Digital Pathology Enabled Study of CD8 T Cell Prevalence in NSCLC Using CD39/TCF7/CD8 Multiplex IHC Assay
Biswajeeta Saha 1*, Abhishek Shukla 2, Manjunath Nayak 1, Kavana Nadahalli 1,
Chetana Basavaraj 3, Scott Ely 4
*Correspondence to: Biswajeeta Saha. Biocon Bristol-Myers Squibb Research & Development Centre (BBRC), Syngene International Ltd., Bangalore – 560 099, Karnataka, India.
Copyright
© 2026 Biswajeeta Saha. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Received: 07 September 2026
Published: 01 October 2026
DOI: https://doi.org/10.5281/zenodo.23080276
Introduction
CD8+ T cells, also known as cytotoxic T cells, play a critical role in adaptive immunity against pathogens and cancer. This cytotoxic property of CD8+ T cells against tumour cells is utilized in immunotherapy, specifically immune checkpoint inhibitor (ICI) therapy. Assessment of CD8+ T cells[1,2] and ICI target markers, such as PD1[1] and PD-L1, [1,2] are found to be useful in predicting the probable responders and thus stratifying them. Although this stratifying strategy showed some success, there is still a high proportion of patients who do not respond to these ICI therapies. This may be because of many reasons, such as difference in the tumour mutational burden, immune contexture, genetic and epigenetic alterations, and microbiota. [3,4]
With respect to tumour immune contexture, the different physiological states exhibited by CD8+ T cells influence response to ICI therapy.[5,6] In patients with CD8+ T cells that express TCF7, a transcription factor known to be an important player in differentiation and also considered an early exhaustion marker, better ICI therapy response was observed.[6] Whereas in patients with CD8+ cells with gene profiles with CD39, a terminal exhaustion marker expression, no response was observed.[6] Furthermore, TCF7 expressing CD8+ T cells are shown to exhibit proliferation or self-renewal, effector functions, and further differentiation potential.[7] In addition, in preclinical models, CD8+ T cells expressing TCF7 were shown to be required for anti-tumour functions, and in melanoma tumours, these TCF7+ CD8+ T cells were found to be preferentially localized in the stroma.[7] On the contrary, presence of CD8+ T cells that express CD39 are shown to have positive implications in ICI therapy.[8,9] Furthermore, co-staining with CD39 is suggested to be a useful strategy for predicting PD-1 therapy response.[8] In addition, CD39 is also considered as another target for checkpoint inhibition therapies in combination with current ICIs.[10,11,12]
Thus, assessment of expression of these markers is essential for bettering the efficiency of ICI therapy. Currently, tumour mutation burden assessment, gene expression profiling, monoplex IHC, and multiplex IHC/IF are used for this purpose. Among these techniques, mIHC has been found to be the most effective with the diagnostic accuracy better than all other modalities or equal to multimodal efforts.[13,14] The advent of digital pathology has added additional value to the advantages of mIHC. Using digital pathology, objectivity could be introduced to the traditional manual image analysis and the turnover time can be reduced.
Numerous studies have reported tumour infiltrating CD8+ T cells to exhibit different physiological states,[5-11] and few studies have indicated that the diverse states of CD8+ T cells are clonal products, which appear to be derived from stem-like progenitor early exhausted CD8+ T cells.[15]
These studies have used techniques where gene and/or marker expression and cell densities are assessed using bulk modalities that does not provide spatial information. To our knowledge, there are no studies correlating the geographical localization of these distinct population of cells. In this study, we hypothesized that subsets of CD8 T cells, having distinct functional states, in the TME, localize to distinct geographical areas. We assessed two subsets of CD8+ T cells in TME namely TCF7+ and CD39+ cells. Using a unique quantification strategy of multi–field of view in digital image analysis, we quantified CD8 T cell subsets.
Materials and Methods
The triplex IHC assay for CD8, CD39, and TCF7 was validated on non-small cell lung cancer (NSCLC) samples. All assays were run on a Leica BOND Rx IHC autostainer.
