to the total procedure time. This delay could be detrimental if MRIFUS was used to treat patients with acute conditions such as stroke where the treatment window is limited.

We propose to utilize newly developed MRI sequences to make precise measurements of the calvarial structure which correlate with those measurements made by CT. This would be the first step in creating more detailed calvarial maps with MRI which could eliminate the need for CT prior to MRIFUS.