Prognostic value of various immune cells and Immunoscore in triple-negative breast cancer

Abstract

Background: This study aimed to evaluate the expression status and prognostic role of various immunoregulatory cells and test in triple-negative breast cancer (TNBC).

Methods: The expression of five markers (CD3/CD4/CD8/CD19/CD163) of tumor immune cells was evaluated retrospectively in tumor sections from 68 consecutive cases of TNBC by immunohistochemistry. Computational image analysis was used to quantify the density and distribution of each immune marker within the tumor region, tumor invasive margin, and expression hotspots. Immunoscores were calculated using an automated approach. Other clinical characteristics were also analyzed.

Results: For all patients, Kaplan-Meier survival analysis showed that high CD3+ signals in the tumor region (disease-free survival (DFS), P=0.0014; overall survival (OS), P=0.0031) and total region (DFS, P=0.0014; OS, P=0.0031) were significantly associated with better survival. High CD4+ levels in the tumor region and total regions were significantly associated with better survival (P<0.05). For Hotspot analysis, CD3+ was associated with significantly better survival for all Top1, Top2, and Top3 densities (DFS and OS, P<0.05). High CD4+ levels were significantly associated with better prognosis for Top1 and Top3 densities (DFS and OS, P<0.05). For stage IIB and IIIC patients, CD3+ in the tumor region and all Top hotspots was found to be significantly correlated with survival (DFS and OS, P<0.05). CD4+ cells were significantly associated with survival in the tumor region, total region, and Top3 density (DFS, P=0.0213; OS, P=0.0728). CD8+ cells were significantly associated with survival in the invasive margin, Top2 density, and Top3 density. Spatial parameter analysis showed that high colocalization of tumor cells and immune cells (CD3+, CD4+, or CD8+) was significantly associated with patient survival.

Conclusion: Computational image analysis is a reliable tool for evaluating the density and distribution of immune regulatory cells and for calculating the Immunoscore in TNBC. The Immunoscore retains its prognostic significance in TNBC later than IIB stage breast cancer. Future studies are required to confirm its potential to predict tumor responses to chemotherapy and immune therapy.

Keywords: Immunoscore; image registration; immune microenvironment; immunohistochemical staining; triple-negative breast cancer.

Figure 1

Schematic Overview of IHC Series Imaging, Registration and Multiplex Visualization. (A) Whole-Slide Digital Images representing serial sections enable assessment of immune microenvironment biomarkers. The images were down-sampled to a 500 pixels × 500 pixels size and then registered to the target images to generate deformation fields. The original size digital images were finally registered with up-sampled deformation fields, and thus serial sections identifying different biomarkers were spatially aligned to study the distributions of multiple immune cells. (B) Integrated visualization of serial sections by pseudo-coloring. Biomarkers and colors are shown on the right. The gray arrows indicated analytical steps that moves forward.

Figure 2

The flowchart of the experiment. The brown part present image registration and annotation process; The blue part present the analysis process of every IHC markers; The green part present process of colocalization of immune marker of tumor cells.

Figure 3

The overall survival curve. (A) CD3 expression in tumor region in all patients; (B) CD8-CD163 colocalization level in all patients; (C) Hotspot 1 CD3 expression in IIB and III stage patients; (D) CK-CD8 colocalization level in IIB and III stage patients.