Clinical relevance of the characterisation and quantification of CD8+ T-cell infiltration in high-grade serous ovarian carcinoma using machine learning-based image analysis of dual immunostaining.

INTRODUCTION

High-grade serous carcinoma (HGSC) is the most common and aggressive subtype of ovarian cancer. Although tumour stage and complete cytoreduction are key prognostic factors, the tumour immune microenvironment also influences prognosis, with CD8+ T lymphocytes representing the main effector component.

OBJECTIVE

To evaluate the clinical relevance of quantifying intraepithelial (ieTILs) and stromal (sTILs) CD8+ tumour-infiltrating lymphocytes using digital image analysis.

MATERIAL AND

METHODS

A retrospective study was conducted on 74 patients with stage III-IV HGSC treated with cytoreductive surgery and chemotherapy. Tissue microarrays (TMAs) were constructed, and dual immunostaining for cytokeratin and CD8 was performed. Sections were analysed using supervised machine-learning algorithms to quantify ieTILs and sTILs. These parameters correlated with age, cytoreduction status, platinum-free interval (PFI), overall survival (OS), neoadjuvant therapy, and BRCA mutation status.

RESULTS

Patients with higher intraepithelial CD8+ T-cell infiltration showed longer overall survival and a trend toward a longer platinum-free interval, whereas stromal density was not associated with prognosis. In multivariate analyses, ieTILs remained as an independent prognostic factor for both OS and PFI.

CONCLUSIONS

Digital quantification of intraepithelial CD8+ TILs using dual immunostaining and machine-learning analysis provides an accurate assessment of the lymphocytic infiltrate and represents an independent prognostic factor in HGSC. The lack of prognostic value of stromal CD8+ density highlights the importance of assessing the spatial localisation of effector cells in tumour-immune interactions.