TIAger: Tumor-Infiltrating Lymphocyte Scoring in Breast Cancer for the TiGER Challenge

Adam Shephard,Mostafa Jahanifar,Ruoyu Wang,Muhammad Dawood,Simon Graham,Kastytis Sidlauskas,Syed Ali Khurram,Nasir Rajpoot,Shan E Ahmed Raza

DOI: https://doi.org/10.48550/arXiv.2206.11943

2022-06-24

Abstract:The quantification of tumor-infiltrating lymphocytes (TILs) has been shown to be an independent predictor for prognosis of breast cancer patients. Typically, pathologists give an estimate of the proportion of the stromal region that contains TILs to obtain a TILs score. The Tumor InfiltratinG lymphocytes in breast cancER (TiGER) challenge, aims to assess the prognostic significance of computer-generated TILs scores for predicting survival as part of a Cox proportional hazards model. For this challenge, as the TIAger team, we have developed an algorithm to first segment tumor vs. stroma, before localising the tumor bulk region for TILs detection. Finally, we use these outputs to generate a TILs score for each case. On preliminary testing, our approach achieved a tumor-stroma weighted Dice score of 0.791 and a FROC score of 0.572 for lymphocytic detection. For predicting survival, our model achieved a C-index of 0.719. These results achieved first place across the preliminary testing leaderboards of the TiGER challenge.

Image and Video Processing,Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to automatically quantify tumor - infiltrating lymphocytes (TILs) in breast cancer tissue - section images. Specifically, the authors developed an algorithm aimed at first segmenting the tumor from the stroma, then locating the tumor area to detect TILs, and finally generating a TILs score for each case. This work is part of the Tumor Infiltrating Lymphocytes in breast cancER (TiGER) challenge, whose goal is to evaluate the prognostic value of computer - generated TILs scores in predicting survival rates as part of the Cox proportional - hazards model.

The paper mentions that the number of TILs has been proven to be an independent predictor of prognosis in breast cancer patients. Normally, pathologists estimate the proportion of the tumor stroma area containing TILs to obtain a TILs score. However, this method relies on manual assessment and has problems of subjectivity and being time - consuming. Therefore, developing a method that can automatically and objectively assess the number of TILs is of great significance for improving the prognostic assessment of breast cancer patients.

To achieve this goal, the authors proposed a deep - learning - based pipeline, using the Efficient - UNet architecture for image segmentation and detection tasks, and optimizing the model performance through a two - stage training process. The experimental results show that this method achieved good results in the preliminary tests, including excellent performance in terms of the tumor - stroma - weighted Dice score and the FROC score for lymphocyte detection, and reaching a C - index of 0.719 in predicting survival rates, which enabled them to obtain the first - place result on the preliminary leaderboard of the TiGER challenge.

Overall, the main contribution of this study lies in providing an efficient and accurate automated TILs - scoring solution, which not only helps to improve the efficiency of clinical diagnosis and treatment decision - making in breast cancer, but also provides a valuable reference for future related research.