Description
Purpose
Transcranial Sonography (TCS) is a non-invasive diagnostic tool for evaluation of cerebral arteries and brain parenchyma. Besides having non-uniform resolution and inferior visual appearance, TCS works only in persons with sufficient trans-temporal acoustic bone window, which limits the diagnostic potential and prevents accurate imaging of the entire brain.
Compared to TCS, transcranial ultrasound tomography does not depend on the existence of trans-temporal acoustic windows in the patient’s skull and provides uniformly high resolution over the entire brain image.
Approach
A 256-element circular ultrasound tomography system originally developed for breast cancer imaging was used in this study. A custom cylindrical head phantom, 200 mm in diameter, was engineered to fit tightly into the system’s ultrasound ring array probe. The outer shell of the phantom simulates an adult human skull of realistic thickness and porosity. The phantom’s brain tissue model is homogeneous and contains hyper- and hypoechogenic foreign inclusions with sound speed and attenuation values of typical cranial lesions.
The CT images of the head phantom were converted into cross-sectional sound speed maps and served as input models for a finite difference-based acoustic field simulator. The latter numerically propagates ultrasound waves though the head phantom, naturally taking into account the refraction, scattering, and attenuation by the skull bone. The acoustic delays from each array element to each image pixel, provided by the numerical model, were used to correct phase aberrations in the recorded signals.
Results
The tomographic cross-sectional images of the head phantom reconstructed from experimental ultrasound data in water provide accurate size, shape, and positional resolution of embedded lesions. The relative merits and deficiencies of the reflection, attenuation, and sound speed imaging modes with respect to targets of different echogenicity are discussed.
Conclusions
The technical feasibility of ultrasound tomography imaging of strongly and weakly scattering brain lesions through a thick, curved, porous skull bone phantom has been experimentally demonstrated. The detrimental effects of skull-induced phase aberration and insufficient signal strength can be effectively reduced by implementing model-based adaptive focusing, thus opening a possibility for anatomically correct ultrasound brain imaging through thick skull bones.
