What happened
Darmiyan of Berkeley, California, makes BrainSee, software that compares a patient's MRI, age, sex and cognitive scores with patients used to develop and train it. It returns a similarity score from 0 to 100, and scores above 50 resemble patients who progressed to Alzheimer's dementia within five years. The company validated it retrospectively on 198 patients with amnestic mild cognitive impairment from 27 North American sites. A consensus of three physicians confirmed which patients progressed. FDA reviewers report a specificity of 96%, a sensitivity of 73%, a positive predictive value of 97% and a negative predictive value of 71%. The lower confidence bound for sensitivity fell below the 70% goal, which FDA accepted. FDA granted the De Novo request on 11 January 2024. Another 209 eligible patients lacked five years of follow-up and were left out. Most patients were white, and FDA required a postmarket study.
Sources
-
Regulatory document, by authors with no tie to the system.
U.S. Food and Drug Administration, Center for Devices and Radiological Health. De Novo classification request for BrainSee (DEN220066): decision summary. Granted 2024-01-11.
This analysis resulted in specificity of 96% (92%, 100%), sensitivity of 73% (65%, 81%), PPV of 97% (93%, 100%), and NPV of 71% (62%, 79%).
Details
AI system
- Workflow
-
The software combined brain volumes from T1-weighted MRI with MMSE and CDR Sum of Boxes scores, age and sex. A score above 50 counted as predicted progression.
Follow-up
- Conditions for replication
-
The patient is 55 to 95 years old with a diagnosis of amnestic mild cognitive impairment. Scores near 50 call for further follow-up, and the score never stands alone.
Entry details
- Sector
-
Healthcare.
- Purpose
-
Darmiyan built BrainSee to help clinicians judge how likely patients aged 55 to 95 with amnestic mild cognitive impairment are to progress to Alzheimer's dementia within five years.
Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). BrainSee foresaw which memory patients would develop dementia (AIBD-0774). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0774.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)