BoneFormer-KG:联邦知识引导的3D Transformer用于部分标注CT的骨肿瘤分割
Federated knowledge-guided 3D Transformer learning for bone tumor segmentation from partially annotated computed tomography data
BoneFormer-KG 框架以联邦知识引导的3D Transformer,在部分标注、非独立同分布的多中心CT数据上分割骨肿瘤及转移灶。Spine-Mets-CT-SEG病灶DSC为0.793±0.049、BM-Seg的BM DSC为0.845±0.065、ULS23 Bone病灶DSC为0.672±0.078,较最佳基线UFPS分别提高2.5%、1.7%和2.4%。
- Article
- Open access
- Published: 05 October 2026
- Yongan Liu1 na1,
- Tao Yu4 na1,
- Dan Su3 na1,
- Ming Ni4 na1,
- Shenyu Wang2 na1,
- Yuan Hong4 na1,
- Xinchen Li4 na1,
- Jingci Xu2 na1,
- Liying Yang3 na1 &
- …
- Mu Hu4 na1
Scientific Reports (2026) Cite this article
We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.
Abstract
Accurate 3D delineation of bone tumors and metastatic lesions from volumetric CT is essential for diagnosis, staging, radiotherapy planning, and treatment monitoring, yet manual contouring is time consuming and variable across observers. Although modern 3D CNN/Transformer segmenters achieve strong centralized performance, they typically assume pooled multi-institution data and consistent label definitions. In practice, clinical imaging archives are privacy-isolated, exhibit pronounced non-IID appearance shift (scanner/protocol and case-mix variation), and provide heterogeneous and partially annotated supervision (label-set mismatch and voxel-level incompleteness), where naive training and FedAvg-style aggregation can induce client drift and systematic false negatives. We propose federated knowledge-guided 3D Transformer segmentation (BoneFormer-KG) framework that explicitly addresses these constraints. BoneFormer-KG adapts a pretrained promptable backbone to volumetric CT using a parameter-efficient 3D pipeline with factorized in-plane/axial patch embedding, a learnable depth positional table, and lightweight 3D adapters to inject through-plane inductive bias while keeping memory and communication costs tractable. To handle label-set mismatch and partial voxel supervision, local optimization uses masked hybrid supervision (Dice + cross-entropy) computed only over annotated voxels and locally available labels, avoiding false-negative gradients on unobserved classes. On the server, we introduce effective labeled-signal weighting to aggregate client updates proportional to supervised voxel evidence, reducing bias under uneven annotation density. To further mitigate heterogeneity and supervision gaps, BoneFormer-KG integrates a coupled global-local distillation strategy: confidence-gated global distillation regularizes missing labels/unlabeled regions using the current global teacher, while class-sparse peer distillation transfers specialist knowledge from label-rich clients to label-poor clients with controlled bandwidth. An Auto Prompt Generator (APG) replaces manual prompting by synthesizing volumetric prompts from learned features for scalable prompt-free inference. We evaluate BoneFormer-KG on three CT datasets under non-IID federated splits and heterogeneous/partial supervision. BoneFormer-KG achieves competitive macro-averaged performance, reaching \(0.793\!\pm \!0.049\) lesion DSC and \(7.4\!\pm \!2.7\) mm HD95 on Spine-Mets-CT-SEG, \(0.845\!\pm \!0.065\) BM DSC with \(5.6\!\pm \!2.1\) mm BM HD95 (and \(0.964\!\pm \!0.014\) bone DSC) on BM-Seg, and \(0.672\!\pm \!0.078\) lesion DSC with \(11.3\!\pm \!4.1\) mm HD95 and \(0.751\!\pm \!0.069\) lesion \(F_1\) on ULS23 Bone VOIs, outperforming the best baseline (UFPS) by 2.5% DSC on Spine-Mets-CT-SEG, 1.7% on BM-Seg, and 2.4% on ULS23 Bone baselines.
Author information
Author notes
These authors contributed equally: Yongan Liu, Tao Yu, Dan Su, Ming Ni, Shenyu Wang, Yuan Hong, XinchenLi, Jingci Xu, Liying Yang and Mu Hu.
Authors and Affiliations
Department of Intensive Care Unit, Shanghai Jiao Tong University School of Medicine, Ruijin Hospital, Shanghai, 201801, China
Yongan Liu
Department of Radiology, Shanghai Jiao Tong University School of Medicine, Ruijin Hospital, Shanghai, 201801, China
Shenyu Wang & Jingci Xu
Department of Operating Room, Shanghai Jiao Tong University School of Medicine, Ruijin Hospital, Shanghai, 201801, China
Dan Su & Liying Yang
Department of Orthopedics, Shanghai Jiao Tong University School of Medicine, Ruijin Hospital, Shanghai, 201801, China
Tao Yu, Ming Ni, Yuan Hong, Xinchen Li & Mu Hu
Authors
- Yongan Liu
- Tao Yu
- Dan Su
- Ming Ni
- Shenyu Wang
- Yuan Hong
- Xinchen Li
- Jingci Xu
- Liying Yang
- Mu Hu
Corresponding author
Correspondence to Mu Hu.
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A
Reproducibility and implementation details
This appendix summarizes the principal training, architectural, and dataset-partitioning settings used for BoneFormer-KG.
Training hyperparameters
Table 8 reports the main optimization and federated training settings used in the experiments.
Architecture specifications
Table 9 lists the principal architectural specifications of the parameter-efficient 3D adaptation pipeline.
Dataset partitioning
Table 10 summarizes the patient-level client construction used to simulate non-IID multi-institutional training.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Liu, Y., Yu, T., Su, D. et al. Federated knowledge-guided 3D Transformer learning for bone tumor segmentation from partially annotated computed tomography data. Sci Rep (2026). https://doi.org/10.1038/s41598-026-71869-5
Received: 22 January 2026
Accepted: 11 September 2026
Published: 05 October 2026
DOI: https://doi.org/10.1038/s41598-026-71869-5
Keywords
来源:Nature 健康护理主题 · nature.com