@article{27756e515fdd4c02b3d8a820441dd66d,
title = "De novo prediction of cancer-associated T cell receptors for noninvasive cancer detection",
abstract = "The adaptive immune system recognizes tumor antigens at an early stage to eradicate cancer cells. This process is accompanied by systemic proliferation of the tumor antigen–specific T lymphocytes. While detection of asymptomatic early-stage cancers is challenging due to small tumor size and limited somatic alterations, tracking peripheral T cell repertoire changes may provide an attractive solution to cancer diagnosis. Here, we developed a deep learning method called DeepCAT to enable de novo prediction of cancer-associated T cell receptors (TCRs). We validated DeepCAT using cancer-specific or non-cancer TCRs obtained from multiple major histocompatibility complex I (MHC-I) multimer-sorting experiments and demonstrated its prediction power for TCRs specific to cancer antigens. We blindly applied DeepCAT to distinguish over 250 patients with cancer from over 600 healthy individuals using blood TCR sequences and observed high prediction accuracy, with area under the curve (AUC) ≥ 0.95 for multiple early-stage cancers. This work sets the stage for using the peripheral blood TCR repertoire for noninvasive cancer detection.",
author = "Daria Beshnova and Jianfeng Ye and Oreoluwa Onabolu and Benjamin Moon and Wenxin Zheng and Fu, {Yang Xin} and James Brugarolas and Jayanthi Lea and Bo Li",
note = "Funding Information: Acknowledgments: We thank J. Han (HudsonAlpha Institute and iRepertoire) for sharing the PBMC TCR-seq samples for healthy donors and the RCC samples profiled using the iRepertoire platform; D. Pardoll, K. Smith, and H. Ji (Johns Hopkins) for sharing the early-stage lung cancer samples profiled with Adaptive Biotechnologies; and T. Aguilera and the Pancreatic Cancer Prevention Program at UT Southwestern Medical Center for providing the PBMC samples from patients with early-stage pancreatic cancer or benign cyst. We also thank C. Arteaga and J. Minna for useful discussions during manuscript preparation and S. Rajaram for useful suggestions to help the methodology development. Funding: This work is supported by CPRIT RR170079 (B.L.) and NCI SPORE 1P50CA196516 (J.B.). Author contributions: B.L. conceived the project and developed the DeepCAT method. B.L. and D.B. performed data analysis and wrote the manuscript. J.L., O.O., and J.B. contributed clinical samples. J.Y. performed gDNA extraction and immune repertoire sequencing. B.M., W.Z., and Y.-X.F. helped write the manuscript. Competing interests: J.L. has an advisory affiliation with Clovis Oncology. B.L. is the inventor on patent application (133892-020601/ PCT) submitted by UT Southwestern Medical Center that covers “Computerized system and method for antigen-independent de novo prediction of cancer-associated TCR repertoire.” All other authors declare that they have no competing interests. Data and materials availability: Source code of DeepCAT and AdaBoost and the training and testing datasets are available at https://github.com/s175573/DeepCAT. All the raw TCR-seq data generated in this study and the post-processed datasets associated with this study are available at Zenodo DOI: 10.5281/zenodo.3894880. Publisher Copyright: Copyright {\textcopyright} 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works",
year = "2020",
doi = "10.1126/SCITRANSLMED.AAZ3738",
language = "English (US)",
volume = "12",
journal = "Science translational medicine",
issn = "1946-6234",
publisher = "American Association for the Advancement of Science",
number = "557",
}