Does Machine Unlearning Preserve Clinical Safety? A Risk Analysis for Medical Image Classification
Resumo
The application of Deep Learning in medical diagnosis must balance patient safety with compliance with data protection regulations. Machine Unlearning enables the selective removal of training data from deployed models. However, most methods are validated primarily through efficiency and privacy-oriented metrics, with limited attention to clinically asymmetric error costs. In this work, we investigate how unlearning affects clinical risk in binary medical image classification. We show that standard unlearning strategies (Fine-Tuning, Random Labeling, and SalUn) may reduce test utility while increasing false-negative rates, thereby amplifying clinical risk. To mitigate this, we propose SalUn-CRA (Clinical Risk-Aware), a variant of SalUn that replaces random relabeling with entropy-based forgetting for malignant samples in the forget set, preventing the model from learning harmful benign associations. We evaluate on DermaMNIST and PathMNIST medical image datasets under 20% and 50% data removal. Using Global Risk metrics with asymmetric costs, SalUn-CRA achieves lower or comparable clinical risk to full retraining while preserving unlearning effectiveness. These results suggest that clinical risk should be an integral component of unlearning validation in medical systems.
Referências
Bourtoule, L., Chandrasekaran, V., Choquette-Choo, C. A., Jia, H., Travers, A., Zhang, B., Lie, D., and Papernot, N. (2021). Machine unlearning. In 2021 IEEE symposium on security and privacy (SP), pages 141–159. IEEE.
Brazil (2018). Brazilian general data protection law (law no. 13,709/2018). [link]. Accessed: 25 Feb. 2026.
Chan, H.-P., Samala, R. K., Hadjiiski, L. M., and Zhou, C. (2020a). Deep learning in medical image analysis. In Deep Learning in Medical Image Analysis: Challenges and Applications, pages 3–21. Springer.
Chan, H.-P., Samala, R. K., Hadjiiski, L. M., and Zhou, C. (2020b). Deep learning in medical image analysis. Deep learning in medical image analysis: challenges and applications, pages 3–21.
Deng, Z. et al. (2025). Maverick: Collaboration-free federated unlearning for medical privacy. In Lecture Notes in Computer Science. Springer.
European Parliament and Council of the European Union (2016). Regulation (eu) 2016/679 (general data protection regulation – gdpr). [link]. Accessed: 25 Feb. 2026.
Falcao, A. and Cordeiro, F. (2025). Análise de desaprendizado de máquina em modelos de classificação de imagens médicas. In Anais Estendidos do XXV Simpósio Brasileiro de Computação Aplicada à Saúde, pages 43–48, Porto Alegre, RS, Brasil. SBC.
Fan, C., Liu, J., Zhang, Y., Wong, E., Wei, D., and Liu, S. (2024). Salun: Empowering machine unlearning via gradient-based weight saliency in both image classification and generation. In International Conference on Learning Representations (ICLR).
Golatkar, A., Achille, A., and Soatto, S. (2020). Eternal sunshine of the spotless net: Selective forgetting in deep networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9304–9312.
Graves, L., Nagisetty, V., and Ganesh, V. (2021). Amnesiac machine learning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 11516–11524.
Haimerl, M. and Reich, C. (2025). Risk-based evaluation of machine learning-based classification methods used for medical devices. BMC Medical Informatics and Decision Making, 25(1):126.
Hardan, S., Taratynova, D., Essofi, A., Nandakumar, K., and Yaqub, M. (2025). Forget-mi: Machine unlearning for forgetting multimodal information in healthcare settings. In International Conference on Medical Image Computing and Computer-Assisted Intervention, pages 204–213. Springer.
Hoofnagle, C. J., van der Sloot, B., and Borgesius, F. Z. (2019). The european union general data protection regulation: What it is and what it means. Information & Communications Technology Law, 28(1):65–98.
Li, N., Zhou, C., Gao, Y., Chen, H., Fu, A., Zhang, Z., and Yu, S. (2024). Machine unlearning: Taxonomy, metrics, applications, challenges, and prospects. ACM Computing Surveys.
Ling, C. X. and Sheng, V. S. (2010). Cost-Sensitive Learning and the Class Imbalance Problem, pages 231–235. Springer.
Mester, S. and et al. (2024). Machine unlearning for medical imaging. ResearchGate.
Nasirigerdeh, R., Razmi, N., Schnabel, J. A., Rueckert, D., and Kaissis, G. (2024). Machine unlearning for medical imaging. arXiv preprint arXiv:2407.07539. Acesso em: 23 fev. 2025.
Sakib, S. K. and Xie, M. (2024). Machine unlearning in digital healthcare: Addressing technical and ethical challenges. In Proceedings of the AAAI Symposium Series, volume 4, pages 319–322.
Scholz, R. and et al. (2024). Imbalance-aware loss functions improve medical image classification. In Proceedings of Machine Learning Research.
Warnecke, A. et al. (2021). Machine unlearning of features and labels. arXiv preprint arXiv:2108.11577.
Wu, Z., Shen, C., and Van Den Hengel, A. (2019). Wider or deeper: Revisiting the resnet model for visual recognition. Pattern Recognition, 90:119–133.
Yang, J., Shi, R., Wei, D., Liu, Z., Wang, L., Zhou, Y., Zhou, S., Bian, C., Li, L., Wang, X., et al. (2021). Medmnist: A lightweight automl benchmark for medical image analysis. [link]. Accessed: February 13, 2025.
Zhang, H., Nakamura, T., Isohara, T., and Sakurai, K. (2023). A review on machine unlearning. SN Computer Science, 4(4):337.
