三阴性乳腺癌超声表型与分割模型比较
单中心回顾性研究纳入102例三阴性乳腺癌,按FUSCC免疫组化替代分型分为LAR、IM、BLIS三型,分析临床病理与超声异质性。在二维超声上比较UNet++、DeepLabV3+、nnU-Net对乳腺原发灶及腋窝淋巴结的自动分割性能,并评估超声与病理大小一致性。
要点
- LAR型患者年龄更大,后方回声特征在三型间存在差异。
- DeepLabV3+与nnU-Net分割乳腺病灶和腋窝淋巴结的Dice数值较优。
- 超声与病理大小平均偏差为-1.28 mm,报告测量时需注意。
核心结论
- 三阴性乳腺癌FUSCC分型病例表现出临床病理和超声异质性。
- 在探索性固定划分队列中,DeepLabV3+与nnU-Net分割性能数值上较优,可作为候选感兴趣区生成方法,待更大独立数据集验证。
研究设计与样本
- 单中心回顾性研究,纳入102例三阴性乳腺癌,均有乳腺病灶超声图像和手动掩膜。
- 其中54例有可评估的腋窝淋巴结图像和掩膜。
- 乳腺病灶固定划分72/15/15,腋窝淋巴结固定划分38/8/8。
- 比较UNet++、DeepLabV3+、nnU-Net,指标为Dice、IoU、Precision、Recall、Specificity。
- 采用Bootstrap置信区间和配对Wilcoxon符号秩检验,92例进行Bland-Altman超声病理大小一致性分析。
关键影像表现/征象
- FUSCC分型84例中,LAR型35例、IM型16例、BLIS型33例。
- 后方回声特征在三阴性乳腺癌FUSCC分型间存在差异。
诊断标准与关键数值
- 乳腺病灶Dice:UNet++为0.7866,DeepLabV3+为0.8698,nnU-Net为0.8704。
- 腋窝淋巴结Dice:UNet++为0.6005,DeepLabV3+为0.8007,nnU-Net为0.8097。
- DeepLabV3+在乳腺病灶和腋窝淋巴结两项任务中均显著优于UNet++,DeepLabV3+与nnU-Net无显著差异。
- 超声与病理大小平均偏差为-1.28 mm。
临床应用与阅片要点
- 三阴性乳腺癌不同FUSCC分型可能具有不同超声后方回声表现,描述时应结合分型信息。
- 自动分割模型生成感兴趣区时可优先考虑DeepLabV3+或nnU-Net,但需进一步独立验证后再用于临床流程。
局限性
- 单中心回顾性设计。
- 样本量有限,尤其腋窝淋巴结可评估病例仅54例。
- 采用固定划分与探索性比较,未在大规模独立数据集验证。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
RATIONALE AND OBJECTIVES: To describe clinicopathologic and ultrasonographic heterogeneity across immunohistochemistry-based FUSCC surrogate subtypes in triple-negative breast cancer (TNBC) and compare 3 automatic segmentation models for primary breast lesions and axillary lymph nodes (ALNs) on two-dimensional ultrasound.
MATERIALS AND METHODS: This single-center retrospective study included 102 TNBC cases with breast lesion ultrasound images and manual masks; 54 also had evaluable ALN images and masks. Subtype analyses included classified luminal androgen receptor (LAR), immunomodulatory (IM), and basal-like immune-suppressed (BLIS) cases. Fixed splits were 72/15/15 for breast lesions and 38/8/8 for ALNs. UNet++, DeepLabV3+, and nnU-Net were compared using Dice, IoU, Precision, Recall, and Specificity. Bootstrap confidence intervals and paired Wilcoxon signed-rank tests were used. Bland-Altman analysis assessed ultrasound-pathology size agreement in 92 cases.
RESULTS: Among 84 classified cases, 35 were LAR, 16 IM, and 33 BLIS. LAR cases were older, and posterior acoustic patterns differed across subtypes. Breast lesion Dice scores were 0.7866, 0.8698, and 0.8704 for UNet++, DeepLabV3+, and nnU-Net; ALN Dice scores were 0.6005, 0.8007, and 0.8097. DeepLabV3+ significantly outperformed UNet++ in both tasks, whereas DeepLabV3+ and nnU-Net did not differ significantly. Mean ultrasound-pathology size bias was -1.28 mm.
CONCLUSION: Classified TNBC cases showed clinicopathologic and ultrasonographic heterogeneity. DeepLabV3+ and nnU-Net showed numerically favorable segmentation performance in this exploratory fixed-split cohort and may be considered candidate region-of-interest (ROI)-generation approaches, pending validation in larger independent datasets.
引用
- Li, Bian, Xia, et al. Ultrasonographic Phenotypes and Segmentation Model Comparison for Primary Breast Lesions and Axillary Lymph Nodes in Triple-negative Breast Cancer. Acad Radiol. 2026. PMID: 42309955. doi:10.1016/j.acra.2026.05.014
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。