A novel multi–field of view (FOV) approach was devised and used in this study to assess the prevalence of CD8 T-cell subsets. To check the feasibility of this approach, one selected WSI was viewed, and all the cells were annotated by trained pathologists using HALO Link. The same WSI was divided into N grids of 4 mm2, ensuring that the heterogeneity across the slide is captured. This size of the grids was selected because it is approximately equal to the area viewed under a microscope using a 20x objective lens. Assuming that the cell distribution in such small grids is homogeneous, two FOVs, each of 0.25 mm2 size, in each grid were subjectively selected such that the FOVs were representative of the grid. Cells were annotated in those FOVs using pin annotation on HALO LINK to provide counts of each cell type (????_1 ????????????????1 ????????????????????, ????_2 ????????????????2 ????????????????????, …) in both tumor and stromal compartments. Percentage scores per unit area corresponding to the above compartments were generated. Spatial distribution and correlation between different cellular phenotypes were analyzed.
Analysis was performed and concordance between (manual) WSI and FOV scores was calculated.
Results
Data analysis was performed and concordance between WSI and multi-FOV manual scores was evaluated. Results showed high concordance (89.7%) between the two methods, validating the multi-FOV scoring technique.
A total of 28 WSIs stained with TCF7/CD39/CD8 triplex IHC assay using yellow, blue, and red chromogens respectively, yielded 7 different cell phenotypes as follows: CD8+/TCF7-/CD39-, CD8+/TCF7+/CD39-, CD8+/TCF7-/CD39+, CD8-/TCF7+/CD39+, CD8-/TCF7+/CD39-, CD8-/TCF7-/CD39+, and CD8+/TCF7+/CD39+ (Fig 2). All these cell types were observed and studied in tumor and stromal compartment within tumor ROI, which included tumor parenchyma, tumor associated stroma, and invasive margin. The novel FOV based approach was validated on 1 WSI, where entire WSI region was annotated for all cell phenotypes on HALO LINK by a trained pathologist. The scores from WSI and multi-FOV approach were comparable, i.e., sum of errors of each class was 10.3% (Fig 3).
In this cohort, the ratio of TCF7+ CD8+ T cells to CD8+ T cells was higher in the invasive margins and tumor stroma in comparison with that in tumor parenchyma. On the other hand, the ratio of CD39+ CD8+ T cells to CD8+ T cells was higher in the tumor parenchyma.
CD8+ T cells were observed predominantly in both tumor and stroma, whereas only TCF7+ cells were found chiefly in the stroma. The renewable CD8+ T cell subset with TCF7 positivity was observed in the stroma with minimal infiltration in the tumor, whereas the exhausted CD8+ T cell subset with CD39 positivity was found to be well distributed in both tumor and stroma.
Discussion
Analysis of CD8 topology and expression of checkpoint markers such as PD-L1 is the current practice to stratify the patients for ICI therapies, such as anti PD-1 therapy. However, CD8 topology, which checks only the presence of CD8+ T cells in the TME do not provide the in-depth information on the state of CD8+ T cells. Thereby, multiplex Immunohistochemistry (IHC) has become an integral tool in the study of CD8 T cell subsets, such as TCF7 and CD39 positive cells. This technique allows for the simultaneous detection of multiple markers in a single tissue section, thereby providing a comprehensive view of the cellular landscape of the tumor microenvironment. A study by Mitra et al. (2023) highlighted the use of multiplex IHC to evaluate the co-expression of CD38 and CD39 on CD4+ T helper cells within tumor-infiltrating lymphocytes, emphasizing the potential of this approach to inform checkpoint inhibitor resistance.[16]
Furthermore, study by Zhang et al. (2023) investigated the potential of targeting CD39 to overcome radiotherapy resistance in lung cancer through utilizing IHC to examine the distribution of CD39-expressing cells within lung cancer tissue microarrays.[11] Another study by Im et al. (2023) used multiplex IHC to discern the characteristics and anatomic location of PD-1+TCF1+ stem-like CD8 T cells within chronic viral infections and cancer.[17]
In lung cancer patients, CD8 T cell subsets' distribution is believed to have significant implications for the efficacy of immunotherapies. The tumor microenvironment (TME) of lung cancer is often infiltrated by these T cells, and their presence has been linked with improved prognosis and response to therapies such as immune checkpoint inhibitors.
