发表机构
Nanjing University of Information Science and Technology; Taizhou University; Harbin Medical University（南京信息工程大学; 泰州大学; 哈尔滨医科大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究针对计算病理学全模态切片表征学习的模态坍塌问题，提出Φ-Omni协同信息解缠框架，经多数据集预训练后在少样本任务中优于基线方法。
AI 中文摘要
在计算病理学（CPath）领域，开发整合组织学、基因组学与临床报告的全模态自监督学习（SSL）模型，可为全切片图像（WSI）提供可迁移的表征学习能力。现有方法通过对比对齐将异质模态隐含地强制映射到统一隐空间，导致模态坍塌，即独特的协同诊断信号（记为Φ）被舍弃，转而保留琐碎冗余信息。我们假设，最强的任务无关SSL训练信号源于对协同交互的提炼，而非仅对齐共享冗余。为此，我们提出基于部分信息分解（PID）理论的协同信息解缠框架——Φ-Omni，用于切片表征学习。与标准对比方法不同，Φ-Omni采用由所提ΦID目标调控的协同信息瓶颈（SIB），该目标明确抑制边缘冗余，同时最大化不可约协同性，从而提炼高阶跨模态交互。在乳腺癌队列（n=1031）与肺癌队列（n=919）上进行预训练后，Φ-Omni在涵盖8项任务的5个独立外部数据集上，相较于监督与SSL基线方法，展现出更优的少样本性能。源代码可在此获取。
英文摘要
In computational pathology (CPath), developing omni-modal self-supervised learning (SSL) models that integrate histology, genomics, and clinical reports enables transferable representation learning for whole slide images (WSIs). Existing approaches implicitly force heterogeneous modalities into a uniform latent space by contrastive alignment, causing modality collapse where unique, synergistic diagnostic signals (termed as $\mathrmΦ$) are discarded in favor of trivial redundancy. We hypothesize that the strongest task-agnostic SSL training signal stems from distilling the synergistic interactions over merely aligning shared redundancy. To this end, we introduce \textsc{$\mathrmΦ$-Omni}, a synergistic information disentanglement framework grounded in Partial Information Decomposition (PID) theory for slide representation learning. Unlike standard contrastive approaches, \textsc{$\mathrmΦ$-Omni} employs a Synergistic Information Bottleneck (SIB) regulated by the proposed $\mathrmΦ\text{ID}$ objective, which explicitly suppresses marginal redundancy while maximizing irreducible synergy, thereby distilling high-order cross-modal interactions. Following pretraining on breast ($n$=1031) and lung ($n$=919) cohorts, \textsc{$\mathrmΦ$-Omni} demonstrates superior few-shot performance across five independent external datasets spanning eight tasks compared to supervised and SSL baselines. Source code is available here.
CommentsNeeds further revision
