TORONTO – Edmonton’s University of Alberta is lending its neuromotor research smarts to a wearable, machine learning driven sensors platform developed by Menlo Park, Calif., and Calgary, Alberta-based Protxx Inc. for better managing the future care of patients suffering from neurodegenerative conditions such as multiple sclerosis (MS). The “tail wind” for the partnership said Protxx CEO John Ralston has been early federal funding for fighting COVID-19 “to address the fact patients can’t get out to see their doctors because their offices and clinics are closed.”
“The goal is really to begin leveraging much more seriously wearable devices to expand virtual health care and remote patient monitoring,” Ralston told BioWorld. His platform will replace expensive equipment and time-consuming testing for earlier detection of MS symptoms, “particularly in Canada which has one of the highest incidence reports for MS in the world.”
Small and mighty
Protxx began by developing a wearable sensor called the Protxx PTS which sits just behind the ear where it detects multiple physiological impairments stemming from concussion in high contact sports. Its goal now: apply what it has learned to patients suffering from stroke, Parkinson’s Disease, diabetic nerve damage and more recently MS, which causes progressive disability in motor, visual, sensory and cognitive functions.
Currently, keeping track of all the equipment used to detect and monitor multiple impairments from conditions like these is like watching the credits roll from the last “Lord of the Rings” movie, from computerized dynamic posturography to video motion capture systems to electromyography and functional MRI. Virtually all of it is big and bulky and costly, said Ralston, leaving a doctor whose patient has multiple impairments only two options.
“You can spend the next twelve weeks scheduling your patient for a battery of these very expensive tests. Or you can pull out our Protxx PTS, put it on the patient and one minute later get 90% of the results right then and there so you can begin treating, monitoring and rehabbing the patient.”
The key here are sensors Protxx has adapted from system engineering tools for detecting and analyzing mechanical vibration in nuclear reactors, commercial aircraft and robots that mimic human movements. Protxx’s sensor is seated close to the vestibular balance organs in your ear and is built small to avoid distorting the physical motion being measured.
“Our physiological vibration acceleration or ‘phybrata’ sensor detects features and patterns in normal and pathological motions of the human body at a level that’s not previously been observed,” Ralston said. Ultimately, this will help doctors identify and quantify combinations of multiple impairments anywhere in the body’s physiological systems.
Machine learning models, meantime, automate standard statistical analytic tools for high levels of diagnostic predictions with “sensitivities and specificities above 90%,” said Ralston. This, coupled with the phybrata sensor, “is a pretty major breakthrough for a non-invasive, wearable device that costs a tiny fraction of today’s clinical systems,” said Ralston.
No magic boxes, please
The principal investigator measuring the accuracy of all this is Hossein Rouhani, a mechanical engineer and assistant professor in the University of Alberta’s Neuromuscular Control and Biomechanics Laboratory. The first phase will investigate how best to extract recorded, clinically meaningful data from the sensors and discover biomarkers that can identify changes in a patient’s medical condition.
That first phase will help determine the number of potential patients in a future randomized, clinical validation of the technology to determine its application in patients suffering from neurodegenerative disease, including MS. Rouhani said Protxx’s platform is far better than “black box” modeling which has a limited knowledge base from which to predict future changes in human physiology.
“We know which physiological parameters can be indicators of a change in disease conditions, in this case MS,” Rouhani told BioWorld. “Instead of black box modeling we target physiological parameters as they are related to motion, for example.”
Moreover, less data will be needed to cull results and unlike “a magic box which says Yes or No,” physicians will see the activity patterns within individuals, said Rouhani. “The physician will have a more active role in evaluating a medical condition and in decision-making.”
Reluctant to disclose development costs to date or funding sources, Ralston confirmed Ottawa is investing three dollars for every dollar Protxx invests in development of the technology.