Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Federated learning for predicting histological response to neoadjuvant chemotherapy in triple-negative breast cancer

Abstract

Triple-negative breast cancer (TNBC) is a rare cancer, characterized by high metastatic potential and poor prognosis, and has limited treatment options. The current standard of care in nonmetastatic settings is neoadjuvant chemotherapy (NACT), but treatment efficacy varies substantially across patients. This heterogeneity is still poorly understood, partly due to the paucity of curated TNBC data. Here we investigate the use of machine learning (ML) leveraging whole-slide images and clinical information to predict, at diagnosis, the histological response to NACT for early TNBC women patients. To overcome the biases of small-scale studies while respecting data privacy, we conducted a multicentric TNBC study using federated learning, in which patient data remain secured behind hospitals’ firewalls. We show that local ML models relying on whole-slide images can predict response to NACT but that collaborative training of ML models further improves performance, on par with the best current approaches in which ML models are trained using time-consuming expert annotations. Our ML model is interpretable and is sensitive to specific histological patterns. This proof of concept study, in which federated learning is applied to real-world datasets, paves the way for future biomarker discovery using unprecedentedly large datasets.

This is a preview of subscription content, access via your institution

Access options

Buy article

Get time limited or full article access on ReadCube.

$32.00

All prices are NET prices.

Fig. 1: TNBC treatment pathway and interhospital collaborative FL study overview.
Fig. 2: Univariate and multivariate associations between diagnosis variables and NACT histological response.
Fig. 3: Performance comparison between local and collaborative FL training on clinical or WSI data to predict treatment response in TNBC.
Fig. 4: Performance of WSI-based methods versus the best models using clinical data on the separate test.
Fig. 5: Interpretation of the WSI-based collaborative FL model.

Data availability

Data underpinning this study are under restricted access and are not freely available as they contain patients’ data, and specific clearance from the ethics committee is required in each center. Data can, however, be made available upon reasonable request from Centre Léon Bérard, Institut Curie, Institut Gustave Roussy and Institut Universitaire du Cancer Toulouse (IUCT) Oncopole. Specific conditions and restrictions of access to the datasets are to be discussed directly with the main investigators in each center: P.E.H. for Centre Léon Bérard, G.B. or A. Leopold for Institut Curie, C.F. for IUCT Oncopole and M.L.-T. for Institut Gustave Roussy.

Code availability

All federated learning experiments relied on Substra, an open source software available at https://github.com/Substra/substra. The scripts to run these experiments are not public as they rely heavily on the intellectual property of Owkin, Inc. and on proprietary libraries that are used outside this research project. Those libraries include the TilingTool, a tool to preprocess and tile raw whole-slide images; ClassicAlgos, a library of weakly supervised machine learning models; and Ruche, a library to launch specific federated learning algorithms using the Substra framework. Detailed algorithmic information is presented in the Methods section and should be sufficient to replicate all experiments. For any questions regarding the replication of the experiments, the corresponding author J.O.d.T. should be contacted.

References

  1. Portha, H. et al. Nonmetastatic triple-negative breast cancer in 2016: definitions and management. Gynecol. Obstet. Fertil. 44, 492–504 (2016).

    Article  CAS  Google Scholar 

  2. Hortobagyi, G. N. Treatment of breast cancer. N. Engl. J. Med. 339, 974–984 (1998).

    Article  CAS  Google Scholar 

  3. Waks, A. G. & Winer, E. P. Breast cancer treatment: a review. JAMA 321, 288–300 (2019).

    Article  CAS  Google Scholar 

  4. Penault-Llorca, F. et al. 2014 update of the GEFPICS’ recommendations for HER2 status determination in breast cancers in France. Ann. Pathol. 34, 352–365 (2014).

    Article  Google Scholar 

  5. Fujii, T. et al. New threshold of ER positivity in early stage HER2− breast cancer. J. Clin. Oncol. 34(15_suppl), 1067 (2016).

