CoChat
Sign Up
Cited by 41

Predicting EGFR mutational status from pathology images using a real-world dataset

Pao, James J., Biggs, Mikayla, Duncan, Daniel, Lin, Douglas I., Davis, Richard, Huang, Richard S. P., Ferguson, Donna, Janovitz, Tyler +7 more·2023·Scientific Reports
DOI 10.1038/s41598-023-31284-6PMC PMC10020556
Open this paper in CoChat — free account

Abstract

Treatment of non-small cell lung cancer is increasingly biomarker driven with multiple genomic alterations, including those in the epidermal growth factor receptor ( EGFR ) gene, that benefit from targeted therapies. We developed a set of algorithms to assess EGFR status and morphology using a real-world advanced lung adenocarcinoma cohort of 2099 patients with hematoxylin and eosin (H&E) images exhibiting high morphological diversity and low tumor content relative to public datasets. The best performing EGFR algorithm was attention-based and achieved an area under the curve (AUC) of 0.870, a negative predictive value (NPV) of 0.954 and a positive predictive value (PPV) of 0.410 in a validation cohort reflecting the 15% prevalence of EGFR mutations in lung adenocarcinoma. The attention model outperformed a heuristic-based model focused exclusively on tumor regions, and we show that although the attention model also extracts signal primarily from tumor morphology, it extracts additional signal from non-tumor tissue regions. Further analysis of high-attention regions by pathologists showed associations of predicted EGFR negativity with solid growth patterns and higher peritumoral immune presence. This algorithm highlights the potential of deep learning tools to provide instantaneous rule-out screening for biomarker alterations and may help prioritize the use of scarce tissue for biomarker testing.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41598-023-31284-6

Cites

3 of 3 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Integrating AI-based analytical tools directly into the operative workflow may refine intraoperative judgment by complementing surgical expertise with rapid computational insight.

    “In this way, integrating AI-based analytical tools directly into the operative workflow may refine intraoperative judgment by complementing surgical expertise with rapid computational insight [53].”

    From Artificial Intelligence in Oncologic Thoracic Surgery: Clinical Decision Support and Emerging Applications (2026) · cites this paper as [53]

  • Pao et al.'s study highlights that deep learning models can detect fine-grained pathological patterns during surgery and deliver immediate feedback about tissue characteristics.

    “Their study highlights how deep learning models can detect fine-grained pathological patterns during surgery and deliver immediate feedback about tissue characteristics.”

    From Artificial Intelligence in Oncologic Thoracic Surgery: Clinical Decision Support and Emerging Applications (2026) · cites this paper as [53]

  • Real-time histologic assessment allows surgeons to adjust operative strategy as needed, promoting complete tumor removal while avoiding unnecessary excision of healthy structures.

    “Such real-time histologic assessment allows surgeons to adjust their operative strategy as needed, promoting complete tumor removal while avoiding unnecessary excision of healthy structures.”

    From Artificial Intelligence in Oncologic Thoracic Surgery: Clinical Decision Support and Emerging Applications (2026) · cites this paper as [53]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 28

