Inferring cancer disease response from radiology reports using large language models with data augmentation and prompting.
Ryan Shea Ying Cong TanQian LinGuat Hwa LowRuixi LinTzer Chew GohChristopher Chu En ChangFung Fung LeeWei Yin ChanWei Chong TanHan Jieh TeyFun Loon LeongHong Qi TanWen Long NeiWen Yee ChayDavid Wai Meng TaiGillianne Geet Yi LaiLionel Tim-Ee ChengFuh Yong WongMatthew Chin Heng ChuaMelvin Lee Kiang ChuaDaniel Shao Weng TanChoon Hua ThngIain Bee Huat TanHwee Tou NgPublished in: Journal of the American Medical Informatics Association : JAMIA (2023)
Large clinical language models demonstrate potential to infer cancer disease response from radiology reports at scale. Data augmentation techniques are useful to further improve performance. Prompt-based fine-tuning can significantly reduce the size of the training dataset.