Multimodal GPT-5 for Predicting Poor Functional Outcomes After Intracerebral Hemorrhage in the Emergency Department: Validation Study
Koutarou Matsumoto, Kazuaki Ishihara, Ryota Tamba, Yusuke Fujiyoshi, Koki Tokunaga, Katsuhiko Matsuda, Yasunobu Nohara, Jenhui Chen, Shigeo Yamashiro, Naoki Nakashima, Masahiro Kamouchi · 2026
Read Multimodal GPT-5 for Predicting Poor Functional Outcomes After Intracerebral Hemorrhage in the Emergency Department: Validation Study by Koutarou Matsumoto, Kazuaki Ishihara, Ryota Tamba, Yusuke Fujiyoshi, Koki Tokunaga, Katsuhiko Matsuda, Yasunobu Nohara, Jenhui Chen, Shigeo Yamashiro, Naoki Nakashima, Masahiro Kamouchi (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/pmc-13216710.txt.
Description
Koutarou Matsumoto and co-authors present a 2026 validation study from a Japanese emergency department, testing whether general-purpose AI chatbots can predict poor recovery after bleeding inside the brain. Fast prognosis shapes early decisions, and specialists are not always at hand.
Using admission data and CT images, GPT-4.1 and GPT-5 scored discrimination close to a purpose-built machine learning model, with areas under the curve around 0.84 to 0.85. Their predictions were highly reproducible across runs. Calibration, however, lagged: the chatbots systematically underestimated risk, and their overall accuracy measures trailed the specialist model. Feeding model outputs into the prompts helped modestly. The authors position the chatbots as explanation aids beside, not instead of, dedicated prognostic models.
Catalog record
| CREATOR | Koutarou Matsumoto, Kazuaki Ishihara, Ryota Tamba, Yusuke Fujiyoshi, Koki Tokunaga, Katsuhiko Matsuda, Yasunobu Nohara, Jenhui Chen, Shigeo Yamashiro, Naoki Nakashima, Masahiro Kamouchi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |