Though the bulk of recent attention has been directed toward drugs and devices to treat diagnosed cases of Alzheimer’s disease—like Biogen’s controversy-mired Aduhelm—Altoida, for one, has set its sights much earlier in the disease progression timeline.
Altoida’s technology uses artificial intelligence to analyze cognitive test results to determine whether a case of mild cognitive impairment will escalate into Alzheimer’s within a year.
Diagnosing the condition that early, before symptoms have even begun to appear, could help physicians begin to treat at-risk patients right away, potentially delaying the onset or lessening the severity of the neurodegenerative disease.
Clearly excited by this prospect is the FDA, which has awarded its breakthrough designation to Altoida’s predictive system. The breakthrough label will clear some hurdles from the company’s regulatory pathway, upping its chances of getting to market sooner—so long as it proves successful in ongoing clinical trials.
“This designation is recognition of the strong clinical need to accurately and reliably predict conversion to Alzheimer’s disease before irreversible damage can occur. Altoida’s device could enable predictive diagnosis of neurodegenerative disorders at the population level, which can in turn enable preventative and therapeutic intervention in the earliest stages to delay onset and improve clinical outcomes,” said Ioannis Tarnanas, Ph.D., Altoida’s chief scientific officer.
The system is the product of more than 20 years of cognitive research, Tarnanas said. It comprises a slate of neurological tests for users aged 55 and up and AI software to analyze the results of those tests.
The assessment portion of the system takes 10 minutes to complete and can be accessed on a user’s own smartphone or tablet. It focuses on measuring 11 specific areas of the brain research has suggested are linked to the onset of Alzheimer’s.
The augmented reality tests are designed to feel like video games. In one, users are asked to hide three virtual items around the room they’re in, then relocate them at random. Another tasks users with learning the tools and actions needed to simulate a fire evacuation, and the third main activity has them simultaneously locate virtual tools in their environment while a dynamic sound plays intermittently.
Once the tests are complete, Altoida’s AI assesses the results to score the user’s risk of developing Alzheimer’s within the next year. It produces a full cognitive report based on hand and gait movements and errors, eye tracking and pupil dilation, voice parameters and more.
Studies published within the last year have found that the Altoida program has a prognostic accuracy of 94%, while also offering about 91% sensitivity and 82% specificity in predicting the progression of dementia. That makes it approximately 2.6 times more sensitive than traditional tests for Alzheimer’s—in addition to being much more efficient because of its 10-minute span.
If approved by the FDA, Altoida’s device would be the first diagnostic tool cleared to predict Alzheimer’s onset before symptoms arise, but there are plenty of other tech developers hot on its tail.
Just last month, Boston-based startup Linus Health raised $55 million to continue building its own early diagnosis software, which offers a digital version of the traditional clock-drawing test as well as a suite of tools to monitor the progression of a patient’s cognitive decline over time.
And in late 2020, IBM and Pfizer published a study of their own digital diagnostic, which uses an AI model to predict Alzheimer’s development using only a basic language test. According to the study, the model was able to accurately identify which healthy people would eventually develop dementia about 74% of the time.