Emerging evidence has demonstrated that tumor microenvironment (TME) could modulate cancer progression whereas the characteristics of TME may be used for molecular classification to predict treatment response and cancer survival (1, 2). Among different cell types observed in TME, tumor-infiltrating lymphocytes (TILs), including macrophages, dendritic cells, mast cells, natural killer cells, naive and memory lymphocytes, B cells, and effector T cells, are suggested as the main players in modulating cancer progression (3–5). Recent studies have indeed shown an association of TILs with the prognosis of several cancer types (5–7). For instance, a high density of infiltrating CD3+ T cells, CD8+ cytotoxic T cells and CD45RO+ memory T cells was reported to indicate a long overall survival (OS) and progression-free survival among patients with gastric, liver, lung, nasopharyngeal, and colorectal cancers (1, 8–11). Contradicting findings do also exist. For instance, one study reported that high density of CD8+ cytotoxic T cells was associated with larger tumor size in glioblastoma patients (12). Regardless, the existing literature does seem to suggest that TILs might have a generic role in the progression of many solid tumors.

Neuroblastoma (NB) is a common tumor deriving from sympathoadrenal progenitor cells within the neural crest and has substantially varying clinical courses ranging from spontaneous regression to widespread metastasis despite intensive therapy (13, 14). The International Neuroblastoma Staging System (INSS) is currently used to determine treatment strategy and assessment of NB prognosis. NB patients are also commonly classified as at high-, intermediate-, or low- risk of death according to the Children’s Oncology Group (COG) risk classification (15). The different risk groups might demonstrate specific genetic features. TERT rearrangement and inactivating mutations in ATRX have for instance been found predominantly among patients with high-risk NB (16). However, neither the staging systems nor the genetic variations seem to explain completely the extremely divergent clinical course of NB.

With the wide application of cancer immunotherapies, it becomes increasingly important to understand the immunological landscape of the TME to understand the divergent course of NB and to predict treatment response. A recent study showed that CD4+ T cells and macrophages are involved in the oncogenesis of NB (17). Marco et al. showed that tumor-infiltrating T cells, mainly CD3+, CD4+, and CD8+ T cells, improved clinical outcomes of patients with therapy-resistant NB (18) whereas Riyue et al. demonstrated that patients with high-risk NB showed a better survival when the tumor was T cell inflamed (19). A detailed description of the phenotypes of all infiltrating T cells, especially CD3+ T cells, CD8+ cytotoxic T cells, and CD45RO+ memory T cells, using a large collection of NB specimens is however currently not available.

To this end, we investigated the densities of tumor-infiltrating CD3+, CD8+ and CD45RO+ T cells in a large sample of NB patients using digital pathology and developed a new immunoscore, summarizing such densities, to assess its role in predicting OS and event-free survival (EFS) of NB patients.

In a study of 244 incident patients with NB, we performed the IHC analysis and digital pathology to assess the prognostic values of the densities of tumor-infiltrating CD3+, CD8+ and CD45RO+ T cells in NB survival. We found a high density of CD3+ and CD8+ T cells to indicate longer OS and EFS whereas an immunoscore of low densities of both CD3 and CD8 to indicate the highest risk of death or any progression event.

A previous study has reported that TME within tumor regions, including tumor interior and invasive margins, may be defined by the type, functional orientation, density, and location of immune cells (26). Therefore, tumor-infiltrating immune cells have obtained considerable research interest, especially in terms of their prognostic values in treatment response and caner survival (27). Several studies have shown that the density and distribution of TILs based on H&E slides are valuable prognostic makers in breast (28), non-small cell lung (29) and gastric (30) cancers. Moreover, studies have shown that the density of TILs is independently associated with treatment response of neoadjuvant chemotherapy in triple-negative breast cancer with HER2-overexpressing disease and homologous recombination deficient status (31, 32).

Galon et al. reported however that using H&E slides to evaluate density of TILs was not reproducible and highly subjective, and suggested that immunoscore through quantifying specific T cells such as CD3+, CD8+ and CD45RO+ T cells might be more reproducible and objective (1). Such immunoscore has indeed been reported as prognostic indicators in many cancers. Frank et al. showed for instance that CD3+, CD8+ and CD45RO+ T cells were correlated with cancer recurrence and OS in patients with rectal cancer and an immunoscore integrating the three was a powerful prognostic tool and might supplement the TNM staging system (33). In this study, we used digital pathology and showed that CD3+ T cells and CD8+ T cells mirrored each other as previously shown in non-small cell lung cancer and colorectal cancer (1, 10), similarly suggesting that high densities of CD3+ and CD8+ T cells were associated with a longer OS in NB.

