用跨序列迁移学习减少膝关节MRI软骨分割标注量
Reducing Annotation Burden for Femoral Cartilage Segmentation in Knee MRI via Cross-Sequence Transfer Learning
- 用DESS图像预训练模型，再迁移到Cube序列上微调
- 仅需9个Cube样本即可达到DESS的分割精度（DSC 0.903）
- 病变对不同序列影响不同，需注意序列选择
目的：开发并评估跨序列迁移学习在股骨软骨自动分割中的应用，测试双回波稳态（DESS）与矢状面质子密度加权3D快速自旋回波（Cube）序列间的双向迁移。方法：在来自骨关节炎进展研究计划（OAI）的507张DESS图像上优化改进的2D U-Net。在意大利博洛尼亚伊斯坦特奥托佩迪科·里佐利研究所采集的44张DESS和44张Cube图像子集上建立同序列基线，每组含22例无损伤与22例损伤病例。通过逐步增加目标序列训练样本数，对预训练模型进行微调以研究收敛性，同时保持验证与测试集不变。使用骰子相似系数（DSC）与平均表面距离（ASD）评估分割效果。采用双侧曼-惠特尼U检验结合邦弗朗尼校正评估损伤影响。结果：同序列训练中，DESS表现优于Cube（DSC分别为0.900与0.830，P<0.001）。Cube到DESS的迁移学习性能与原DESS相当（DSC 0.903±0.032 vs 0.900±0.027），在9个训练样本时达到性能平稳期。而DESS到Cube的迁移表现较差（联合DSC 0.802±0.049 vs 0.830±0.042），在24个样本时达平稳期。损伤对DESS无显著影响（P≥0.39），但降低Cube分割精度（DSC 0.805 vs 0.856，P<0.001）。结论：跨序列迁移学习可显著降低目标序列的标注需求，但性能具有方向性和序列依赖性，且损伤对不同序列的影响各异。
原文摘要 · Abstract (English)
Purpose: To develop and evaluate cross-sequence transfer learning for automatic femoral cartilage segmentation, testing bidirectional transfer between dual-echo steady-state (DESS) and sagittal proton density-weighted 3D fast spin-echo (Cube) sequences. Materials and Methods: We optimized a modified 2D U-Net on 507 DESS images from the Osteoarthritis Initiative (OAI). We then established same-sequence baselines using subject-level cross-validation on a subset of 44 OAI DESS images and 44 Cube images acquired at the Istituto Ortopedico Rizzoli, Bologna, Italy. Each subset included 22 non-lesioned and 22 lesioned subjects. Finally, we performed transfer learning across sequences by fine-tuning the pretrained models on the target sequence with increasing training set sizes to study convergence, while keeping validation and test sets fixed. Segmentations were evaluated using Dice similarity coefficient (DSC) and average surface distance (ASD). Lesion effects were assessed with two-sided Mann-Whitney U tests with Bonferroni correction. Results: Same-sequence training yielded higher accuracy on DESS than Cube (DSC, $0.900$ vs $0.830$; $P < .001$). Cube-to-DESS transfer matched DESS performance (DSC, $0.903 \pm 0.032$ vs $0.900 \pm 0.027$), reaching a performance plateau at 9 training subjects. DESS-to-Cube yielded a lower combined DSC ($0.802 \pm 0.049$ vs $0.830 \pm 0.042$), reaching a plateau at 24 training subjects. Lesions did not affect DESS ($P \ge .39$) but reduced Cube accuracy (DSC, $0.805$ vs $0.856$; $P < .001$). Conclusion: Transfer learning across sequences can substantially reduce target-sequence annotation requirements for femoral cartilage segmentation, but performance is direction- and sequence-dependent, and the effects of lesions on segmentation may vary across MRI sequences.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。