TY  - JOUR
AU  - Long, Yuehong
AU  - Zeng, Yuliang
AU  - Zhang, Tao
AU  - Zhou, Jiancun
AU  - He, Qian
AU  - Liu, Xingqi
AU  - Wang, Ke
PY  - 2026
DA  - 2026/01/04
TI  - Radiomic Evaluation Model on the Efficacy of Neoadjuvant Chemotherapy for Non-small Cell Lung Cancer A Multicenter Collaborative Research Based on Privacy Protection
JO  - Chinese Journal of Information Fusion
T2  - Chinese Journal of Information Fusion
JF  - Chinese Journal of Information Fusion
VL  - 3
IS  - 1
SP  - 17
EP  - 30
DO  - 10.62762/CJIF.2025.125241
UR  - https://www.icck.org/article/abs/CJIF.2025.125241
KW  - radiomics
KW  - federated learning
KW  - deep learning
KW  - distributed learning
AB  - Background: Practical implementation of radiomics research faces significant data accessibility challenges due to privacy and ethical restrictions on multicenter data aggregation. Federated Learning (FL) provides a secure distributed framework that preserves data privacy through cryptographic techniques. Its adoption in radiomics is an emerging trend, enabling collaborative training without sharing sensitive imaging data. However, the inherently Non-IID data distribution across clients in FL often leads to class imbalance, which can substantially degrade global model performance. Purpose: To develop a privacy-preserving, multicenter collaborative CT-radiomics model for evaluating neoadjuvant chemotherapy efficacy in non‑small cell lung cancer (NSCLC). Methods: To mitigate FL performance degradation caused by data imbalance, we propose a parameter‑sharing federated aggregation algorithm (FedPS), where model parameters are sequentially shared via the server. Results: On an imbalanced NSCLC NAC efficacy dataset, centralized learning achieved an AUC of 0.92. FedPS attained competitive performance (AUC = 0.88), approaching the centralized benchmark while preserving privacy. Common FL algorithms performed lower: FedAvg (AUC = 0.84), FedSGD (0.85), and FedProx (0.85). On extremely imbalanced data, FedPS maintained good performance (AUC = 0.86), compared to FedAvg (0.80), FedSGD (0.83), and FedProx (0.85). Conclusions: The proposed FedPS algorithm demonstrates promising classification and generalization performance in imbalanced federated learning scenarios.
SN  - 2998-3371
PB  - Institute of Central Computation and Knowledge
LA  - English
ER  - 

@article{Long2026Radiomic,
  author = {Yuehong Long and Yuliang Zeng and Tao Zhang and Jiancun Zhou and Qian He and Xingqi Liu and Ke Wang},
  title = {Radiomic Evaluation Model on the Efficacy of Neoadjuvant Chemotherapy for Non-small Cell Lung Cancer A Multicenter Collaborative Research Based on Privacy Protection},
  journal = {Chinese Journal of Information Fusion},
  year = {2026},
  volume = {3},
  number = {1},
  pages = {17-30},
  doi = {10.62762/CJIF.2025.125241},
  url = {https://www.icck.org/article/abs/CJIF.2025.125241},
  abstract = {Background: Practical implementation of radiomics research faces significant data accessibility challenges due to privacy and ethical restrictions on multicenter data aggregation. Federated Learning (FL) provides a secure distributed framework that preserves data privacy through cryptographic techniques. Its adoption in radiomics is an emerging trend, enabling collaborative training without sharing sensitive imaging data. However, the inherently Non-IID data distribution across clients in FL often leads to class imbalance, which can substantially degrade global model performance. Purpose: To develop a privacy-preserving, multicenter collaborative CT-radiomics model for evaluating neoadjuvant chemotherapy efficacy in non‑small cell lung cancer (NSCLC). Methods: To mitigate FL performance degradation caused by data imbalance, we propose a parameter‑sharing federated aggregation algorithm (FedPS), where model parameters are sequentially shared via the server. Results: On an imbalanced NSCLC NAC efficacy dataset, centralized learning achieved an AUC of 0.92. FedPS attained competitive performance (AUC = 0.88), approaching the centralized benchmark while preserving privacy. Common FL algorithms performed lower: FedAvg (AUC = 0.84), FedSGD (0.85), and FedProx (0.85). On extremely imbalanced data, FedPS maintained good performance (AUC = 0.86), compared to FedAvg (0.80), FedSGD (0.83), and FedProx (0.85). Conclusions: The proposed FedPS algorithm demonstrates promising classification and generalization performance in imbalanced federated learning scenarios.},
  keywords = {radiomics, federated learning, deep learning, distributed learning},
  issn = {2998-3371},
  publisher = {Institute of Central Computation and Knowledge}
}