病理基础模型图谱中的信号由什么承载？乳腺癌的患者级对照基准测试
What Carries the Signal in Pathology Foundation-Model Atlases? A Patient-Level Controlled Benchmark in Breast Cancer
浏览论文内容
中文总结 AI 辅助
本研究以乳腺癌患者为单位开展对照基准测试，发现病理基础模型的信号由嵌入而非几何机制承载，嵌入在多数基因程序预测中优于组织组成，驱动基因计数指标无显著预测性。
中文摘要 AI 辅助
已有研究表明病理基础模型可编码组织形态中的分子程序，但证据通常是队列范围内的排名基因列表，而非对保留患者的预测。本研究以患者为证据单位重建此类分析，探究哪个流水线组件承载信号。研究使用11个冻结骨干网络、4个预先指定的基因程序、285例TCGA-BRCA患者（配对切片与RNA-seq，含44个细胞），采用按患者分组的GroupKFold划分，所有预处理在折内完成。对平均池化嵌入的岭回归预测保留程序得分，Spearman相关系数为0.25-0.56，UNI2在四个程序中表现最强（免疫程序为0.556）。匹配的置换零假设显示，每10000次置换的原始p值约为1e-4，经Holm校正后p=0.0044，表明信号真实但并非均匀分布在形态中。在相同患者和折的竞争模型对比中，嵌入在ER/管腔、增殖和免疫程序上优于组织组成（分别提升+0.280、+0.284、+0.479，p≤0.003），但在基底程序上不占优，其中仅区室分数就达到0.469，而嵌入为0.493（p=0.77）。54个可解释的细胞计数特征在每个程序上的表现介于0.043-0.085之间。几何机制无显著贡献，原因在于测地线图通过欧几里得最近邻搜索选择邻居，仅对已选边重新加权，拓扑结构本质为欧几里得（黎曼减去欧几里得的增益为+0.0010，95%置信区间[-0.0007, +0.0029]），一致应用几何后表现更差（-0.0117）。岭回归比图与度量解码器提升+0.097（置信区间[+0.069, +0.127]）。该文献中常见的驱动基因计数指标在此几乎无信息：91.8%的随机6基因面板可恢复≥5/6个驱动基因。
英文摘要
Pathology foundation models are reported to encode molecular programmes in tissue morphology, but the evidence is usually a cohort-wide ranked gene list rather than a prediction for a held-out patient. We rebuild such an analysis with the patient as the unit of evidence and ask which pipeline component carries signal. Across 11 frozen backbones, four pre-specified gene programmes and 285 TCGA-BRCA patients with paired slides and RNA-seq (44 cells; GroupKFold by patient, all preprocessing fitted inside the fold), ridge regression on mean-pooled embeddings predicts held-out programme scores at Spearman rho = 0.25-0.56, UNI2 strongest on all four (immune 0.556). A matched permutation null gives raw p ~ 1e-4 at 10,000 permutations for every cell; Holm-adjusted p = 0.0044. The signal is real but not uniformly morphological. Against competing models on the same patients and folds, embeddings beat tissue composition for ER/luminal, proliferation and immune (+0.280, +0.284, +0.479; p <= 0.003) but not basal, where compartment fractions alone reach 0.469 against the embedding's 0.493 (p = 0.77). Fifty-four interpretable cell-count features come within 0.043-0.085 on every programme. The geometric machinery contributes nothing measurable, and we identify why: the geodesic graph selects neighbours by Euclidean nearest-neighbour search and only reweights edges already chosen, so the topology is Euclidean by construction (Riemannian minus Euclidean = +0.0010, 95% CI [-0.0007, +0.0029]). Applied consistently the geometry is worse (-0.0117). Ridge regression beats the graph-and-metric decoder by +0.097 (CI [+0.069, +0.127]). The driver-count metric common in this literature is near-uninformative here: 91.8% of random six-gene panels recover >=5/6 drivers.
发表机构
- University of Missouri（密苏里大学）
机构由 AI 辅助整理，请以论文原文为准。
