Pitfalls in machine learning‐based assessment of tumor‐infiltrating lymphocytes in breast cancer: A report of the International Immuno‐Oncology Biomarker Working Group on Breast Cancer

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. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Application of AI in detection of breast cancer with laboratory results monitoring

“…One common method used in machine learning-based assessment of TILs is image segmentation, where AI algorithms delineate regions of interest corresponding to TILs in tissue slides. By distinguishing TILs from other cell types and structures, machine learning models can quantify the percentage of TILs present in the tumor, as well as assess their spatial organization and relationship to other tumor features [49]. Artificial Intelligence-based mitosis scoring in breast cancer is a cutting-edge approach that utilizes advanced algorithms to automate the assessment of mitotic activity in tumor samples.…”

Early Breast Cancer Risk Assessment: Integrating Histopathology with Artificial Intelligence

“…Additionally, these challenges are further complicated by technical factors, such as the lack of standardized parameters for data acquisition. [ 56 ]. AI-assisted TIL quantification remains a promising tool when handled by an experienced pathologist; however, forecasting clinical outcomes based on pre-treatment histopathologic images remains a challenging endeavor, hindered by the incomplete comprehension of the tumor immune microenvironment [ 52 , 56 , 57 ].…”

Exploring Biomarkers in Breast Cancer: Hallmarks of Diagnosis, Treatment, and Follow-Up in Clinical Practice

“…Li et al's meta-analysis of 36 studies indicated that CA15-3 levels differ among breast cancer subtypes, suggesting its utility in differentiating TNBC from other forms [133]. Fu and Li emphasized the importance of combining CA15-3 with other tumor markers for more accurate clinical assessments [134], while Zhu et al highlighted its role in monitoring therapy outcomes and disease progression in metastatic breast cancer [135]. Additionally, Wang et al identified CA15-3 as a key biomarker in nipple discharge, aiding in diagnosis and prognosis [136], and Oliveira et al's development of a microfluidic device for CA15-3 detection underscores its practical application in clinical settings [137].…”