    Article  Google Scholar 

  6. Moo, T.-A., Sanford, R., Dang, C. & Morrow, M. Overview of breast cancer therapy. PET Clin. 13, 339–354 (2018).

    Article  Google Scholar 

  7. Yin, L., Duan, J.-J., Bian, X.-W. & Yu, S.-C. Triple-negative breast cancer molecular subtyping and treatment progress. Breast Cancer Res. 22, 61 (2020).

    Article  Google Scholar 

  8. Sakuma, K. et al. Pathological tumor response to neoadjuvant chemotherapy using anthracycline and taxanes in patients with triple-negative breast cancer. Exp. Ther. Med. 2, 257–264 (2011).

    Article  Google Scholar 

  9. Fraser Symmans, W. et al. Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy. J. Clin. Oncol. 25, 4414–4422 (2007).

    Article  Google Scholar 

  10. Pandy, J. G. P., Balolong-Garcia, J. C., Cruz-Ordinario, M. V. B. & Que, F. V. F. Triple negative breast cancer and platinum-based systemic treatment: a meta-analysis and systematic review. BMC Cancer 19, 1065 (2019).

    Article  Google Scholar 

  11. Hwang, S.-Y., Park, S. & Kwon, Y. Recent therapeutic trends and promising targets in triple negative breast cancer. Pharmacol. Ther. 199, 30–57 (2019).

    Article  CAS  Google Scholar 

  12. Vikas, P., Borcherding, N. & Zhang, W. The clinical promise of immunotherapy in triple-negative breast cancer. Cancer Manag. Res. 10, 6823–6833 (2018).

    Article  CAS  Google Scholar 

  13. Schmid, P. et al. Event-free survival with pembrolizumab in early triple-negative breast cancer. N. Engl. J. Med. 386, 556–567 (2022).

    Article  CAS  Google Scholar 

  14. Gass, P. et al. Prediction of pathological complete response and prognosis in patients with neoadjuvant treatment for triple-negative breast cancer. BMC Cancer 18, 1051 (2018).

    Article  CAS  Google Scholar 

  15. Abuhadra, N. et al. Beyond TILs: predictors of pathologic complete response (pCR) in triple-negative breast cancer (TNBC) patients with moderate tumor-infiltrating lymphocytes (TIL) receiving neoadjuvant therapy. J. Clin. Oncol. 37(15_suppl), 572 (2019).

    Article  Google Scholar 

  16. Jung, Y. Y. et al. Histomorphological factors predicting the response to neoadjuvant chemotherapy in triple-negative breast cancer. J. Breast Cancer 19, 261–267 (2016).

    Article  Google Scholar 

  17. Mao, Y. et al. The prognostic value of tumor-infiltrating lymphocytes in breast cancer: a systematic review and meta-analysis. PLoS One 11, e0152500 (2016).

    Article  Google Scholar 

  18. Stanton, S. E. & Disis, M. L. Clinical significance of tumor-infiltrating lymphocytes in breast cancer. J. Immunother. Cancer 4, 59 (2016).

    Article  Google Scholar 

  19. Denkert, C. et al. Tumour infiltrating lymphocytes and prognosis in different subtypes of breast cancer: a pooled analysis of 3771 patients treated with neoadjuvant therapy. Lancet Oncol. 19, 40–50 (2018).

    Article  Google Scholar 

  20. Luen, S. J. et al. Prognostic implications of residual disease tumor-infiltrating lymphocytes and residual cancer burden in triple-negative breast cancer patients after neoadjuvant chemotherapy. Ann. Oncol. 30, 236–242 (2019).

    Article  CAS  Google Scholar 

  21. Penault-Llorca, F. & Radosevic-Robin, N. Ki67 assessment in breast cancer: an update. Pathology 49, 166–171 (2017).

    Article  CAS  Google Scholar 

  22. Pistelli, M. et al. Prognostic factors in early-stage triple-negative breast cancer: lessons and limits from clinical practice. Anticancer Res. 33, 2737–2742 (2013).

