文献速递乳腺MRUS

基于纵向超声和MR深度学习的乳腺癌新辅助治疗反应早期无创预测:一项多中心研究

Early and noninvasive prediction of response to neoadjuvant therapy for breast cancer via longitudinal ultrasound and MR deep learning: A multicentre study.
PubMed 文献 · Acad Radiol · 2024-11-13

该多中心回顾性研究纳入三个中心448例乳腺癌患者,提取新辅助化疗前及第2周期后MR和/或超声图像的最大截面深度学习特征,构建堆叠模型整合四种单模态模型与HER2状态,预测病理完全缓解的AUC达0.951-0.979、准确率91.55%-92.65%、敏感度90.63%-93.94%、特异度89.47%-94.44%。

要点

  • 纵向多模态深度学习模型可在新辅助化疗早期无创预测病理完全缓解。
  • 堆叠模型整合四种单模态模型与HER2状态,AUC达0.951-0.979。
  • 不同单模态深度学习模型间诊断性能无显著差异(DeLong检验,p>0.05)。

核心结论

  • 纵向多模态深度学习可能有助于预测病理完全缓解;堆叠模型表现出优良性能,可作为新辅助化疗反应早期无创预测的新工具,从而辅助制定乳腺癌个体化治疗策略。

研究设计与样本

  • 回顾性多中心研究,纳入三个中心共448例乳腺癌患者,回顾新辅助化疗前及第2周期后乳腺MR和/或超声图像,采用ResNet50从肿瘤最大截面提取深度学习特征,经T检验、Pearson相关分析和LASSO回归筛选特征,构建单模态及堆叠模型并评估诊断性能。

诊断标准与关键数值

  • 全部患者病理完全缓解率为36.65%;整合四种单模态深度学习模型与HER2状态的堆叠模型在各队列中AUC为0.951-0.979、准确率为91.55%-92.65%、敏感度为90.63%-93.94%、特异度为89.47%-94.44%;不同单模态深度学习模型间诊断性能无显著差异(DeLong检验,p>0.05)。

临床应用与阅片要点

  • 作者认为纵向多模态深度学习可用于预测病理完全缓解,堆叠模型可作为新辅助化疗反应早期无创预测的新工具,以辅助制定乳腺癌个体化治疗策略。

局限性

  • 原文未报告局限性。

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: The early prediction of response to neoadjuvant chemotherapy (NAC) will aid in the development of personalized treatments for patients with breast cancer. This study investigated the value of longitudinal multimodal deep learning (DL) based on breast MR and ultrasound (US) in predicting pathological complete response (pCR) after NAC.

MATERIALS AND METHODS: We retrospectively reviewed the pre-NAC and post-2nd-NAC MR and/or US images of 448 patients enrolled from three centers and extracted DL features from the largest section of the breast tumour using ResNet50. T test, Pearson correlation analysis and least absolute shrinkage and selection operator regression were used to select the most significant DL features for the pre-NAC and post-2nd-NAC MR and US DL models. The stacking model integrates different single-modality DL models and meaningful clinical data. The diagnostic performance of the models was evaluated.

RESULTS: In all the patients, the pCR rate was 36.65%. There was no significant difference in diagnostic performance between the different single-modality DL models (DeLong test, p > 0.05). The stacking model integrating the above four DL models with HER2 status yielded areas under the curves of 0.951-0.979, accuracies of 91.55%-92.65%, sensitivities of 90.63%-93.94%, and specificities of 89.47%-94.44% in the cohorts.

CONCLUSION: Longitudinal multimodal DL can be useful in predicting pCR. The stacking model can be used as a new tool for the early noninvasive prediction of the response to NAC, as evidenced by its excellent performance, and therefore aid the development of personalized treatment strategies for patients with breast cancer.

来源:Acad Radiol,查看原文。版权归原作者所有。

引用

  1. Zeng Q, Liu L, He C, et al. Early and noninvasive prediction of response to neoadjuvant therapy for breast cancer via longitudinal ultrasound and MR deep learning: A multicentre study. Acad Radiol. 2024. PMID: 39542804. doi:10.1016/j.acra.2024.10.033

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。