High densities of TILs appear to decrease the risk of cancer recurrence, however, the driving forces for the differential densities remain unknown. It has been speculated that both host and tumor factors might contribute to the density of TILs. Therefore the densities of CD3+, CD8+, and CD45RO+ T cells in NB may also differ by clinical variables. In our study, we found that the density of CD3+ T cells was higher than that of CD8+ and CD45RO+ T cells. Further, density of CD3+ T cells was higher among NB patients with MYCN nonamplification and male, density of CD8+ T cells was higher among NB patients with differentiated pathology and male, whereas density of CD45RO+ T cells was higher among NB patients diagnosed above 18 months of age. These suggest that CD3+, CD8+ and CD45RO+ T cells might have different roles in the prognosis of NB, namely that CD3+ T cells are mostly strongly associated with OS and EFS of NB, followed by CD8+ T cells, as suggested in a previous study (18), whereas CD45RO+ T cells might have little or a harmful effect on NB survival. In general, a high density of CD45RO has been related to good clinical outcomes of several solid tumors, including melanoma, head and neck cancer, lung cancer, and colorectal cancer (34). The contradictory finding of NB and the previous studies might suggest that the role of CD45RO might be specific to tumor type and TME. Indeed, in contrast to other solid tumors, NB is a pediatric cancer with low mutation burden, little infiltrating immune cells, and poor response to immune checkpoint inhibitors (35, 36).

It is known that the tumor microenvironment is both diverse and complicated as different immune cells are capable of infiltrating tumor tissues, indicating that a synergistic view of different immune cells might provide better understanding of the immune microenvironment. For example, one study reported that the number of CD8+ TILs alone could not predict survival of patients with glioma, whereas the combination of having both low density of CD8+ and high density of CD4+ predicted survival (33). In the present study, we classified the NB patients into four groups based on the densities of CD3+ and CD8+ T cells and found that NB patients with CD3-CD8- had the highest risk of death and progression event, compared with other patients. Our results add therefore important new knowledge to existing literature suggesting a beneficial effect of cytotoxic T lymphocytes in tumors of diverse origin (11, 26, 31).

Digital pathology is gaining momentum in the analysis of pathologic tissue samples due to its accuracy and quantification of the full-view slides, providing a tool for more automatic and objective evaluation (37). Another advantage is the possibility to facilitate complex spatial patterns and standard metrics (38). In our study, we conducted digital pathology to evaluate the densities of CD3+, CD8+ and CD45RO+ T cells in NB tumor tissues. Due to the lower overall densities of infiltrating immune cells in NB tumors, compared to other cancers, we evaluated the densities of these markers in the indicated slides as a whole. Franck et al. have also used digital pathology and presented an immunoscore based on the densities of CD3+ and CD8+ T cells measured in the core and invasive margins of tumors, and found it to predict the survival of patients with rectal cancer (33). Therefore, immunoscore using digital pathology made it possible to determine the immunotype for each NB patient, strongly indicative of prognosis.

Novel tumor classification has also been proposed based on the immunoscore, as hot (high immunoscore), altered (intermediate immunoscore), and cold (low immunoscore) tumors (39). Immunotherapies are the most rapidly growing field in cancer treatment and have a major impact in clinical outcomes of cancer patients. A consensus was made that effector T cells play the central role in antitumor responses (40). Several studies have reported that hot tumors show a fertile ground for effective immune checkpoint inhibitors (ICIs) based on monotherapy or combination therapies, whereas cold tumors featured by low immunoscore should be treated with combined priming therapy, which may enhance T cell responses to convert an immune cold tumor to an immune hot tumor (41). NB is in general characterized as low immunogenicity due to the lower tumor mutation burden and lower infiltrating of immune cells (42), impeding the effective engagement during immunotherapy (43). Thus, a different or more extensive biomarker should be developed for successful implementation of novel immunotherapies in NB patients to improve survival.

There are several limitations in this study. First, we did not evaluate TILs density using H&E slides. However, it remains interesting to incorporate information on the presence of tertiary lymphoid structures (TLSs) and TILs to the immunoscore, which will likely lead to better understanding of the functional role of immune contexture in NB. Another limitation is the lack of independent validation. Future studies of independent samples are therefore needed to validate our findings.

In conclusion, our work shows that high densities of CD3+ and CD8+ T cells are indicative of longer OS and EFS in NB patients. Additionally, an immunoscore based on the densities of CD3+ and CD8+ T cells may help to classify patients with NB into different prognostic profiles. Together, these findings might pave the way for a novel implementation of immunotherapy in NB patients.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Institute Research Ethics Committee of Guangzhou Women and Children’s Medical Center [2021]078A01. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

H-YW and NL performed study concept and design. LZ, S-HL, KC, and L-JQ conducted experiment. LZ and KC reviewed IHC slides. S-YX conducted digital pathology. H-YW, LZ, X-YL, S-YX, and RW performed data analysis and interpretation. FW, SF, S-HL, F-HW, LL, and LM retrieved clinical data. H-YW, LZ, and X-YL wrote manuscript drafting. All authors have reviewed and approved the final manuscript for publication.

This work was partially supported by the start-up program of Guangzhou Women and Children’s Medical Center (No. 3001151-04) and the National Natural Science Foundation of China (81602468).

Author S-YX was employed by Bio-totem Pte. Ltd.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer SW declared a shared affiliation with the authors X-YL, FW, LD, and NL to the handling editor at the time of review.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.