Machine learning can aid in prediction of IDH mutation from H&E-stained histology slides in infiltrating gliomas.
Journal
Scientific reports
ISSN: 2045-2322
Titre abrégé: Sci Rep
Pays: England
ID NLM: 101563288
Informations de publication
Date de publication:
31 12 2022
31 12 2022
Historique:
received:
30
06
2022
accepted:
12
12
2022
entrez:
31
12
2022
pubmed:
1
1
2023
medline:
4
1
2023
Statut:
epublish
Résumé
While Machine Learning (ML) models have been increasingly applied to a range of histopathology tasks, there has been little emphasis on characterizing these models and contrasting them with human experts. We present a detailed empirical analysis comparing expert neuropathologists and ML models at predicting IDH mutation status in H&E-stained histology slides of infiltrating gliomas, both independently and synergistically. We find that errors made by neuropathologists and ML models trained using the TCGA dataset are distinct, representing modest agreement between predictions (human-vs.-human κ = 0.656; human-vs.-ML model κ = 0.598). While no ML model surpassed human performance on an independent institutional test dataset (human AUC = 0.901, max ML AUC = 0.881), a hybrid model aggregating human and ML predictions demonstrates predictive performance comparable to the consensus of two expert neuropathologists (hybrid classifier AUC = 0.921 vs. two-neuropathologist consensus AUC = 0.920). We also show that models trained at different levels of magnification exhibit different types of errors, supporting the value of aggregation across spatial scales in the ML approach. Finally, we present a detailed interpretation of our multi-scale ML ensemble model which reveals that predictions are driven by human-identifiable features at the patch-level.
Identifiants
pubmed: 36587030
doi: 10.1038/s41598-022-26170-6
pii: 10.1038/s41598-022-26170-6
pmc: PMC9805452
doi:
Substances chimiques
Isocitrate Dehydrogenase
EC 1.1.1.41
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
22623Informations de copyright
© 2022. The Author(s).
Références
Capper, D. et al. DNA methylation-based classification of central nervous system tumours. Nature 555, 469–474. https://doi.org/10.1038/nature26000 (2018).
doi: 10.1038/nature26000
Lastowska, M. et al. Molecular identification of CNS NB-FOXR2, CNS EFT-CIC, CNS HGNET-MN1 and CNS HGNET-BCOR pediatric brain tumors using tumor-specific signature genes. Acta Neuropathol. Commun. 8, 105. https://doi.org/10.1186/s40478-020-00984-9 (2020).
doi: 10.1186/s40478-020-00984-9
Johann, P. D. et al. Atypical teratoid/rhabdoid tumors are comprised of three epigenetic subgroups with distinct enhancer landscapes. Cancer Cell 29, 379–393. https://doi.org/10.1016/j.ccell.2016.02.001 (2016).
doi: 10.1016/j.ccell.2016.02.001
Taylor, M. D. et al. Molecular subgroups of medulloblastoma: The current consensus. Acta Neuropathol. 123, 465–472. https://doi.org/10.1007/s00401-011-0922-z (2012).
doi: 10.1007/s00401-011-0922-z
Reinhardt, A. et al. Anaplastic astrocytoma with piloid features, a novel molecular class of IDH wildtype glioma with recurrent MAPK pathway, CDKN2A/B and ATRX alterations. Acta Neuropathol. 136, 273–291 (2018).
doi: 10.1007/s00401-018-1837-8
Miller, K. D. et al. Brain and other central nervous system tumor statistics, 2021. CA Cancer J. Clin. 71, 381–406. https://doi.org/10.3322/caac.21693 (2021).
doi: 10.3322/caac.21693
Ostrom, Q. T. et al. CBTRUS statistical report: primary brain and other central nervous system tumors diagnosed in the united states in 2011–2015. Neuro Oncol. 20, iv1–iv86. https://doi.org/10.1093/neuonc/noy131 (2018).
doi: 10.1093/neuonc/noy131
Stupp, R., Hegi, M. E., Gilbert, M. R. & Chakravarti, A. Chemoradiotherapy in malignant glioma: Standard of care and future directions. J. Clin. Oncol. 25, 4127–4136 (2007).
doi: 10.1200/JCO.2007.11.8554
WHO Classification of Tumours Editorial Board. Central nervous system tumours [Internet]. Lyon (France): International agency for research on cancer; 2021 [cited November 14, 2022]. (WHO classification of tumours series, 5th ed.; vol. 6). Available from: https://tumourclassification.iarc.who.int/chapters/45 . (2021).
