Case Input
Have your own head CT? Send the full DICOM study as a .zip — Dean personally reviews it and emails back a PDF result.
Ground rules. Not a medical device or diagnosis. De-identified data only, as a DICOM study .zip (all slices, zipped — not a single image). By submitting, you agree to indemnify and hold harmless AEGIS and Dr Loh Kah Meng from claims arising from use of AEGIS outputs. Not for emergencies — seek immediate care for symptoms of concern. PDPA-compliant handling.
Goes to a real person, not an automated system — typically 1–2 business days. Research-stage software only; not a diagnosis.
Submission received. Your report will be sent to the email you provided.
or explore real de-identified sample cases
Real case gallery — 24 de-identified RSNA studies. Click any scan to load it and see the board’s result instantly.
GREEN · 6 real cases
YELLOW · 6 real cases
ORANGE · 6 real cases
RED · 6 real cases
Result
Select a real case from the gallery above to see the board’s result.
How this result was produced. Every case is scored by the AEGIS AI ensemble — 3 decision engines voting by majority, cross-checked against 4 reference engines. Engine identity, architecture, and internal weighting are protected AEGIS research IP and are never disclosed; only the ensemble’s governance-zone result is shown here. System AUC on RSNA validation: 90.15%.