    CAS  Google Scholar 

  23. Ahn, K. J., Park, J. & Choi, Y. Lymphovascular invasion as a negative prognostic factor for triple-negative breast cancer after surgery. Radiat. Oncol. J. 35, 332–339 (2017).

    Article  Google Scholar 

  24. Dimitriou, N., Arandjelović, O. & Caie, P. D. Deep learning for whole slide image analysis: an overview. Front. Med. 6, 264 (2019).

    Article  Google Scholar 

  25. Campanella, G. et al. Clinical-grade computational pathology using weakly supervised deep learning on whole slide images. Nat. Med. 25, 1301–1309 (2019).

    Article  CAS  Google Scholar 

  26. Echle, A. et al. Deep learning in cancer pathology: a new generation of clinical biomarkers. Br. J. Cancer 124, 686–696 (2021).

    Article  Google Scholar 

  27. Courtiol, P. et al. Deep learning-based classification of mesothelioma improves prediction of patient outcome. Nat. Med. 25, 1519–1525 (2019).

    Article  CAS  Google Scholar 

  28. Saillard, C. et al. Predicting survival after hepatocellular carcinoma resection using deep learning on histological slides. Hepatology 72, 2000–2013 (2020).

    Article  Google Scholar 

  29. Couture, H. D. et al. Image analysis with deep learning to predict breast cancer grade, ER status, histologic subtype, and intrinsic subtype. NPJ Breast Cancer 4, 30 (2018).

    Article  Google Scholar 

  30. Turkki, R. et al. Breast cancer outcome prediction with tumour tissue images and machine learning. Breast Cancer Res. Treat. 177, 41–52 (2019).

    Article  Google Scholar 

  31. Naik, N. et al. Deep learning-enabled breast cancer hormonal receptor status determination from base-level H&E stains. Nat. Commun. 11, 5727 (2020).

    Article  CAS  Google Scholar 

  32. Saltz, J. et al. Spatial organization and molecular correlation of tumor-infiltrating lymphocytes using deep learning on pathology images. Cell Rep. 23, 181–193 (2018).

    Article  CAS  Google Scholar 

  33. Binder, A. et al. Morphological and molecular breast cancer profiling through explainable machine learning. Nat. Mach. Intell. 3, 355–366 (2021).

    Article  Google Scholar 

  34. Cancer du sein triple négatif: la has autorise le trodelvy en accès précoce. Haute Autorité de Santé https://www.has-sante.fr/jcms/p_3284628/fr/cancer-du-sein-triple-negatif-la-has-autorise-le-trodelvy-en-acces-precoce (2021).

  35. Naylor, P., Boyd, J., Laé, M., Reyal, F. & Walter, T. Predicting residual cancer burden in a triple negative breast cancer cohort. In Proc. 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019), 933–937 (IEEE, 2019).

  36. McMahan, B., Moore, E., Ramage, D., Hampson, S. & Aguera y Arcas, B. Communication-efficient learning of deep networks from decentralized data. In Proc. 20th International Conference of Artificial Intelligence and Statistics (AISTATS) 54,1273–1282 (JMLR, 2017).

  37. Rieke, N. et al. The future of digital health with federated learning. NPJ Digital Med. 3, 119 (2020).

    Article  Google Scholar 

  38. Sheller, M. J., Reina, G. A., Edwards, B., Martin, J. & Bakas, S. Multi-institutional deep learning modeling without sharing patient data: a feasibility study on brain tumor segmentation. In International MICCAI Brain Lesion Workshop, 92–104 (Springer, 2018).

  39. Chang, K. et al. Distributed deep learning networks among institutions for medical imaging. J. Am. Med. Inform. Assoc. 25, 945–954 (2018).

    Article  Google Scholar 

  40. Lee, J. et al. Privacy preserving patient similarity learning in a federated environment: development and analysis. JMIR Med. Inform. 6, e7744 (2018).