Reuss, D. E. et al. IDH mutant diffuse and anaplastic astrocytomas have similar age at presentation and little difference in survival: A grading problem for WHO. Acta Neuropathol. 129, 867–873. https://doi.org/10.1007/s00401-015-1438-8 (2015).
doi: 10.1007/s00401-015-1438-8
Yan, H. et al. IDH1 and IDH2 Mutations in Gliomas. N. Engl. J. Med. 360, 765–773 (2009).
doi: 10.1056/NEJMoa0808710
Horbinski, C., Kofler, J., Kelly, L. M., Murdoch, G. H. & Nikiforova, M. N. Diagnostic use of IDH1/2 mutation analysis in routine clinical testing of formalin-fixed, paraffin-embedded glioma tissues. J. Neuropathol. Exp. Neurol. 68, 1319–1325 (2009).
doi: 10.1097/NEN.0b013e3181c391be
Olar, A. & Aldape, K. D. Using the molecular classification of glioblastoma to inform personalized treatment. J. Pathol. 232, 165–177. https://doi.org/10.1002/path.4282 (2014).
doi: 10.1002/path.4282
Olar, A. et al. IDH mutation status and role of WHO grade and mitotic index in overall survival in grade II-III diffuse gliomas. Acta Neuropathol. 129, 585–596. https://doi.org/10.1007/s00401-015-1398-z (2015).
doi: 10.1007/s00401-015-1398-z
LeCun, Y., Bengio, Y. & Hinton, G. Deep learning. Nature 521, 436–444. https://doi.org/10.1038/nature14539 (2015).
doi: 10.1038/nature14539
He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In: Proceedings of the IEEE conference on computer vision and pattern recognition, 770–778 (2016).
Szegedy, C., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., Rabinovich, A. Going Deeper with Convolutions. In: Proceedings of the IEEE conference on computer vision and pattern recognition, 1–9 (2015).
De Fauw, J. et al. Clinically applicable deep learning for diagnosis and referral in retinal disease. Nat. Med. 24, 1342–1350. https://doi.org/10.1038/s41591-018-0107-6 (2018).
doi: 10.1038/s41591-018-0107-6
Esteva, A. et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature 542, 115–118. https://doi.org/10.1038/nature21056 (2017).
doi: 10.1038/nature21056
Coudray, N. et al. Classification and mutation prediction from non-small cell lung cancer histopathology images using deep learning. Nat. Med. 24, 1559–1567. https://doi.org/10.1038/s41591-018-0177-5 (2018).
doi: 10.1038/s41591-018-0177-5
Diao, J. A. et al. Human-interpretable image features derived from densely mapped cancer pathology slides predict diverse molecular phenotypes. Nat. Commun. 12, 1613. https://doi.org/10.1038/s41467-021-21896-9 (2021).
doi: 10.1038/s41467-021-21896-9
Heather D. Couture, J. S. M., Perou, C. M., Troester, M. A. Marc Niethammer. in International conference on medical image computing and computer-assisted intervention (2018).
Lu, M. Y. et al. AI-based pathology predicts origins for cancers of unknown primary. Nature 594, 106–110. https://doi.org/10.1038/s41586-021-03512-4 (2021).
doi: 10.1038/s41586-021-03512-4
Lu, M. Y. et al. Data-efficient and weakly supervised computational pathology on whole-slide images. Nat. Biomed. Eng. 5, 555–570. https://doi.org/10.1038/s41551-020-00682-w (2021).
doi: 10.1038/s41551-020-00682-w
Xu, Z. et al. Deep learning predicts chromosomal instability from histopathology images. iScience 24, 102394. https://doi.org/10.1016/j.isci.2021.102394 (2021).
doi: 10.1016/j.isci.2021.102394
Campanella, G. et al. Clinical-grade computational pathology using weakly supervised deep learning on whole slide images. Nat. Med. 25, 1301–1309. https://doi.org/10.1038/s41591-019-0508-1 (2019).
doi: 10.1038/s41591-019-0508-1
Hollon, T. C. et al. Near real-time intraoperative brain tumor diagnosis using stimulated Raman histology and deep neural networks. Nat. Med. 26, 52–58. https://doi.org/10.1038/s41591-019-0715-9 (2020).
doi: 10.1038/s41591-019-0715-9
Kather, J. N. et al. Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer. Nat. Med. 25, 1054–1056. https://doi.org/10.1038/s41591-019-0462-y (2019).
