An Open Source Automated Tumor Infiltrating Lymphocytes Algorithm for Prognosis in Primary Small Bowel Adenocarcinoma Using Routine Hematoxylin and Eosin Stained Sections
Minying Deng, Xin Liu, Yuanlong Luo, Ayizimugu Abuduwaili, Qi Song, Rongkui Luo, Wen Huang, Lei Xu, Xiaolei Zhang, Lin Sun, Jieakesu Su, Chen Xu, Yingyong HouABSTRACT
Objective
While tumor‐infiltrating lymphocytes (TILs) are prognostic in various cancers, their role in small bowel adenocarcinoma (SBA) is unexplored. This study evaluates the prognostic significance of digitally quantified TILs in SBA.
Methods
Using digital pathology (QuPath) on hematoxylin and eosin (H&E)‐stained slides from 62 SBA cases, we quantified tumor cells, TILs (lymphocytes & plasma cells), and stromal components. Derived variables (eTILs, etTILs, esTILs, eaTILs, sTILs, iTILs, easTILs, and tumor‐stroma ratio [TSR]) were correlated with clinicopathological features and survival. To account for multiple comparisons across all statistical tests, false discovery rate (FDR) correction was applied.
Results
Several TIL‐related parameters showed suggestive associations with clinicopathological features ( p < 0.05), though none remained significant after FDR correction. In univariate survival analysis, M stage and TNM stage were strongly associated with disease‐free survival (DFS) and remained highly significant after FDR correction (both q < 0.001). Regarding TIL‐related parameters, higher eTILs ( p = 0.028) and iTILs ( p = 0.032) showed suggestive associations with prolonged overall survival (OS) in univariate analysis, and higher iTILs also showed a suggestive association with DFS ( p = 0.045); however, none of these associations remained significant after FDR correction. In multivariate analysis, M stage and TNM stage were confirmed as independent prognostic factors for DFS, while eTILs ( p = 0.099) and tumor deposit ( p = 0.055) showed trends toward independent prognostic significance without reaching statistical significance.
Conclusion
This study, for the first time, objectively quantifies TILs in SBA, a rare tumor, using digital pathology and explores the potential prognostic significance of TILs in SBA. TIL‐related parameters demonstrate suggestive trends that warrant validation in larger, independent cohorts. These findings provide a foundation for future studies to clarify the role of TILs in SBA.