    Article  Google Scholar 

  41. Lu, M. Y. et al. Federated learning for computational pathology on gigapixel whole slide images. Med. Image Anal. 76, 102298 (2022).

    Article  Google Scholar 

  42. Sheller, M. J. et al. Federated learning in medicine: facilitating multi-institutional collaborations without sharing patient data. Sci. Rep. 10, 1–12 (2020).

    Article  Google Scholar 

  43. Warnat-Herresthal, S. et al. Swarm learning for decentralized and confidential clinical machine learning. Nature 594, 12598 (2021).

    Article  Google Scholar 

  44. Sadilek, A. et al. Privacy-first health research with federated learning. NPJ Digital Med. 4, 132 (2021).

    Article  Google Scholar 

  45. Saldanha, O. L. et al. Swarm learning for decentralized artificial intelligence in cancer histopathology. Nat. Med. 28, 1232–1239 (2022).

    Article  CAS  Google Scholar 

  46. Zou, H. & Hastie, T. Regularization and variable selection via the elastic net. J. R. Stat. Soc. Series B Stat. Methodol. 67, 301–320 (2005).

    Article  Google Scholar 

  47. Ma, D. et al. Integrated molecular profiling of young and elderly patients with triple-negative breast cancer indicates different biological bases and clinical management strategies. Cancer 126, 3209–3218 (2020).

    Article  CAS  Google Scholar 

  48. Tang, Z. et al. Prognostic factors and models for elderly (≥70 years old) primary operable triple-negative breast cancer: analysis from the national cancer database. Front. Endocrinol. (Lausanne) 13, 856268 (2022).

    Article  Google Scholar 

  49. Dietterich, T. G. Ensemble methods in machine learning. In International Workshop on Multiple Classifier Systems, 1–15 (Springer, 2000).

  50. Karimireddy, S. P. et al. Stochastic controlled averaging for federated learning. In Proc. International Conference on Machine Learning, 5132–5143 (PMLR, 2020).

  51. Desa, D. E. et al. Second harmonic generation directionality is associated with neoadjuvant chemotherapy response in breast cancer core needle biopsies. J. Biomed. Opt. 24, 086503 (2019).

    Article  CAS  Google Scholar 

  52. Lehmann, B. D. et al. Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. J. Clin. Investig. 121, 2750–2767 (2011).

    Article  CAS  Google Scholar 

  53. Masuda, H. et al. Differential response to neoadjuvant chemotherapy among 7 triple-negative breast cancer molecular subtypes. Clin. Cancer Res. 19, 5533–5540 (2013).

    Article  CAS  Google Scholar 

  54. Schmauch, B. et al. A deep learning model to predict RNA-Seq expression of tumours from whole slide images. Nat. Commun. 11, 3877 (2020).

    Article  CAS  Google Scholar 

  55. Colin, I., Bellet, A., Salmon, J. & Clémençon, S. Extending gossip algorithms to distributed estimation of U-statistics. In Proc. 28th International Conference on Neural Information Processing Systems 1, 271–279 (MIT Press, 2015).

  56. Lassau, N. et al. Integrating deep learning CT-scan model, biological and clinical variables to predict severity of COVID-19 patients. Nat. Commun. 12, 634 (2021).

    Article  CAS  Google Scholar 

  57. Singletary, S. E. et al. Revision of the American Joint Committee on Cancer staging system for breast cancer. J. Clin. Oncol. 20, 3628–3636 (2002).

    Article  Google Scholar 

  58. Park, Y. H. et al. Clinical relevance of TNM staging system according to breast cancer subtypes. Ann. Oncol. 22, 1554–1560 (2011).

    Article  CAS  Google Scholar 

  59. Salgado, R. et al. The evaluation of tumor infiltrating lymphocytes (TILs) in breast cancer: recommendations by an international TILs working group 2014. Ann. Oncol. 26, 259–271 (2015).

