文献速递儿科影像神经MR

多模态Transformer融合影像与文本鉴别GHD与ISS

Multimodal Transformer Fusion of Clinical Information, Medical Text, and Pituitary MRI 2.5D Deep Learning Features for Differentiating Growth Hormone Deficiency and Idiopathic Short Stature.
PubMed 文献 · Acad Radiol · 2026-08-21

本研究回顾性纳入两个中心共618例儿童为训练队列、164例为独立外部测试队列,构建并比较基于垂体MRI的2.5D、2D、3D深度学习模型及临床与医学文本单模态模型,并建立整合三种模态的多模态Transformer融合模型。结果显示2.5D模型诊断性能及泛化能力优于2D和3D模型,MM_Fusion模型最优,训练队列AUC为0.942,外部测试队列AUC为0.896,显著优于所有单模态模型。

要点

  • 融合MRI影像、临床信息和医学文本的多模态模型对GHD与ISS鉴别效能最优。
  • 基于垂体MRI的2.5D深度学习模型优于2D和3D模型。
  • 该工具可作为儿童GHD与ISS鉴别诊断的非侵入性辅助手段。

核心结论

  • 基于垂体MRI的2.5D深度学习模型可有效鉴别儿童生长激素缺乏症与特发性矮小。
  • 多模态Transformer融合模型进一步提高诊断准确性和泛化能力,可为两种疾病的精确鉴别提供可靠的非侵入性辅助工具。
  • MM_Fusion模型在训练队列AUC为0.942,外部测试队列AUC为0.896,显著优于所有单模态模型,且校准性能和临床效用良好。

研究设计与样本

  • 回顾性双中心研究,训练队列纳入618例儿童,独立外部测试队列纳入164例儿童。
  • 构建并比较基于垂体MRI的2.5D、2D和3D深度学习模型。
  • 开发临床数据和医学文本数据的单模态模型,并建立整合三种模态的多模态Transformer融合模型。

诊断标准与关键数值

  • 2.5D深度学习模型诊断性能及泛化能力显著优于2D和3D模型。
  • MM_Fusion模型训练队列AUC为0.942,外部测试队列AUC为0.896。
  • MM_Fusion模型显著优于所有单模态模型。

临床应用与阅片要点

  • 可将垂体MRI 2.5D深度学习模型作为鉴别儿童GHD与ISS的辅助影像分析工具。
  • 整合临床信息和医学文本的多模态融合方案值得在诊断流程中进一步验证和应用。

局限性

  • 研究为回顾性设计。

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 develop a non-invasive, efficient and accurate auxiliary tool for the precise differential diagnosis between pediatric growth hormone deficiency (GHD) and idiopathic short stature (ISS).

MATERIALS AND METHODS: This retrospective two-center study enrolled 618 children as the internal training cohort and 164 children as the independent external test cohort. We constructed and compared 2.5D, 2D and 3D deep learning (DL) models based on pituitary MRI, developed unimodal models for clinical and medical text data, and established a multimodal Transformer fusion (MM_Fusion) model integrating the three modalities, with systematic evaluation of its diagnostic efficacy.

RESULTS: The 2.5D DL model showed significantly better diagnostic performance and generalization ability than 2D and 3D models. The MM_Fusion model achieved the optimal efficacy, with an AUC of 0.942 in the training cohort and 0.896 in the external test cohort, significantly outperforming all unimodal models, along with excellent calibration performance and clinical utility.

CONCLUSION: The pituitary MRI-based 2.5D DL model can effectively differentiate pediatric GHD and ISS. The multimodal Transformer fusion model further improves diagnostic accuracy and generalization, providing a reliable non-invasive auxiliary tool for the precise differential diagnosis of the two diseases.

来源:Acad Radiol,查看原文。版权归原作者所有。

引用

  1. Zhang, Jia, Lin, et al. Multimodal Transformer Fusion of Clinical Information, Medical Text, and Pituitary MRI 2.5D Deep Learning Features for Differentiating Growth Hormone Deficiency and Idiopathic Short Stature. Acad Radiol. 2026. PMID: 42629286. doi:10.1016/j.acra.2026.08.015

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。