FDA clears NEXT-Cue algorithm to forecast five-year dementia risk from MRI scans

The US Food and Drug Administration has cleared NEXT-Cue, an algorithmic tool designed to forecast five-year dementia risk from brain magnetic resonance imaging scans.

The software evaluates neuroimaging from patients presenting with mild cognitive impairment to project their likelihood of conversion to dementia. Mild cognitive impairment occupies an intermediate clinical space between normal cognitive ageing and full-scale dementia. While a subset of individuals with the condition maintains stable cognitive function for years, others experience rapid decline towards Alzheimer's disease. Clinicians have long lacked non-invasive, accessible methods to separate these patient cohorts reliably.

NEXT-Cue processes standard MRI sequences to evaluate subtle structural biomarkers across the brain. By analysing morphometric variations and patterns of tissue loss, the algorithm generates a five-year prognostic assessment. Radiologists and neurologists receive a standardised output detailing the patient's likelihood of progression. The tool functions as an adjunctive measure, intended to supplement rather than replace comprehensive neurological examinations and cognitive testing.

The decision marks an important regulatory precedent for predictive neuroimaging tools. Software as a medical device in neurology has predominantly centred on retrospective identification, such as detecting acute intracranial haemorrhage or quantifying established hippocampal atrophy. Regulatory clearance for NEXT-Cue reflects the FDA's willingness to sanction algorithmic prognostic models that attempt to foresee clinical trajectory over a multi-year horizon.

Accurate risk stratification has grown more consequential with the emergence of targeted Alzheimer's therapeutics. Disease-modifying treatments depend heavily on identifying patients during the earliest phases of neurodegeneration. Existing biological markers, including cerebrospinal fluid assays and amyloid positron emission tomography, often carry substantial financial costs or require invasive sampling. Algorithmic interpretation of conventional MRI offers a less burdensome screening mechanism for both clinical care and pharmaceutical trial enrolment.

The clearance specifies that NEXT-Cue must operate under specialist oversight. Healthcare providers must interpret its risk projections within the wider context of patient history and symptom severity.

As predictive algorithms transition from computational validation into routine neuroimaging practice, how will clinicians manage the ethical and diagnostic consequences of quantifying dementia risk years before severe symptoms appear?