doi: 10.1038/s41591-019-0462-y
Raciti, P. et al. Novel artificial intelligence system increases the detection of prostate cancer in whole slide images of core needle biopsies. Mod. Pathol. 33, 2058–2066. https://doi.org/10.1038/s41379-020-0551-y (2020).
doi: 10.1038/s41379-020-0551-y
Mobadersany, P. et al. Predicting cancer outcomes from histology and genomics using convolutional networks. Proc. Natl. Acad. Sci. USA 115, E2970–E2979. https://doi.org/10.1073/pnas.1717139115 (2018).
doi: 10.1073/pnas.1717139115
Huang, G. Lin, Z. Van Der Maaten, L. & Weinberger, K. Q. Densely connected convolutional networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition, 4700–4708 (2017).
Cantero, D. et al. TP53, ATRX alterations, and low tumor mutation load feature IDH-wildtype giant cell glioblastoma despite exceptional ultra-mutated tumors. Neurooncol. Adv. 2, vdz059. https://doi.org/10.1093/noajnl/vdz059 (2020).
doi: 10.1093/noajnl/vdz059
Ceccarelli, M. et al. Molecular profiling reveals biologically discrete subsets and pathways of progression in diffuse glioma. Cell 164, 550–563. https://doi.org/10.1016/j.cell.2015.12.028 (2016).
doi: 10.1016/j.cell.2015.12.028
Ferris, S. P., Hofmann, J. W., Solomon, D. A. & Perry, A. Characterization of gliomas: From morphology to molecules. Virchows Arch. 471, 257–269. https://doi.org/10.1007/s00428-017-2181-4 (2017).
doi: 10.1007/s00428-017-2181-4
Kleinschmidt-DeMasters, B. K., Aisner, D. L., Birks, D. K. & Foreman, N. K. Epithelioid GBMs show a high percentage of BRAF V600E mutation. Am. J. Surg. Pathol. 37, 685–698. https://doi.org/10.1097/PAS.0b013e31827f9c5e (2013).
doi: 10.1097/PAS.0b013e31827f9c5e
Korshunov, A. et al. Histologically distinct neuroepithelial tumors with histone 3 G34 mutation are molecularly similar and comprise a single nosologic entity. Acta Neuropathol. 131, 137–146. https://doi.org/10.1007/s00401-015-1493-1 (2016).
doi: 10.1007/s00401-015-1493-1
Neumann, J. E. et al. Distinct histomorphology in molecular subgroups of glioblastomas in young patients. J. Neuropathol. Exp. Neurol. 75, 408–414. https://doi.org/10.1093/jnen/nlw015 (2016).
doi: 10.1093/jnen/nlw015
Ilse, M. Tomczak, J. & Welling, M. Attention-based deep multiple instance learning. In: International conference on machine learning, 2127–2136 (2018).
Jiang, S., Zanazzi, G. J. & Hassanpour, S. Predicting prognosis and IDH mutation status for patients with lower-grade gliomas using whole slide images. Sci. Rep. 11, 16849. https://doi.org/10.1038/s41598-021-95948-x (2021).
doi: 10.1038/s41598-021-95948-x
Tomczak, K., Czerwinska, P. & Wiznerowicz, M. The cancer genome atlas (TCGA): An immeasurable source of knowledge. Contempl Oncol. (Pozn) 19, A68-77. https://doi.org/10.5114/wo.2014.47136 (2015).
doi: 10.5114/wo.2014.47136
Alex Krizhevsky, I. S. & Hinton, G. E. Imagenet classification with deep convolutional neural networks. Adv. Neural Inf. Proc. Syst. 25, 1097–1105 (2021).
Deng, J., Wei Dong, Richard Socher, Li-Jia Li, Kai Li, Li Fei-Fei. in IEEE Conference on Computer Vision and Pattern Recognition (2009).
Diederik, P. & Kingma, JLB. in International Conference on Learning Representations (San Diego, CA, 2015).
Becht, E. et al. Dimensionality reduction for visualizing single-cell data using UMAP. Nat. Biotechnol. https://doi.org/10.1038/nbt.4314 (2018).
doi: 10.1038/nbt.4314
Leland McInnes, J. H., James Melville. UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction. arXiv preprint arXiv:1802.03426v3 (2020).
Gu, Z., Eils, R. & Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 32, 2847–2849 (2016).
doi: 10.1093/bioinformatics/btw313