Fine-tuning of pre-trained models in transfer learning for prediction of radiation pneumonitis
Wang Zhixiang, Wang Guanjie, Wang Qingxin, Li Jia, Ren Pengling, Cai Linkun, Wang Xinghao, Sun Jing, Wang Zhenchang, Lyu Han, Zhang Zhen, Zhao Lujun
Abstract
ObjectiveTo predict radiation pneumonitis more effectively following thoracic radiotherapy, this study aims to fine-tune pre-trained models through transfer learning, to extract deep learning features.
MethodsBetween 2013 and 2018, CT images and dosage data from 314 lung cancer patients undergoing radical radiotherapy in the Tianjin Medical University Cancer Institute and Hospital were collected. To extract deep learning features related to radiation pneumonitis, a 3D ResNet50 model pre-trained with UCF101 video dataset is applied and fine-tuned against task adaptation. Evaluation strategies Principal Component Analysis (PCA) is used to contrast the feature distribution differences before and after fine-tuning, to validate their discriminative capacity. A logistic regression model is applied to compare the predictive performance using features post-fine-tuning, and the original pre-trained features, with AUC (Area Under the Curve) as the evaluation metric.
ResultsPCA confirms the fine-tuned features show a significant distinction between positive and negative samples of radiation pneumonitis. Moreover, the classification model based on fine-tuned features achieved an AUC score of 0.65, superior to the score of 0.58 using original pre-trained model features, indicating enhanced feature performance post-fine-tuning.
ConclusionsThe study demonstrates that employing transfer learning strategies to fine-tune generic pre-trained models for specific tasks, is feasible in effective deep learning features extraction that is suitable for predicting radiation pneumonitis, thus significantly enhancing predictive accuracy.
Key words:
Radiation pneumonitis; Deep learning; Transfer learning; Feature dimensionality reduction
Contributor Information
Wang Zhixiang
Medical Imaging Center, Beijing Friendship Hospital, Capital Medical University, Beijing Institute of Clinical Medical Research, Beijing 100050, China
Wang Guanjie
Department of Radiation Therapy, Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Tianjin Key Laboratory of Cancer Prevention and Therapy, Tianjin′s Clinical Research Center for Cancer, Tianjin 300060, China
Wang Qingxin
Department of Radiation Therapy, Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Tianjin Key Laboratory of Cancer Prevention and Therapy, Tianjin′s Clinical Research Center for Cancer, Tianjin 300060, China
Li Jia
Medical Imaging Center, Beijing Friendship Hospital, Capital Medical University, Beijing Institute of Clinical Medical Research, Beijing 100050, China
Ren Pengling
Medical Imaging Center, Beijing Friendship Hospital, Capital Medical University, Beijing Institute of Clinical Medical Research, Beijing 100050, China
Cai Linkun
School of Biological Sciences and Medical Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100191, China
Wang Xinghao
Medical Imaging Center, Beijing Friendship Hospital, Capital Medical University, Beijing Institute of Clinical Medical Research, Beijing 100050, China
Sun Jing
Medical Imaging Center, Beijing Friendship Hospital, Capital Medical University, Beijing Institute of Clinical Medical Research, Beijing 100050, China
Wang Zhenchang
Medical Imaging Center, Beijing Friendship Hospital, Capital Medical University, Beijing Institute of Clinical Medical Research, Beijing 100050, China
Lyu Han
Medical Imaging Center, Beijing Friendship Hospital, Capital Medical University, Beijing Institute of Clinical Medical Research, Beijing 100050, China
Zhang Zhen
Department of Radiation Therapy, Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Tianjin Key Laboratory of Cancer Prevention and Therapy, Tianjin′s Clinical Research Center for Cancer, Tianjin 300060, China
Zhejiang Cancer Hospital, Institute of Basic Medical Sciences and Cancer, Chinese Academy of Sciences, Hangzhou 310022, China
Zhao Lujun
Department of Radiation Therapy, Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Tianjin Key Laboratory of Cancer Prevention and Therapy, Tianjin′s Clinical Research Center for Cancer, Tianjin 300060, China