    Article  CAS  Google Scholar 

  60. Levsky, J. M. & Singer, R. H. Fluorescence in situ hybridization: past, present and future. J. Cell Sci. 116, 2833–2838 (2003).

    Article  CAS  Google Scholar 

  61. Geršak, K., Gazic, B., Klevisar Ivancic, A., Ruzic Gorenjec, N. & Grasic Kuhar, C. Intra-and inter-observer variability in tumor infiltrating lymphocyte scoring in breast cancer core needle biopsy. J. Clin. Oncol. 39(15_suppl), e12626 (2021).

    Article  Google Scholar 

  62. Titford, M. The long history of hematoxylin. Biotech. Histochemistry 80, 73–78 (2005).

    Article  CAS  Google Scholar 

  63. Nietner, T., Jarutat, T. & Mertens, A. Systematic comparison of tissue fixation with alternative fixatives to conventional tissue fixation with buffered formalin in a xenograft-based model. Virchows Arch. 461, 259–269 (2012).

    Article  CAS  Google Scholar 

  64. Courtiol, P., Tramel, E. W., Sanselme, M. & Wainrib, G. Classification and disease localization in histopathology using only global labels: a weakly-supervised approach. Preprint at https://doi.org/10.48550/arXiv.1802.02212 (2018).

  65. Norgeot, B. et al. Minimum information about clinical artificial intelligence modeling: the MI-CLAIM checklist. Nat. Med. 26, 1320–1324 (2020).

    Article  CAS  Google Scholar 

  66. Ronneberger, O., Fischer, P. & Brox, T. U-net: convolutional networks for biomedical image segmentation. In Proc.International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), 234–241 (Springer, 2015).

  67. He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In Proc. IEEE Conference on Computer Vision and Pattern Recognition, 770–778 (IEEE, 2016).

  68. Vahadane, A. et al. Structure-preserving color normalization and sparse stain separation for histological images. IEEE Trans. Med. Imaging 35, 1962–1971 (2016).

    Article  Google Scholar 

  69. Abdi, H. & Williams, L. J. Principal component analysis. Wiley Interdiscip. Rev. Comput. Stat. 2, 433–459 (2010).

    Article  Google Scholar 

  70. Maurice Fréchet, M. Sur quelques points du calcul fonctionnel. Rendiconti Circolo Matematico di Palermo 22, 1–72 (1906).

    Article  Google Scholar 

  71. Dowson, D. C. & Landau, B. V. The Fréchet distance between multivariate normal distributions. J. Multivar. Anal. 12, 450–455 (1982).

    Article  Google Scholar 

  72. Heusel, Martin, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, & Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems 30 (Curran Associates, Inc., 2017).

  73. Macenko, M. et al. A method for normalizing histology slides for quantitative analysis. In 2009 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, 1107–1110 (IEEE, 2009).

  74. Tarek Shaban, M., Baur, C., Nÿavab, N. & Albarqouni, S. Staingan: stain style transfer for digital histological images. In Proc. 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019), 953–956 (IEEE, 2019).

  75. Tomczak, A. et al. Multi-task multi-domain learning for digital staining and classification of leukocytes. IEEE Trans. Med. Imaging 40, 2897–2910 (2020).

    Article  Google Scholar 

  76. Andreux, M., Manoel, A., Menuet, R., Saillard, C. & Simpson, C. Federated survival analysis with discrete-time Cox models. Preprint at https://doi.org/10.48550/arXiv.2006.0899 (2020).

  77. Andreux, M., Ogier du Terrail, J., Beguier, C. & Tramel, E. W. Siloed federated learning for multi-centric histopathology datasets. In Domain Adaptation and Representation Transfer, and Distributed and Collaborative Learning, 129–139 (Springer, 2020).

    Chapter  Google Scholar 

  78. Ioffe, S. & Szegedy, C. Batch normalization: accelerating deep network training by reducing internal covariate shift. In Proc.International Conference on Machine Learning, 484–456 (PMLR, 2015).