  1. [1]Hanna, N. H. et al. Therapy for stage IV non-small-cell lung cancer with driver alterations: ASCO and OH (CCO) Joint Guideline Update. J Clin Oncol39, (2021). 33591844
  2. [3]Brown NA Aisner DL Oxnard GR Precision medicine in non-small cell lung cancer: Current standards in pathology and biomarker interpretation Am. Soc. Clin. Oncol. Educ. Book 2018 10.1200/edbk_209089 30231309 · DOI
  3. [4]Kather JN Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer Nat. Med. 2019 25 1054 1056 10.1038/s41591-019-0462-y 31160815 PMC7423299 · DOI · Paper page
  4. [5]Kather JN Pan-cancer image-based detection of clinically actionable genetic alterations Nat. Cancer 2020 1 789 799 10.1038/s43018-020-0087-6 33763651 PMC7610412 · DOI
  5. [7]Reddy GT Hybrid genetic algorithm and a fuzzy logic classifier for heart disease diagnosis Evol. Intell. 2020 13 185 196 10.1007/s12065-019-00327-1 · DOI
  6. [8]Holmberg OG Self-supervised retinal thickness prediction enables deep learning from unlabelled data to boost classification of diabetic retinopathy Nat. Mach. Intell. 2020 2 719 726 10.1038/s42256-020-00247-1 · DOI
  7. [9]Anegondi, N., Yang, Q., Kawczynski, M., Steffen, V. & Gao, S. S. Predicting geographic atrophy growth rate from fundus autofluorescence images using deep neural networks. in (2021). 10.1117/12.2575898. · DOI
  8. [10]Arcadu F Deep learning algorithm predicts diabetic retinopathy progression in individual patients NPJ Digit. Med. 2019 2 92 10.1038/s41746-019-0172-3 31552296 PMC6754451 · DOI · Paper page
  9. [11]Wang G Zhang Y Hegde SS Bottomley PA High-resolution and accelerated multi-parametric mapping with automated characterization of vessel disease using intravascular MRI J. Cardiovasc. Magn. Resonance 2017 19 1 13 10.1186/s12968-017-0399-6 PMC5694914 29157260 · DOI
  10. [13]Wang, D., Khosla, A., Gargeya, R., Irshad, H. & Beck, A. H. Deep Learning for Identifying Metastatic Breast Cancer. https://arxiv.org/pdf/1606.05718.pdf (2016). · arXiv
  11. [14]Coudray N Classification and mutation prediction from non–small cell lung cancer histopathology images using deep learning Nat. Med. 2018 24 1559 1567 10.1038/s41591-018-0177-5 30224757 PMC9847512 · DOI · Paper page
  12. [15]Ilse, M., Tomczak, J. M. & Welling, M. Attention-based deep multiple instance learning. In 35th International Conference on Machine Learning, ICML 2018 vol. 5 (2018).
  13. [16]Campanella G Clinical-grade computational pathology using weakly supervised deep learning on whole slide images Nat. Med. 2019 25 1301 1309 10.1038/s41591-019-0508-1 31308507 PMC7418463 · DOI · Paper page
  14. [17]Naik N Deep learning-enabled breast cancer hormonal receptor status determination from base-level H&E stains Nat. Commun. 2020 11 5727 10.1038/s41467-020-19334-3 33199723 PMC7670411 · DOI
  15. [20]National Institude of Health The Cancer Genome Atlas (TCGA). dbGaP Study Accession (phs000178.v11.p8). https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000178.v11.p8
  16. [21]Agrawal R Imieliński T Swami A Mining association rules between sets of items in large databases ACM SIGMOD Rec. 1993 22 207 216 10.1145/170036.170072 · DOI · Paper page
  17. [22]Sun PL High incidence of EGFR mutations in Korean men smokers with no intratumoral heterogeneity of lung adenocarcinomas: Correlation with histologic subtypes, EGFR/TTF-1 expressions, and clinical features J. Thorac. Oncol. 2012 7 323 330 10.1097/JTO.0b013e3182381515 22237264 · DOI
  18. [23]Motoi N Lung adenocarcinoma: Modification of the 2004 WHO mixed subtype to include the major histologic subtype suggests correlations between papillary and micropapillary adenocarcinoma subtypes, EGFR mutations and gene expression analysis Am. J. Surg. Pathol. 2008 32 810 827 10.1097/PAS.0b013e31815cb162 18391747 · DOI
  19. [24]Shim HS Lee DH Park EJ Kim SH Histopathologic characteristics of lung adenocarcinomas with epidermal growth factor receptor mutations in the international association for the study of lung cancer/american thoracic society/european respiratory society lung adenocarcinoma classification Arch. Pathol. Lab. Med. 2011 135 1329 1334 10.5858/arpa.2010-0493-OA 21970488 · DOI
  20. [25]Ninomiya H Correlation between morphology and EGFR mutations in lung adenocarcinomas: Significance of the micropapillary pattern and the hobnail cell type Lung Cancer 2009 63 235 240 10.1016/j.lungcan.2008.04.017 18571764 · DOI
  21. [26]Dong ZY EGFR mutation correlates with uninflamed phenotype and weak immunogenicity, causing impaired response to PD-1 blockade in non-small cell lung cancer Oncoimmunology 2017 6 e1356145 10.1080/2162402X.2017.1356145 29147605 PMC5674946 · DOI · Paper page
  22. [27]Lin, A., Wei, T., Meng, H., Luo, P. & Zhang, J. Role of the dynamic tumor microenvironment in controversies regarding immune checkpoint inhibitors for the treatment of non-small cell lung cancer (NSCLC) with EGFR mutations. Molecular Cancer vol. 18 Preprint at 10.1186/s12943-019-1062-7 (2019). PMC6745797 31526368 · DOI
  23. [28]Irvine AF Waise S Green EW Stuart B Thomas GJ Characterising cancer-associated fibroblast heterogeneity in non-small cell lung cancer: A systematic review and meta-analysis Sci. Rep. 2021 11 1 5 10.1038/s41598-021-81796-2 33580106 PMC7881148 · DOI
  24. [29]Sakai T Link between tumor-promoting fibrous microenvironment and an immunosuppressive microenvironment in stage I lung adenocarcinoma Lung Cancer 2018 126 64 71 10.1016/j.lungcan.2018.10.021 30527194 · DOI
  25. [30]Kalluri, R. & Zeisberg, M. Fibroblasts in cancer. Nature Reviews Cancer vol. 6 Preprint at 10.1038/nrc1877 (2006). 16572188 · DOI · Paper page
  26. [31]Mariathasan S TGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells Nature 2018 554 544 548 10.1038/nature25501 29443960 PMC6028240 · DOI · Paper page
  27. [34]Kingma, D. P. & Ba, J. L. Adam: A method for stochastic optimization. in 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings (2015).
  28. [35]Wu, Y. & Johnson, J. Rethinking ‘Batch’ in BatchNorm. https://arxiv.org/pdf/2105.07576.pdf (2021). · arXiv

Cited by

Sorted by citations

Authors
Pao, J. J., Biggs, M., Duncan, D., Lin, D. I., Davis, R., Huang, R. S. P., Ferguson, D., Janovitz, T., Hiemenz, M. C., Eddy, N. R., Lehnert, E., Cabili, M. N., Frampton, G. M., Hegde, P. S., & Albacker, L. A.
Venue
Scientific Reports
Publisher
Springer Science and Business Media LLC
Published
2023-03-16
Volume
13
Issue
1
ISSN
2045-2322
DOI
10.1038/s41598-023-31284-6
PMC
PMC10020556
Cited by
41
References
3
Language
en
Licence

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review