Deep learning-based dose prediction in proton beam therapy for hepatocellular carcinoma: comparison of network architectures and loss functions
Shuta Ogawa, Noriyuki Kadoya, Takahiro Kato, Ryohei Kato, Ryota Tozuka, Yuki Narita, Sho Oyama, Masao Murakami, Keiichi Jingu · 2026
Read Deep learning-based dose prediction in proton beam therapy for hepatocellular carcinoma: comparison of network architectures and loss functions by Shuta Ogawa, Noriyuki Kadoya, Takahiro Kato, Ryohei Kato, Ryota Tozuka, Yuki Narita, Sho Oyama, Masao Murakami, Keiichi Jingu (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-13202326.txt.
Description
Shuta Ogawa and co-authors compare deep learning setups for predicting radiation dose in proton therapy for liver cancer. A predicted dose map can guide planners right away and let hospitals without proton facilities judge whether a patient might benefit. Their method needs only scans and drawn outlines, not beam settings.
Using data from many patients, they tested twenty model and loss combinations, including a new loss that pays attention to dose gradients near the sharp proton dose edge. Models with that loss made the smallest errors on target dose and low-dose region size, and matched clinical dose patterns most closely. Even simple networks did well when paired with the new loss.
Catalog record
| CREATOR | Shuta Ogawa, Noriyuki Kadoya, Takahiro Kato, Ryohei Kato, Ryota Tozuka, Yuki Narita, Sho Oyama, Masao Murakami, Keiichi Jingu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |