One person's account of using Claude Code to analyze their own MRI.
One person's account of using Claude Code to analyze their own MRI.
The author gave Claude Code (Opus 4.8, xhigh) a roughly 266 MB DICOM export of a shoulder MRI, let it install packages, and told it only the symptom (2–3 weeks of right shoulder pain). It planned and read the scans for about an hour. A second run used subagents to arbitrate between its report and the doctor's, with extra context supplied.
Opus found an intact tendon, while the doctor reported a Grade III partial tear. The arbitration sided with Opus at moderate-to-high confidence. The author still couldn't fully trust either and is continuing physical therapy.
Each run took about an hour, and the author stresses the result is unverified and not medical advice. The AI disagreed sharply with the radiologist, so a third human opinion is still needed.
I have MRI images and the radiologist's report at [folder path]. Walk through the report in plain language, point out what each finding means, and write a list of questions I should ask my doctor. Be clear about what you can't determine from images alone, and don't present anything as a diagnosis.
A Claude Code-built agent checked ESPN every 30 minutes and sent each 2026 World Cup final score, written up in a persona, to a 3-person Signal group, with a code check on scores.
The poster exports Garmin data into a Claude project and meal planner, where Claude spotted a link between eating more on one day and reduced medication efficacy.
A Claude Cowork skill tracks diet and medical metrics in a spreadsheet, and meals told to Claude on mobile are logged in the daily record.