Description

Intracranial pressure (ICP) is routinely monitored in patients suffering from traumatic brain injury (TBI). Raised ICP can result in compression of the cerebral vasculature and subsequent reduction in oxygen and nutrient delivery to the brain leading to significant morbidity and mortality. In fact, raised ICP is the most common cause of death in patients with severe TBI.

Standard ICP monitoring requires insertion of a cranial bolt into the skull through which an electrical transducer is inserted. Alternatively, an intra-ventricular catheter is inserted through a burr hole. Both of these monitoring methods are associated with risks including haemorrhage and infection, as well as delay in establishing emergency monitoring and limiting it to hospitals that have neurosurgery.

There has been much research in recent years to find a method for measuring intracranial pressure noninvasively (nICP), including measurement of pressure in the retinal veins, measurement of eardrum displacement, transcranial Doppler ultrasonography and imaging-based solutions. These methods all require considerable user intervention and are non-continuous.

This project aims to collect cerebral photoplethysmogram signals and concurrent invasive ICP measurements from patients with traumatic brain injury to develop Crainio machine learning (ML) algorithms. The core intellectual property (IP) of this continuous external monitoring ICP system was originally developed by academics in the lab of Professor Kyriacou at City, University of London. Crainio is a spin-out company that was created to industrialise and commercialise this research on an exclusive basis.

The device comprises a forehead-mounted sensor containing infrared light sources that can illuminate the deep brain tissue of the frontal lobe. Photodetectors in the sensor detect the backscattered light, which is modulated by pulsation of the cerebral arteries. A control unit processes the backscattered light (called the photoplethysmogram, PPG) and transmits it to a computer device to train ML models that estimate an absolute value of ICP.

The basic science behind this method for measuring ICP is that changes in the extramural arterial pressure affect the morphology of the recorded optical pulse, so analysis of the acquired signal using an appropriate algorithm will enable calculation of nICP. The reported nICP will provide screening at the triage stage, indicating the need for imaging or rapid intervention (such as haematoma evacuation) and guide head injury management, notably ICP-targeted treatment regimes. Ultimately this could lead to significant improvements in secondary injury-related mortality, length of hospital stay and reduced post-trauma disability.

This feasibility study aims to collect the clinical data with which to train the nICP algorithms to the point that they can detect raised intracranial pressure (ICP\>20 mmHg) with sufficient sensitivity and specificity that Crainio device can be regulated for clinical use.