Artificial intelligence for detection of microsatellite instability in colorectal cancer-a multicentric analysis of a pre-screening tool for clinical application.
Artificial intelligence for detection of microsatellite instability in colorectal cancer-a multicentric analysis of a pre-screening tool for clinical application.
复制标题
DOI：
10.1016/j.esmoop.2022.100400
复制
发表时间：
2022-04
期刊：
影响因子：
7.3
通讯作者：
Kather, J. N.
中科院分区：
文献类型：
--
作者：
Echle, A.;Laleh, N. Ghaffari;Quirke, P.;Grabsch, H., I;Muti, H. S.;Saldanha, O. L.;Brockmoeller, S. F.;van den Brandt, P. A.;Hutchins, G. G. A.;Richman, S. D.;Horisberger, K.;Galata, C.;Ebert, M. P.;Eckardt, M.;Boutros, M.;Horst, D.;Reissfelder, C.;Alwers, E.;Brinker, T. J.;Langer, R.;Jenniskens, J. C. A.;Offermans, K.;Mueller, W.;Gray, R.;Gruber, S. B.;Greenson, J. K.;Rennert, G.;Bonner, J. D.;Schmolze, D.;Chang-Claude, J.;Brenner, H.;Trautwein, C.;Boor, P.;Jaeger, D.;Gaisa, N. T.;Hoffmeister, M.;West, N. P.;Kather, J. N.
关键词：
微卫星不稳定性(MSI)/错配修复缺陷(DMMR)是结直肠癌(CRC)患者的关键遗传特征，应根据医学指南对其进行检测。人工智能(AI)方法可以直接在常规病理切片中检测MSI/dMMR，但尚未系统地研究其测试性能。我们培训和验证了基于AI的MSI/dMMR检测器，并使用来自不同国家和种族的8343名患者的9个患者队列评估了预定义的性能指标。分类器实现了临床级别的性能，在不使用任何手动注释的情况下，接收器操作曲线(AUROC)下的面积高达0.96。随后，我们发现AI系统可以作为一种排除测试：通过使用队列特定的阈值，每个手术队列中平均52.73%的肿瘤[MSI/dMMR总数=1020，微卫星稳定(MSS)/熟练错配修复(PMMR)=7323名患者]可以被识别为MSS/pMMR，灵敏度固定在95%。在另一组N=1530(MSI/dMMR=211，MSS/pMMR=1319)的内窥镜活检样本中，系统获得的AUROC为0.89，队列特异性阈值排除了44.12%的肿瘤，其固定敏感度为95%。作为队列特定阈值的一个更可靠的替代，我们表明，对于所有队列，如果阈值固定为0.25，我们可以排除手术标本中25.51%和活检中6.10%的可能性。当应用于临床时，这意味着人工智能系统可以在四分之一(具有全局阈值)或一半的CRC患者(通过局部微调)排除MSI/dMMR，从而降低分子图谱的成本和周转时间。我们描述了迄今为止最大的人工智能在分子癌症诊断中的多中心研究。我们的开源AI系统实现了对MSI检测的高预筛性能。我们的数据突显了大规模验证AI生物标记物在不同地理和种族人群中的作用。
Microsatellite instability (MSI)/mismatch repair deficiency (dMMR) is a key genetic feature which should be tested in every patient with colorectal cancer (CRC) according to medical guidelines. Artificial intelligence (AI) methods can detect MSI/dMMR directly in routine pathology slides, but the test performance has not been systematically investigated with predefined test thresholds. We trained and validated AI-based MSI/dMMR detectors and evaluated predefined performance metrics using nine patient cohorts of 8343 patients across different countries and ethnicities. Classifiers achieved clinical-grade performance, yielding an area under the receiver operating curve (AUROC) of up to 0.96 without using any manual annotations. Subsequently, we show that the AI system can be applied as a rule-out test: by using cohort-specific thresholds, on average 52.73% of tumors in each surgical cohort [total number of MSI/dMMR = 1020, microsatellite stable (MSS)/ proficient mismatch repair (pMMR) = 7323 patients] could be identified as MSS/pMMR with a fixed sensitivity at 95%. In an additional cohort of N = 1530 (MSI/dMMR = 211, MSS/pMMR = 1319) endoscopy biopsy samples, the system achieved an AUROC of 0.89, and the cohort-specific threshold ruled out 44.12% of tumors with a fixed sensitivity at 95%. As a more robust alternative to cohort-specific thresholds, we showed that with a fixed threshold of 0.25 for all the cohorts, we can rule-out 25.51% in surgical specimens and 6.10% in biopsies. When applied in a clinical setting, this means that the AI system can rule out MSI/dMMR in a quarter (with global thresholds) or half of all CRC patients (with local fine-tuning), thereby reducing cost and turnaround time for molecular profiling. We describe the largest multicentric study of AI in molecular cancer diagnostics to date. Our open-source AI system achieves a high pre-screening performance for detection of MSI. Our data highlight the role of large-scale validation of AI biomarkers in different geographic and ethnic populations.
登录
查看更多内容
影响因子：
29.4
作者：
Boland CR;Goel A
通讯作者：
Goel A
影响因子：
11.5
作者：
Grabsch, H;Dattani, M;Mueller, W
通讯作者：
Mueller, W
影响因子：
2.7
作者：
Lee, Hyunseok;Seo, Jihyun;Hong, Ayoung
通讯作者：
Hong, Ayoung
影响因子：
168.9
作者：
Gray, Richard;Barnwell, Jennifer;Kerr, David J.
通讯作者：
Kerr, David J.
影响因子：
22.7
作者：
Kather JN;Heij LR;Grabsch HI;Loeffler C;Echle A;Muti HS;Krause J;Niehues JM;Sommer KA;Bankhead P;Kooreman LF;Schulte JJ;Cipriani NA;Buelow RD;Boor P;Ortiz-Brüchle NN;Hanby AM;Speirs V;Kochanny S;Patnaik A;Srisuwananukorn A;Brenner H;Hoffmeister M;van den Brandt PA;Jäger D;Trautwein C;Pearson AT;Luedde T
通讯作者：
Luedde T