arXiv 2408.158232024-12-08eess.IVcs.CV

作为弱监督计算病理学特征提取器的基础模型基准测试

Benchmarking foundation models as feature extractors for weakly-supervised computational pathology

Technical University Dresden（德累斯顿工业大学）；StratifAI GmbH（StratifAI有限公司）；University Hospital and Faculty of Medicine Carl Gustav Carus, Technische Universität Dresden（德累斯顿工业大学卡尔·古斯塔夫·卡鲁斯医学院附属医院）；National Center for Tumor Diseases Dresden (NCT/UCC)（德累斯顿国家肿瘤疾病中心（NCT/UCC））；German Cancer Research Center (DKFZ)（德国癌症研究中心（DKFZ））；National Center for Tumor Diseases (NCT)（国家肿瘤疾病中心（NCT））；German Cancer Consortium (DKTK)（德国癌症联盟（DKTK））；Kepler University Hospital, Johannes Kepler University Linz（约翰内斯·开普勒大学林茨开普勒大学医院）；University of Bern（伯尔尼大学）；University Hospital Schleswig-Holstein（石勒苏益格-荷尔斯泰因大学医院）；University Medical Center Mainz（美因茨大学医学中心）；University Hospital Aachen（亚琛大学医院）；European Institute of Oncology IRCCS（欧洲肿瘤研究所IRCCS）；University Hospital Heidelberg（海德堡大学医院）

Peter Neidlinger, Omar S. M. El Nahhas, Hannah Sophie Muti, Tim Lenz, Michael Hoffmeister, Hermann Brenner, Marko van Treeck, Rupert Langer, Bastian Dislich, Ha… 展开作者

Peter Neidlinger, Omar S. M. El Nahhas, Hannah Sophie Muti, Tim Lenz, Michael Hoffmeister, Hermann Brenner, Marko van Treeck, Rupert Langer, Bastian Dislich, Hans Michael Behrens, Christoph Röcken, Sebastian Foersch, Daniel Truhn, Antonio Marra, Oliver Lester Saldanha, Jakob Nikolas Kather

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本研究在多癌种大规模临床队列上基准测试19个组织病理学基础模型，发现视觉语言模型CONCH表现最优，且模型特征融合可提升性能，证实数据多样性比数据量对模型更重要，为病理学基础模型改进指明方向。

摘要：

人工智能的发展推动了众多能够提取临床相关信息的病理学基础模型的研发。然而，目前鲜有文献在真正的外部队列和临床相关任务上独立评估这些基础模型，以探明未来改进的调整方向。本研究中，我们在涵盖肺癌、结直肠癌、胃癌和乳腺癌的13个患者队列（共6818名患者、9528张切片）上对19个组织病理学基础模型开展基准测试。评估围绕生物标志物、形态学特征和预后结局相关的弱监督任务进行。研究表明，视觉-语言基础模型CONCH（一种视觉语言病理学基础模型）的表现优于仅视觉的基础模型，Virchow2紧随其后位居第二。实验显示，在不同队列上训练的基础模型会学习到互补特征以预测同一标签，将这些特征融合后性能可超越当前最优水平。结合CONCH和Virchow2预测结果的集成模型在55%的任务中表现优于单个模型，充分利用了二者在分类场景下的互补优势。此外，研究结果表明对于基础模型而言，数据多样性比数据量更重要。本研究明确了病理学基础模型改进的可行调整方向。

英文摘要：

Advancements in artificial intelligence have driven the development of numerous pathology foundation models capable of extracting clinically relevant information. However, there is currently limited literature independently evaluating these foundation models on truly external cohorts and clinically-relevant tasks to uncover adjustments for future improvements. In this study, we benchmarked 19 histopathology foundation models on 13 patient cohorts with 6,818 patients and 9,528 slides from lung, colorectal, gastric, and breast cancers. The models were evaluated on weakly-supervised tasks related to biomarkers, morphological properties, and prognostic outcomes. We show that a vision-language foundation model, CONCH, yielded the highest performance when compared to vision-only foundation models, with Virchow2 as close second. The experiments reveal that foundation models trained on distinct cohorts learn complementary features to predict the same label, and can be fused to outperform the current state of the art. An ensemble combining CONCH and Virchow2 predictions outperformed individual models in 55% of tasks, leveraging their complementary strengths in classification scenarios. Moreover, our findings suggest that data diversity outweighs data volume for foundation models. Our work highlights actionable adjustments to improve pathology foundation models.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。