Imaging the brain with ultrasound full-waveform inversion

Rapid brain imaging is central to the diagnosis and treatment of acute neurological conditions - for example stroke or head trauma. Existing imaging methods require large, immobile, high-power instruments that are near-impossible to deploy outside specialized environments, leading to unnecessarily delayed diagnosis and treatment, and consequent increased disability and higher fatality rates. This project will create a device that can be simply and rapidly applied to any patient, any time, any place, exploiting advances that have already revolutionised imaging in geophysics. We will image the brain using ultrasound waves, transmitted across the head, applying advanced computer modelling to remove the distorting effects of the skull, thereby enabling high-resolution high-contrast imaging of the brain unachievable by conventional ultrasound.

The petroleum industry has spent large sums developing advanced geophysical algorithms to image oil and gas deposits in three dimensions. Foremost among these is "full-waveform inversion" (FWI), a computationally intensive technique in which accurate modelling of soundwave propagation through a three-dimensional object is used to recover the detailed internal properties of that object. This project will adapt and transfer that technology across disciplines so that it can be applied directly for medical imaging of the brain, leading to cheaper, faster, more-accurate clinical diagnosis and treatment.

The main existing technologies used in three-dimensional medical imaging are magnetic resonance imaging (MRI), x-ray computed tomography (CT), and pulse-echo ultrasound. MRI is high resolution and high accuracy but is time consuming, expensive and immobile; it cannot be applied safely without a preliminary detailed investigation to ensure the absence of ferromagnetic bodies within any new patient. X-ray CT is cheaper and faster, but it is typically lower resolution than MRI, with poor soft-tissue contrast, and it uses harmful ionising radiation. Conventional pulse-echo ultrasound is cheap, fast, portable and universally safe, but it uses high frequencies that have limited penetration, and that are especially attenuated and distorted by the bones of the skull. Consequently, existing ultrasound technology is unable to image the adult brain successfully within an intact human skull.

Ultrasound at frequencies below those normally used for imaging does however have the penetration required to travel right across the head. Full-waveform inversion is able to produce accurate high-resolution images using lower-frequency data than is possible using conventional techniques; FWI is also able to compensate accurately for all the distortions generated by the skull. Consequently, the combination of low-frequency transmitted ultrasound with full-waveform inversion is able to produce well-resolved accurate images of the entire human brain. The potential of this approach has already been demonstrated in computer and laboratory simulations; this project now seeks to replicate that success in the laboratory on a live human subject.

Safe, fast, quantitative, universally applicable, deployable continuously, and above all portable by paramedics, our device and our approach aim to revolutionise brain imaging, in health and disease. The technology has particular relevance to stroke - globally the second-commonest cause of premature death and a major, growing cause of adult disability - and to brain imaging in resource-limited and inaccessible environments.