Transformer-based biomarker prediction from colorectal cancer histology: A large-scale multicentric study.
Transformer-based biomarker prediction from colorectal cancer histology: A large-scale multicentric study.
复制标题
DOI：
10.1016/j.ccell.2023.08.002
复制
发表时间：
2023-09-11
期刊：
影响因子：
50.3
通讯作者：
Kather, Jakob Nikolas
中科院分区：
文献类型：
--
作者：
Wagner, Sophia J.;Reisenbuechler, Daniel;West, Nicholas P.;Niehues, Jan Moritz;Zhu, Jiefu;Foersch, Sebastian;Veldhuizen, Gregory Patrick;Quirke, Philip;Grabsch, Heike I.;van den Brandt, Piet A.;Hutchins, Gordon G. A.;Richman, Susan D.;Yuan, Tanwei;Langer, Rupert;Jenniskens, Josien C. A.;Offermans, Kelly;Mueller, Wolfram;Gray, Richard;Gruber, Stephen B.;Greenson, Joel K.;Rennert, Gad;Bonner, Joseph D.;Schmolze, Daniel;Jonnagaddala, Jitendra;Hawkins, Nicholas J.;Ward, Robyn L.;Morton, Dion;Seymour, Matthew;Magill, Laura;Nowak, Marta;Hay, Jennifer;Koelzer, Viktor H.;Church, David N.;Matek, Christian;Geppert, Carol;Peng, Chaolong;Zhi, Cheng;Ouyang, Xiaoming;James, Jacqueline A.;Loughrey, Maurice B.;Salto-Tellez, Manuel;Brenner, Hermann;Hoffmeister, Michael;Truhn, Daniel;Schnabel, Julia A.;Boxberg, Melanie;Peng, Tingying;Kather, Jakob Nikolas
深度学习（DL）可以加速预测结直肠癌（CRC）常规病理切片中的预后生物标志物。然而，目前的方法依赖于卷积神经网络（CNN），并且大多在小型患者队列中得到验证。在这里，我们开发了一种新的基于transformer的管道，通过将预训练的Transformer编码器与用于补丁聚合的Transformer网络相结合，用于从病理切片进行端到端生物标志物预测。我们的变压器为基础的方法大大提高了性能，概括性，数据效率和可解释性相比，目前国家的最先进的算法。在对来自16个结直肠癌队列的超过13，000名患者的大型多中心队列进行培训和评估后，我们实现了0.99的灵敏度，阴性预测值超过0.99，用于预测手术切除标本上的微卫星不稳定性（MSI）。我们证明，切除手术仅培训达到临床级的内镜活检组织的性能，解决了一个长期存在的诊断问题。基于AI的生物标志物预测（MSI，BRAF和KRAS）使用变压器MSI预测达到结肠直肠癌活检的临床级性能基于变压器的生物标志物预测概括得更好，数据效率更高瓦格纳等人对来自16个队列的13，000多名患者进行了大规模多队列评估。显示了与当前最先进算法相比，基于变换器的来自组织学的生物标志物的预测实质上提高了性能、可推广性、数据效率和可解释性。该方法显著优于现有的手术切除中微卫星不稳定性检测方法，并在结直肠癌活检中达到临床级性能，解决了长期存在的诊断问题。
Deep learning (DL) can accelerate the prediction of prognostic biomarkers from routine pathology slides in colorectal cancer (CRC). However, current approaches rely on convolutional neural networks (CNNs) and have mostly been validated on small patient cohorts. Here, we develop a new transformer-based pipeline for end-to-end biomarker prediction from pathology slides by combining a pre-trained transformer encoder with a transformer network for patch aggregation. Our transformer-based approach substantially improves the performance, generalizability, data efficiency, and interpretability as compared with current state-of-the-art algorithms. After training and evaluating on a large multicenter cohort of over 13,000 patients from 16 colorectal cancer cohorts, we achieve a sensitivity of 0.99 with a negative predictive value of over 0.99 for prediction of microsatellite instability (MSI) on surgical resection specimens. We demonstrate that resection specimen-only training reaches clinical-grade performance on endoscopic biopsy tissue, solving a long-standing diagnostic problem. AI-based prediction of biomarkers (MSI, BRAF, and KRAS) using transformers MSI prediction reaches clinical-grade performance on biopsies of colorectal cancer Transformer-based biomarker prediction generalizes better and is more data efficient Large-scale multi-cohort evaluation on over 13,000 patients from 16 cohorts Wagner et al. show that transformer-based prediction of biomarkers from histology substantially improves the performance, generalizability, data efficiency, and interpretability as compared with current state-of-the-art algorithms. The method significantly outperforms existing approaches for microsatellite instability detection in surgical resections and reaches clinical-grade performance on biopsies of colorectal cancer, solving a long-standing diagnostic problem.
登录
查看更多内容
DOI：
10.1056/nejmoa2201445
发表时间：
2022-06-23
期刊：
The New England journal of medicine
影响因子：
--
作者：
通讯作者：
--
在一项基于人群的队列研究中，他汀类药物的使用，候选甲酸盐途径生物标志物和结肠癌存活。
影响因子：
8.8
作者：
Gray RT;Loughrey MB;Bankhead P;Cardwell CR;McQuaid S;O'Neill RF;Arthur K;Bingham V;McGready C;Gavin AT;James JA;Hamilton PW;Salto-Tellez M;Murray LJ;Coleman HG
通讯作者：
Coleman HG
影响因子：
11.5
作者：
Grabsch, H;Dattani, M;Mueller, W
通讯作者：
Mueller, W
影响因子：
78.8
作者：
Bera, Kaustav;Schalper, Kurt A.;Madabhushi, Anant
通讯作者：
Madabhushi, Anant
影响因子：
168.9
作者：
Gray, Richard;Barnwell, Jennifer;Kerr, David J.
通讯作者：
Kerr, David J.