Papers

研究剖析病理基础模型在非典型乳腺病变分类上的失效原因

Where slide-level foundation-model classifiers fail on atypical breast lesions

Why we picked it

做计算病理的人该看看这篇:它用 BRACS 和四种病理基础模型拆出了失效根源——不是模型不行,是切片级标签只有 39% 对得上病灶。病灶级标签训练就能救回来,方法可直接搬到自己的分诊流水线。

一项预注册研究发现,基于病理基础模型(PFM)的切片级分类器在非典型导管增生(ADH)与低级别导管原位癌的鉴别上表现最差,而这一判别正日益决定是否手术。研究者在 BRACS 数据集上用四种 PFM 做交叉验证,发现由于切片按最严重病变标注,仅 39% 的含 ADH 标注的切片被标为 ADH,标签噪声是主要失效来源之一。注意力机制虽然能将病灶排在较高位置,但很少排第一;只把病灶喂给模型也帮助有限。

Body

相比之下,用病灶级标签按决策逐项训练的分类器改善了全部四个非典型边界。将 ADH 区域插入良性切片后,模型更倾向上调恶性概率而非非典型概率。在一家由病理医生标注单一亚型切片的医院数据及 HISTAI 上,良性切片得分更接近病灶,分诊特异性从 0.48-0.53 跌至 0.00-0.07 和 0.22。论文指出标签粒度而非模型能力是当前 PFM 切片级分类的瓶颈。

Original · arXiv cs.AI

Where slide-level foundation-model classifiers fail on atypical breast lesions

Whether a ductal breast lesion is called atypical hyperplasia or low-grade carcinoma in situ increasingly determines surgery. However, slide-level classifiers built on pathology foundation models (PFMs) perform worst on these lesions. Using four PFMs, cross-validation on BRACS and preregistered hypotheses, we studied where the classification of atypia fails. Because slides are labelled with their most severe lesion, only 39% of the slides with annotated atypical ductal hyperplasia (ADH) were labelled ADH. Attention ranked the lesion high but rarely first, and showing the model only the lesion helped little, whereas classifiers trained per decision with lesion-level labels improved all four atypical boundaries. ADH regions inserted into benign slides raised the probability of malignancy more than that of atypia. In a hospital with pathologist-labelled single-subtype slides and in HISTAI, benign slides were scored closer to lesions and triage specificity fell from 0.48-0.53 to 0.00-0.07 and 0.22.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk. Key entities and technical topics under continuous tracking include BRACS, ADH, medical AI.

Source attribution: curated from arXiv cs.AI (@Nerea Hernandez, Luis J. Herrera, Cesar Ramirez, Olga Valenzuela, Ignacio Rojas, Francisco M. Ortuño).