Pitfalls in machine learning-based assessment of tumor-infiltrating lymphocytes in breast cancer: a report of the international immuno-oncology biomarker working group

Jeppe Thagaard, Glenn Broeckx, David B. Page, Chowdhury Arif Jahangir, Sara Verbandt, Zuzana Kos, Rajarsi Gupta, Reena Khiroya, Khalid Abduljabbar, Gabriela Acosta Haab, Balazs Acs, Guray Akturk, Jonas S. Almeida, Isabel Alvarado-Cabrero, Mohamed Amgad, Farid Azmoudeh-Ardalan, Sunil Badve, Nurkhairul Bariyah Baharun, Eva Balslev, Enrique R. BellolioVydehi Bheemaraju, Kim R.M. Blenman, Luciana Botinelly Mendonça Fujimoto, Najat Bouchmaa, Octavio Burgues, Alexandros Chardas, Maggie Chon U Cheang, Francesco Ciompi, Lee A.D. Cooper, An Coosemans, Germán Corredor, Anders B. Dahl, Flavio Luis Dantas Portela, Frederik Deman, Sandra Demaria, Johan Doré Hansen, Sarah N. Dudgeon, Thomas Ebstrup, Mahmoud Elghazawy, Claudio Fernandez-Martín, Stephen B. Fox, William M. Gallagher, Jennifer M. Giltnane, Sacha Gnjatic, Paula I. Gonzalez-Ericsson, Anita Grigoriadis, Niels Halama, Matthew G. Hanna, Aparna Harbhajanka, Steven N. Hart, Johan Hartman, Søren Hauberg, Stephen Hewitt, Akira I. Hida, Hugo M. Horlings, Zaheed Husain, Evangelos Hytopoulos, Sheeba Irshad, Emiel A.M. Janssen, Mohamed Kahila, Tatsuki R. Kataoka, Kosuke Kawaguchi, Durga Kharidehal, Andrey I. Khramtsov, Umay Kiraz, Pawan Kirtani, Liudmila L. Kodach, Konstanty Korski, Anikó Kovács, Anne Vibeke Laenkholm, Corinna Lang-Schwarz, Denis Larsimont, Jochen K. Lennerz, Marvin Lerousseau, Xiaoxian Li, Amy Ly, Anant Madabhushi, Sai K. Maley, Vidya Manur Narasimhamurthy, Douglas K. Marks, Elizabeth S. McDonald, Ravi Mehrotra, Stefan Michiels, Fayyaz ul Amir Afsar Minhas, Shachi Mittal, David A. Moore, Shamim Mushtaq, Hussain Nighat, Thomas Papathomas, Frederique Penault-Llorca, Rashindrie D. Perera, Christopher J. Pinard, Juan Carlos Pinto-Cardenas, Giancarlo Pruneri, Lajos Pusztai, Arman Rahman, Nasir Mahmood Rajpoot, Bernardo Leon Rapoport, Tilman T. Rau, Jorge S. Reis-Filho, Joana M. Ribeiro, David Rimm, Anne Roslind, Anne Vincent-Salomon, Manuel Salto-Tellez, Joel Saltz, Shahin Sayed, Ely Scott, Kalliopi P. Siziopikou, Christos Sotiriou, Albrecht Stenzinger, Maher A. Sughayer, Daniel Sur, Susan Fineberg, Fraser Symmans, Sunao Tanaka, Timothy Taxter, Sabine Tejpar, Jonas Teuwen, E. Aubrey Thompson, Trine Tramm, William T. Tran, Jeroen van der Laak, Paul J. van Diest, Gregory E. Verghese, Giuseppe Viale, Michael Vieth, Noorul Wahab, Thomas Walter, Yannick Waumans, Hannah Y. Wen, Wentao Yang, Yinyin Yuan, Reena Md Zin, Sylvia Adams, John Bartlett, Sibylle Loibl, Carsten Denkert, Peter Savas, Sherene Loi, Roberto Salgado, Elisabeth Specht Stovgaard

Research output: Contribution to journalReview articlepeer-review

13 Citations (Scopus)

Abstract

The clinical significance of the tumor-immune interaction in breast cancer is now established, and tumor-infiltrating lymphocytes (TILs) have emerged as predictive and prognostic biomarkers for patients with triple-negative (estrogen receptor, progesterone receptor, and HER2-negative) breast cancer and HER2-positive breast cancer. How computational assessments of TILs might complement manual TIL assessment in trial and daily practices is currently debated. Recent efforts to use machine learning (ML) to automatically evaluate TILs have shown promising results. We review state-of-the-art approaches and identify pitfalls and challenges of automated TIL evaluation by studying the root cause of ML discordances in comparison to manual TIL quantification. We categorize our findings into four main topics: (1) technical slide issues, (2) ML and image analysis aspects, (3) data challenges, and (4) validation issues. The main reason for discordant assessments is the inclusion of false-positive areas or cells identified by performance on certain tissue patterns or design choices in the computational implementation. To aid the adoption of ML for TIL assessment, we provide an in-depth discussion of ML and image analysis, including validation issues that need to be considered before reliable computational reporting of TILs can be incorporated into the trial and routine clinical management of patients with triple-negative breast cancer.

Original languageEnglish
Pages (from-to)498-513
Number of pages16
JournalJournal of Pathology
Volume260
Issue number5
DOIs
Publication statusPublished - Aug 2023
Externally publishedYes

Keywords

  • deep learning
  • digital pathology
  • guidelines
  • image analysis
  • machine learning
  • pitfalls
  • prognostic biomarker
  • triple-negative breast cancer
  • tumor-infiltrating lymphocytes

Fingerprint

Dive into the research topics of 'Pitfalls in machine learning-based assessment of tumor-infiltrating lymphocytes in breast cancer: a report of the international immuno-oncology biomarker working group'. Together they form a unique fingerprint.

Cite this