  79. Tomczak, K., Czerwińska, P. & Wiznerowicz, M. The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge. Contemp. Oncol. 19, 68–77 (Termedia, 2015).

  80. Chen, X., Fan, H., Girshick, R. & He, K. Improved baselines with momentum contrastive learning. Preprint at https://doi.org/10.48550/arXiv.2003.04297 (2020).

  81. Maron, O. & Lozano-Pérez, T. A framework for multiple-instance learning. In Advances in Neural Information Processing Systems, 570–576 (MIT Press, 1998).

  82. Durand, T, Thome, N & Cord, M. WELDON: weakly supervised learning of deep convolutional neural networks. In Pattern Recognition CVPR 2016, 4743–4752 (IEEE, 2016).

  83. Ilse, M., Tomczak, J. & Welling, M. Attention-based deep multiple instance learning. In Proc. International Conference on Machine Learning, 2127–2136 (PMLR, 2018).

  84. Kingma, D. P. & Ba, J. Adam: a method for stochastic optimization. In Proc. 3rd International Conference on Learning Representations, (ICLR, 2015).

  85. Galtier M. N. & Marini C. Substra: a framework for privacy-preserving, traceable and collaborative machine learning. Preprint at https://doi.org/10.48550/arXiv.1910.11567 (2019).

  86. Androulaki, E. et al. Hyperledger fabric: a distributed operating system for permissioned blockchains. In Proc. Thirteenth EuroSys Conference 1–15 (ACM SIGOPS, 2018).

  87. McMahan, B., Moore, E., Ramage, D., Hampson, S. & Aguera y Arcas, B. Communication-efficient learning of deep networks from decentralized data. In Proc. 20th International Conference of Artificial Intelligence and Statistics (AISTATS) 54, 1273–1282 (JMLR, 2017).

  88. Steinberg, D. & Colla, P. Cart: classification and regression trees. Top. Ten Algorithms Data Min. 9, 179 (2009).

    Article  Google Scholar 

  89. Chen, T. & Guestrin, C. Xgboost: a scalable tree boosting system. In Proc. 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 785–794 (ACM, 2016).

    Chapter  Google Scholar 

  90. Lahat, D., Adali, T. & Jutten, C. Multimodal data fusion: an overview of methods, challenges, and prospects. Proc. IEEE 103, 1449–1477 (2015).

    Article  Google Scholar 

  91. Baltrušaitis, T., Ahuja, C. & Morency, L.-P. Multimodal machine learning: a survey and taxonomy. IEEE Trans. Pattern Anal. Mach. Intell. 41, 423–443 (2018).

    Article  Google Scholar 

  92. Bergeron, A. et al. Triple negative breast lobular carcinoma: a luminal androgen receptor carcinoma with specific esrra mutations. Mod. Pathol. 34, 1282–1296 (2021).

    Article  CAS  Google Scholar 

  93. Aickin, M. & Gensler, H. Adjusting for multiple testing when reporting research results: the Bonferroni vs Holm methods. Am. J. Public Health 86, 726–728 (1996).

    Article  CAS  Google Scholar 

  94. Seabold, S. & Perktold, J. Statsmodels: econometric and statistical modeling with python. In Proc. 9th Python in Science Conference 57, 10–25080 (Scipy, 2010).

  95. Wilcoxon, F. Individual comparisons by ranking methods. In Breakthroughs in Statistics 202, 196–983. (Springer, 1992).

  96. Virtanen, P. et al. Scipy 1.0: fundamental algorithms for scientific computing in python. Nat. Methods 17, 261–272 (2020).

    Article  CAS  Google Scholar 

  97. Viechtbauer, W. Conducting meta-analyses in R with the metafor package. J. Stat. Softw. 36, 1–48 (2010).

    Article  Google Scholar 

  98. Fisher, R. A. On the interpretation of χ2 from contingency tables, and the calculation of p. J. R. Stat. Soc. 85, 87–94 (1922).

