Federated Learning with Adaptive Intermediate Model Selection for Predicting IVIG Resistance in Kawasaki Disease
DOI:
https://doi.org/10.3991/ijoe.v22i02.58737Keywords:
Federated Learning, Kawasaki Disease, Intravenous Immunoglobulin Resistance, ADASYN, Flower Framework, Convolutional Neural Network, Adaptive Model SelectionAbstract
Kawasaki disease (KD), a rare pediatric illness affecting children under five, is treated with intravenous immunoglobulin (IVIG). But 10–20% of patients are resistant to IVIG, and these resistant kids face a higher risk of coronary artery abnormalities. Identifying resistance early is vital, yet data scarcity, class imbalance, and the disease’s rarity necessitate nationwide collaboration, which is often hindered by country-specific privacy policies. Federated learning (FL) provides a practical way for different parties to collaborate on training a model while keeping their raw data private and secure. To enhance model adaptability across diverse clinical populations, we propose an adaptive intermediate model selection strategy in federated learning. Each client retains the version—global or locally fine-tuned—that performs best on its own data, using customizable performance metrics such as F1-score or recall. The system was implemented using the Flower FL framework, with three simulated clients and a shared convolutional neural network (CNN) architecture. Experiments demonstrated that the global model achieved stronger performance than conventional models, and several clients obtained further gains by selecting intermediate models aligned with their data. This approach introduces a novel balance between worldwide collaboration and local personalization in FL, offering a flexible and clinically meaningful solution for IVIG resistance prediction.
References
[1] J. Schaefer, M. Lehne, J. Schepers, F. Prasser, and S. Thun, “The use of machine learning in rare diseases: A scoping review,” Orphanet J Rare Dis, vol. 15, no. 1, pp. 1–10, Jun. 2020, doi: 10.1186/S13023-020-01424-6/FIGURES/4.
[2] J. Lee et al., “Deep learning for rare disease: A scoping review,” J Biomed Inform, vol. 135, p. 104227, Nov. 2022, doi: 10.1016/J.JBI.2022.104227.
[3] B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-Efficient Learning of Deep Networks from Decentralized Data,” Apr. 10, 2017, PMLR. Accessed: Jul. 01, 2024. [Online]. Available: https://proceedings.mlr.press/v54/mcmahan17a.html
[4] D. J. Beutel, T. Topal, A. Mathur, X. Qiu, T. Parcollet, and N. Lane, “Flower: A Friendly Federated Learning Research Framework,” ArXiv, 2020.
[5] I. U. Haq, M. Pifarré, and E. Fraca, “Natural Language Processing Approach to Evaluate Real-Time Flexibility of Ideas to Support Collaborative Creative Process,” International Journal of Emerging Technologies in Learning (iJET), vol. 19, no. 05, pp. 93–107, Jun. 2024, doi: 10.3991/IJET.V19I05.47465.
[6] D. J. Challacombe and E. N. McElhiney, “Phishing Susceptibility Among Healthcare Workers: The Impact of Awareness, Email Type, and Location,” International Journal of Advanced Corporate Learning (iJAC), vol. 18, no. 1, pp. 4–15, Mar. 2025, doi: 10.3991/IJAC.V18I1.51671.
[7] S. Alshattnawi and H. R. Alshboul, “Combined Deep Learning Approaches for Intrusion Detection Systems,” International Journal of Interactive Mobile Technologies (iJIM), vol. 18, no. 19, pp. 144–155, Oct. 2024, doi: 10.3991/IJIM.V18I19.49907.
[8] S. S. Gupta, K. E. Joslyn, K. D. McKenney, and A. M. Comi, “Biomarker development in Sturge-Weber syndrome,” J Neurodev Disord, vol. 17, no. 1, pp. 1–8, Dec. 2025, doi: 10.1186/S11689-025-09640-6/FIGURES/1.
[9] I. Smesseim et al., “Prospective validation of an artificial intelligence assessment in a cohort of applicants seeking financial compensation for asbestosis (PROSBEST),” Eur Radiol Exp, vol. 9, no. 1, pp. 1–8, Dec. 2025, doi: 10.1186/S41747-025-00619-5/FIGURES/2.
[10] K. Saleh et al., “4-025 Facilitating AI-ECG models for rare cardiac diseases: transfer learning and synthetic data generation for brugada ECG classification,” Heart, vol. 111, no. Suppl 3, pp. A139–A140, Sep. 2025, doi: 10.1136/HEARTJNL-2025-BCS.137.
[11] F. Amin, H. Khalid, M. Khalid, M. Talha, and A. Waafira, “Integrative AI-metabolomics: a new frontier in diagnosing pulmonary tumor thrombotic microangiopathy,” Annals of Medicine & Surgery, Aug. 2025, doi: 10.1097/MS9.0000000000003707.
[12] T. Wang, G. Liu, and H. Lin, “A machine learning approach to predict intravenous immunoglobulin resistance in Kawasaki disease patients: A study based on a Southeast China population,” PLoS One, vol. 15, no. 8, p. e0237321, Aug. 2020, doi: 10.1371/JOURNAL.PONE.0237321.
