Beyond skin deep: physiologic oral pigmentation as an overlooked validation gap in AI-assisted oral cancer screening
AI-assisted oral cancer screening tools, exemplified by MeMoSA, have been validated almost exclusively in Asian populations. We propose a testable hypothesis: physiologic oral pigmentation, which varies by race independent of pathology, may cause classifiers to misclassify normal variation as suspicious or miss camouflaged lesions. Following precedent from pulse oximetry and dermatology AI, we call for pigmentation-stratified validation before deployment in racially diverse settings such as the United States.
Authors
- Ashoo Grover (ORCID: https://orcid.org/0000-0001-8027-8608)
- Deepti Khanna (ORCID: https://orcid.org/0009-0008-0385-9722)
Institutions
- Indian Council of Medical Research (IN)
Publication Details
- Journal
- npj Digital Medicine
- Published
- 2026-11-14
- DOI
- https://doi.org/10.1038/s41746-026-03466-9
- Primary Topic
- AI in cancer detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Indian Council of Medical Research