QuTILs: Open-Source Image-Based Infiltrating Immune Cell Detection for Research Application

Area of Science:

• Computational pathology
• Digital pathology
• Breast cancer research

Background:

• Stromal tumor infiltrating lymphocytes (sTILs) are crucial prognostic and predictive biomarkers in triple-negative breast cancer (TNBC).
• Accurate quantification of sTILs from hematoxylin and eosin (H&E) slides is essential for clinical research.
• Digitalization of pathology slides enables computational analysis for sTIL enumeration.

Purpose of the Study:

• To introduce QuTILs, an open-source, research-based computational approach for enumerating sTILs.
• To validate the QuTILs method in large TNBC clinical trials.
• To assess the association of QuTILs-derived sTIL percentage with patient outcomes.

Main Methods:

• Development of a multilayer perceptron-based framework using QuPath software for sTIL identification.
• Training and execution of the QuTILs algorithm on open-source H&E images from TNBC patients.
• Application of QuTILs to H&E slides from two Phase III TNBC clinical trials (CALGB 40502 and 40603).

Main Results:

• QuTILs demonstrated a significant univariate association with improved outcomes in the CALGB 40502 trial.
• Higher sTIL percentages quantified by QuTILs correlated with reduced hazard (HR: 0.75, 95% CI: 0.63-0.91).
• The prognostic significance of QuTILs was confirmed in multivariable models and validated in the CALGB 40603 trial.

Conclusions:

• QuTILs offers a computationally efficient and open-source workflow for sTIL identification in digital H&E images.
• This method facilitates reproducible sTIL quantification for research applications in TNBC.
• QuTILs can aid in understanding the role of sTILs in treatment response and survival in TNBC.