[13] J. Liu et al., “A Machine Learning Model to Predict Intravenous Immunoglobulin-Resistant Kawasaki Disease Patients: A Retrospective Study Based on the Chongqing Population,” Front Pediatr, vol. 9, p. 756095, Nov. 2021, doi: 10.3389/FPED.2021.756095/BIBTEX.
[14] Y. Sunaga and A. Watanabe, “A Simple Scoring Model Based on Machine Learning Predicts Intravenous Immunoglobulin Resistance in Kawasaki Disease,” 2022, doi: 10.21203/rs.3.rs-1215051/v1.
[15] Y. He et al., “Interpretable web-based machine learning model for predicting intravenous immunoglobulin resistance in Kawasaki disease,” Ital J Pediatr, vol. 51, no. 1, pp. 1–17, Dec. 2025, doi: 10.1186/S13052-025-02036-1/TABLES/5.
[16] J. Y. Lam et al., “Intravenous immunoglobulin resistance in Kawasaki disease patients: prediction using clinical data,” Pediatric Research 2023 95:3, vol. 95, no. 3, pp. 692–697, Feb. 2023, doi: 10.1038/s41390-023-02519-z.
[17] H. Huang et al., “Nomogram to predict risk of resistance to intravenous immunoglobulin in children hospitalized with Kawasaki disease in Eastern China,” Ann Med, vol. 54, no. 1, pp. 442–453, Dec. 2022, doi: 10.1080/07853890.2022.2031273;WGROUP:STRING:PUBLICATION.
[18] Y. Yang et al., “Research on Early Identification Model of Intravenous Immunoglobulin Resistant Kawasaki Disease Based on Gradient Boosting Decision Tree,” Pediatric Infectious Disease Journal, vol. 42, no. 7, pp. 537–542, Jul. 2023, doi: 10.1097/INF.0000000000003919.
[19] J. Wang, X. Huang, and D. Guo, “Predictors and a novel predictive model for intravascular immunoglobulin resistance in Kawasaki disease,” Ital J Pediatr, vol. 49, no. 1, pp. 1–8, Dec. 2023, doi: 10.1186/S13052-023-01531-7/FIGURES/4.
[20] R. K. Natarajan, S. V. Bhoopalan, C. Cross, R. Shah, and A. Rothman, “Novel Score to Predict Immunoglobulin Resistance in Kawasaki Disease,” Pediatr Cardiol, vol. 44, no. 7, pp. 1546–1551, Oct. 2023, doi: 10.1007/S00246-023-03175-0/METRICS.
[21] U. Kaya Akca et al., “Comparison of IVIG resistance predictive models in Kawasaki disease,” Pediatr Res, vol. 91, no. 3, pp. 621–626, Feb. 2022, doi: 10.1038/s41390-021-01459-w.
[22] Y. Xia et al., “A machine learning-based model to predict intravenous immunoglobulin resistance in Kawasaki disease,” iScience, vol. 28, no. 3, p. 112004, Mar. 2025, doi: 10.1016/j.isci.2025.112004.
[23] L. Deng et al., “Construction and validation of predictive models for intravenous immunoglobulin–resistant Kawasaki disease using an interpretable machine learning approach,” Clin Exp Pediatr, vol. 67, no. 8, p. 405, Aug. 2024, doi: 10.3345/CEP.2024.00549.
[24] D. Kumar, C. Verma, and Z. Illés, “Federated learning with explainable AI for liver disease prediction: A privacy-preserving approach,” Intell Based Med, vol. 12, p. 100285, Jan. 2025, doi: 10.1016/J.IBMED.2025.100285.
[25] Y. Gao, G. Zhang, C. Zhang, J. Wang, L. T. Yang, and Y. Zhao, “Federated tensor decomposition-based feature extraction approach for industrial IoT,” IEEE Trans Industr Inform, vol. 17, no. 12, pp. 8541–8549, 2021, doi: 10.1109/TII.2021.3074152.
[26] M. Karmakar, A. Hota, and A. Nag, “Convolutional neural network (CNN) and federated learning-based approach for lung disease detection,” Iran Journal of Computer Science, pp. 1–22, Aug. 2025, doi: 10.1007/S42044-025-00320-1/METRICS.
[27] D. M. Jimenez Gutierrez, H. M. Hassan, L. Landi, A. Vitaletti, and I. Chatzigiannakis, “Application of Federated Learning Techniques for Arrhythmia Classification Using 12-Lead ECG Signals,” Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14053 LNCS, pp. 38–65, 2024, doi: 10.1007/978-3-031-49361-4_3.
[28] H. He, Y. Bai, E. A. Garcia, and S. Li, “ADASYN: Adaptive synthetic sampling approach for imbalanced learning,” Proceedings of the International Joint Conference on Neural Networks, pp. 1322–1328, 2008, doi: 10.1109/IJCNN.2008.4633969.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Namitha T N, Raghavendra S, Vinith R
This work is licensed under a Creative Commons Attribution 4.0 International License.