Several studies have explored this association further. For instance, a study by Tumeh et al. (2014) found that the frequency of CD8 T cells within the TME directly correlated with responses to anti-PD1 therapy in non-small cell lung cancer (NSCLC).1 Similarly, Thommen and colleagues (2018) revealed that the presence of CD8 T cells that express PD-1, a key inhibitory receptor, can predict responses to PD-L1 blockade in metastatic NSCLC.[18]
TCF7, a transcription factor, is critical for the maintenance of CD8 T cell stemness and the generation of memory T cells. These TCF7-expressing cells are found to be less exhausted and more responsive to PD-1 blockade, making them prime targets for cancer immunotherapies.[19] In the context of lung cancer, an increased presence of TCF7-expressing CD8 T cells within the tumor microenvironment could potentially enhance the effectiveness of immunotherapies. Recent studies have found that these TCF7-expressing cells are less exhausted and more responsive to immune checkpoint inhibitors, making them prime targets for cancer immunotherapies. For instance, in a study on patients with non-small cell lung cancer undergoing ICI treatment, it was found that the presence of CD8+ T cells expressing TCF7 and inhibitory molecules within the tumor microenvironment could potentially enhance the effectiveness of immunotherapies.[20]
On the other hand, CD39, an ectonucleotidase, is overexpressed on exhausted T cells, including CD8 T cells. CD39, in tandem with CD73, catalyzes the conversion of pro-inflammatory ATP to immune-suppressive adenosine, which can promote tumor growth and resistance to therapy.[10] Therefore, CD39-expressing CD8 T cells might impede the efficacy of immunotherapies in lung cancer patients. Exhausted CD8 T cells are characterized by the sustained expression of multiple inhibitory receptors and loss of effector functions, which is a potential barrier for successful immunotherapies.[21] These studies exemplify the applications of multiplex IHC in understanding the spatial distribution and roles of TCF7 and CD39 positive CD8 T cells in lung cancer, providing valuable insights for enhancing immunotherapy strategies.
In essence, the spatial distribution of TCF7-positive and CD39-positive CD8 T cells in the tumor parenchyma and stroma, respectively, is a critical factor influencing the efficacy of immunotherapies in lung cancers. By understanding and manipulating these factors, it is possible to improve treatment outcomes in lung cancer patients. Understanding the distribution and functional states of CD8 T cell subsets in lung cancer can provide valuable insights for enhancing the efficacy of existing immunotherapies and developing novel therapeutic strategies.
Digital pathology, a rapidly evolving field, offers powerful tools for image analysis, thereby evaluating the spatial distribution of TCF7+ and CD39+ CD8 T cells in multiplex IHC assay images. One key approach is the use of field-of-view (FOV) analysis, which enables detailed examination of specific regions within the tumor parenchyma and stroma.
In the context of TCF7 and CD39, FOV assessment can provide valuable insights into their spatial distribution and relative abundance in different tumor microenvironments. High-resolution imaging techniques can capture detailed FOVs of tumor samples, which can then be digitally analyzed using advanced algorithms. This can help identify the presence and location of TCF7-positive CD8 T cells in the tumor parenchyma and CD39-positive CD8 T cells in the tumor stroma.
Moreover, digital pathology can facilitate a more extensive and detailed analysis than traditional pathology, as it allows for high throughput and automated assessments. Thus leading to robust understanding of the distribution of these T cell subsets. By providing a more comprehensive view of the tumor microenvironment, FOV assessments using digital pathology approaches could potentially inform better therapeutic strategies, enhancing the efficacy of immunotherapies and ultimately improving patient outcomes.
For instance, a study by Sha et al. (2019) utilized digital pathology to predict the status of programmed death-ligand 1 (PD-L1) in non-small cell lung cancer from whole-slide hematoxylin and eosin images. This study emphasized the potential of digital pathology analysis to provide insights that are not readily available through traditional pathology methods.[22]
Another study by Corredor et al. (2018) utilized a watershed and feature-based approach for automated detection of lymphocytes on lung cancer images. They used FOVs extracted from lung cancer whole slide images and were able to automatically detect lymphocytes, demonstrating the potential of digital pathology and FOV analysis.[23]
These studies highlight the potential of using FOV analysis in digital pathology to better understand the distribution and role of specific cell types, such as TCF7 and CD39 positive CD8 T cells, in the tumor microenvironment.
The main advantages of our study are that this method can be used in other indications, it is advantageous when facing challenges in image analysis using digital analysis, and when the cohort of study is small and algorithm development is not feasible.
However, there are few limitations of our study that data for treated vs untreated was unavailable, the cohort used was small, and validation in a large cohort is needed.
Conclusion
The application of an immunohistochemistry panel for CD8 T cell subset identification represents a significant advancement in patient stratification strategies. By overcoming the drawbacks of traditional CD8 phenotyping methods, this approach provides a detailed understanding of the phenotypic heterogeneity within the CD8 T cell pool. The examples presented here highlight the potential clinical impact of this technique across different disease contexts. With further research and validation, the IHC panel can contribute to optimizing treatment outcomes and moving towards precision medicine.
References