    Article  Google Scholar 

Download references

Acknowledgements

This project is supported by Bpifrance as part of the Healthchain project, which resulted from the Digital Investments Program for the major challenges of the future Request For Proposal (RFP). As part of the Healthchain project, a consortium coordinated by Owkin, Inc. (a private company) has been established, including the Substra association, Apricity (a private company), the Assistance Publique des Hôpitaux de Paris, the University Hospital Center of Nantes, the Léon Bérard Center, the French National Center for Scientific Research, the Ecole Polytechnique, the Institut Curie and the University of Paris Descartes. We thank all the Owkin, Inc. teams involved in the project, most notably the Research & Development (R&D) data science team led by Eric Durand, the Information Technology (IT) team led by Jocelyn Dachary, the legal team, the partnerships team and Owkin Inc. founders Thomas Clozel and Gilles Wainrib for their support team.

Author information

Authors and Affiliations

Authors

Contributions

G.B. and P.E.H. conceived the idea of the paper with the help of A. Livartowski and reviewed the medical content of the paper. M.A., C.B., C. Maussion, J.O.d.T., B.S. and E.W.T. implemented the analysis and wrote the article, while G.W. and M.Z. provided supervision. M.A., J.O.d.T. and E.W.T implemented the federated learning operations using the substra backend. C.J., A. Leopold and M.M. were in charge of data curation and harmonization in both training centers. G.B. and J. Gervasoni also annotated the pathology images and provided the interpretations of the histological patterns found. I.D., T.D., C.G., J. Guerin, C. Marini, K.M. and A.S. were in charge of the federated network between the hospitals (Information Technology (IT) infrastructure and software backend using substra). E.B., T.D., M.G., A. Livartowski and A.-L.M. managed the collaboration. D.D. was in charge of data preparation for Institut Gustave Roussy. J.M. managed the partnership relations with external validation institutions. C.F. and M.L.-T. assembled cohorts and medical materials for external validations. R.D. provided the normalization code used for IUCT Oncopole. F.B. performed the meta-analysis of the results.

Corresponding author

Correspondence to Jean Ogier du Terrail.

Ethics declarations

Competing interests

The authors declare the existence of a financial competing interest. As described in the Acknowledgements section, this work received the financial support of the French Banque Publique d’Investissement (BPI) through the HealthChain project (https://www.substra.ai/en/healthchain-project). Some authors are or were employed by Owkin, Inc. during their time on the project (J.O.d.T., C.B., M.A., C. Maussion, B.S., E.B., M.Z., G.W., K.M., C.G., I.D., A.- L.M., C. Marini, M.G., J.M., R.D., A.S. and F.B.). In addition, G.W. is the cofounder of Owkin, Inc. P.E.H. reports grants, personal fees and nonfinancial support from Pfizer; grants and nonfinancial support from Novartis; grants and nonfinancial support from Roche; grants, personal fees and nonfinancial support from AstraZeneca; personal fees and nonfinancial support from Mylan; grants, personal fees, and nonfinancial support from Pierre Fabre; personal fees and nonfinancial support from Amgen; and personal fees and nonfinancial support from Seagen outside the submitted work.

Peer review

Peer review information

Nature Medicine thanks Roberto Salgado, Matthew Hanna and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Michael Basson and Javier Carmona, in collaboration with the Nature Medicine team.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary Figs. 1–7, Tables 1–7 and Note 1.

Reporting Summary

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Ogier du Terrail, J., Leopold, A., Joly, C. et al. Federated learning for predicting histological response to neoadjuvant chemotherapy in triple-negative breast cancer. Nat Med 29, 135–146 (2023). https://doi.org/10.1038/s41591-022-02155-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41591-022-02155-w

Search

Quick links

Nature Briefing: Cancer

Sign up for the Nature Briefing: Cancer newsletter — what matters in cancer research, free to your inbox weekly.

Get what matters in cancer research, free to your inbox weekly. Sign up for Nature Briefing: Cancer