PO.BCS01.07 · 生物信息与计算
利用弱监督transMIL-注意力框架从H&E全切片图像预测胃癌的TCGA分子亚型
Predicting TCGA molecular subtypes of gastric cancer from H&E whole-slide images using a weakly supervised transMIL-attention framework
作者与单位 Authors & Affiliations
摘要 Abstract
中文摘要
背景：
癌症基因组图谱(TCGA)已定义了胃癌的四种分子亚型：Epstein-Barr病毒(EBV)相关型、微卫星不稳定(MSI)相关型、基因组稳定型(GS)和染色体不稳定型(CIN)。这些亚型具有不同的临床病理和治疗意义。然而，这些亚型的常规判定仍依赖于成本高昂且并非普遍可及的多模态分子检测。我们旨在开发一种弱监督深度学习框架，直接从苏木精-伊红(H&E)染色全切片图像(WSI)预测TCGA分子亚型。
方法：
作为基线，我们实现了注意力挑战多示例学习(ACMIL)，其利用多分支注意力(MBA)和随机示例掩蔽进行弱监督WSI分类。随后，我们提出了一种混合多示例学习(MIL)模型，将基于transformer的MIL架构(TransMIL)与MBA相结合。TransMIL-MBA的特征和注意力分数通过加权注意力融合，以生成切片级预测。该模型在来自TCGA ESCA和STAD项目、带有胃癌TCGA分子亚型标签的484张20×放大倍数的H&E WSI上进行训练和评估，训练、验证和测试分别采用80%、10%和10%的划分。模型性能采用四分类受试者工作特征曲线下面积(AUC)、准确率、混淆矩阵，以及用UMAP和V-measure对潜在特征空间进行可视化来评估。
结果：
注意力热图显示，与ACMIL相比，所提出的TransMIL-MBA混合模型持续突出组织学相关的肿瘤区域。在TCGA测试集上，TransMIL-MBA模型在四分类亚型分类中优于ACMIL(AUC：0.89对0.87；准确率：0.74对0.67)。该混合模型在UMAP嵌入中也表现出更好的亚型可分性(TransMIL-MBA的V-measure为0.74，ACMIL为0.65)和更清晰的混淆矩阵。
结论：
将TransMIL与MBA相结合的弱监督多示例学习框架在直接从H&E WSI预测胃癌TCGA分子亚型方面表现出令人鼓舞的性能。通过提供一种可扩展的、基于图像的分子分型替代方法，该方法可能有助于使临床诊疗更接近精准医学，实现更有针对性的治疗，并有可能改善预后不良的胃癌患者的结局。
查看英文原文 English abstract
Background:
The Cancer Genome Atlas (TCGA) has defined four molecular subtypes of gastric cancer: Epstein-Barr virus (EBV)- associated, microsatellite instability (MSI)- associated, genomically stable (GS), and chromosomal instability (CIN). These subtypes have distinct clinicopathologic and therapeutic implications. However, the routine determination of these subtypes still relies on multimodal molecular assays that are costly and not universally available. We aimed to develop a weakly supervised deep learning framework that predicts TCGA molecular subtypes directly from hematoxylin and eosin (H&E)-stained whole-slide images (WSIs).
Methods:
As a baseline, we implemented attention-challenging multiple instance learning (ACMIL), which leverages multi-branch attention (MBA) and stochastic instance masking for weakly supervised WSI classification. We then proposed a hybrid multiple instance learning (MIL) model that combines a transformer-based MIL architecture (TransMIL) with an MBA. The features and attention scores from TransMIL-MBA were fused via weighted attention to generate slide-level predictions. The model was trained and evaluated on 484 H&E WSIs at 20× magnification with TCGA molecular subtype labels for gastric cancer from the TCGA ESCA and STAD projects, using an 80%, 10%, and 10% split for training, validation, and testing, respectively. Model performance was assessed using the four-class area under the receiver operating characteristic curve (AUC), accuracy, confusion matrices, and visualization of the latent feature space with UMAP and V-measure.
Results:
Attention heatmaps indicated that the proposed TransMIL-MBA hybrid model consistently highlighted histologically relevant tumor regions compared to ACMIL. On the TCGA test set, the TransMIL-MBA model outperformed ACMIL in four-class subtype classification (AUC: 0.89 vs. 0.87; accuracy: 0.74 vs. 0.67). The hybrid model also showed improved subtype separability in the UMAP embedding (V-measure 0.74 with TransMIL-MBA vs. 0.65 with ACMIL) and clearer confusion matrices.
Conclusions:
A weakly supervised multiple instance learning framework combining TransMIL with MBA shows promising performance for predicting TCGA molecular subtypes of gastric cancer directly from H&E WSIs. By providing a scalable, image-based surrogate method for molecular subtyping, this approach may help move clinical care closer to precision medicine, enabling more tailored treatments and potentially improving outcomes for patients with poor-prognosis gastric cancer.
利益披露 Disclosure
Y. Jeong, None..
D. M. Bappy, None..
S. Ahn, None..
S. Lee, None.