Multimodal oral cancer detection with embedding-level oversampling of fused image and clinical data
Pablo Ormeñ o-Arriagada, Carla Taramasco, Carlos Toro, Gastó n Má rquez, David Araya, Valentina Z u n ˜ iga, Diego Mellado, Diego Alcocer, David Ruete · 2026
Read Multimodal oral cancer detection with embedding-level oversampling of fused image and clinical data by Pablo Ormeñ o-Arriagada, Carla Taramasco, Carlos Toro, Gastó n Má rquez, David Araya, Valentina Z u n ˜ iga, Diego Mellado, Diego Alcocer, David Ruete (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13572481.txt.
Description
Pablo Ormeño-Arriagada, with eight colleagues, presents a 2026 method for early oral cancer detection that joins photos with basic patient facts. Oral cancer is aggressive and often found late, when survival odds drop. Photos alone can mislead, and confirmed cancer cases are scarce in real data sets, which skews training.
The team trained several models on labelled mouth photos paired with simple clinical data. To fix the imbalance, they made extra synthetic cases at the level of the learned features, not raw images. The best setup beat photo-only models on balanced scores that weigh each class fairly. The gain suggests that patient context helps the model sort benign from malignant sores. The authors aim the tool at screening, where early flags can speed referral and care.
Catalog record
| CREATOR | Pablo Ormeñ o-Arriagada, Carla Taramasco, Carlos Toro, Gastó n Má rquez, David Araya, Valentina Z u n ˜ iga, Diego Mellado, Diego Alcocer, David Ruete |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 77591 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |