In the present study, an IHC-based quantification of TILs was performed. TILs were present in all cases, but we observed that the level of TILs infiltration was considerably variable among patients and even between intra-tumoral and tumor margin areas in the same patient. Indeed, high rates of CD3+ and CD8+ were more likely to be found at the tumor margin with a statistically significant difference (p = 0.03; p = 0.04).

These findings may be in part due to the patient and tumor heterogeneity and the variable molecular subtypes [11]. In fact, in a previous whole-genome gene expression analysis of UC, Sjödahl G. et al. concluded that molecular subtypes of UC display immunological responses at different levels to such an extent that a separate group of “infiltrated/immunogenic” tumors could be observed [12].

Moreover, it could be argued that the protocol for TILs quantification may considerably affect the results. On breast cancer, based on the recommendation of the consensus working group for TILs assessment, quantification is made on hematoxylin-eosin slides, and there is no need for IHC profiling [6]. However, most studies have found that only quantification of stromal TILs is reproducible, and that intra-tumoral TILs are difficult to detect on hematoxylin-eosin-stained slides and are mostly confused with tumor cell apoptosis [6]. In bladder cancer, although some studies have reported important roles for TILs, no convenient and effective counting system is in place. According to some authors, TILs quantification based on the validated recommendations of the consensus working group can be easily performed on UC. However, it is recommended to separately report stromal and intra-tumoral TILs [13]. However, as described in the present paper, most published studies that investigated on TILs in BC used the immunohistochemistry method which has many benefits comparing to the standard morphological method. Indeed, firstly, it could accurately detect the intra-tumoral immune cells component which contributes to TILs [8]. Secondly, with IHC analysis, the different subtypes of TILs are identified which seems to be of great importance since previous studies have advanced that the prognostic impact of TLS depends on the subtypes of the immune cells [14,15,16,17]. Thirdly, immunohistochemical profiling of TILs could be translated into an immunoscore (IS). The latter is a standardized scoring system derived from a measure of CD3+ and CD8+ cell densities at the tumor center and invasive margin. Findings have shown it to be a reliable and statistically significant prognostic tool for predicting the survival of patients with digestive cancers [7, 18].

In BC, Xiang-Dong Li et al. suggested that the immunoscore is also a reliable indicator of an improved prognosis for patients with UC [8]. In the same context, in their reported paper, Mlecnik B. et al. observed that high immunoscore and high density of CD8+ cytotoxic TILs correlated with a decreased risk of metastasis [5].

In the present study, the author’s results were consistent with the previous reported data. Indeed, we noted that a high IS and high levels of CD8+ TILs were significantly associated with good prognostic factors, such as the absence of vascular invasion. We also observed that the CD8+ density was inversely correlated to the age of patients, which is probably due to the aging immune system as stated by Giefing-Kroll C. et al. [19].

While other studies that investigated the prognostic role of TILs in BC suggested that high levels of CD3+ and CD8+ TILs were correlated with a poor outcome, relapse, and reduced overall survival [9, 20]. These inconsistencies may be in part justified by the different samples selection and the variable methods of TILs quantification.

In the present study, the author’s focused on the correlation of TILs and IS to clinical and pathological factors. As previously described, we found that low densities of CD8+ TILs were associated with and advanced age, with a common histological subtype and the presence of vascular invasion. Hence, at the opposite increased, CD8+ lymphocytes seem to reduce tumor extension beyond vessels and protect from vascular micrometastasis. However, we did not establish any significant correlation between TILs populations, IS, and the studied clinicopathological factors. These results were consistent with most published studies [21,22,23,24]. However, some other authors reported a positive association between TILs densities and factors such as histology variants, tumor size, grade, multiplicity, and lymph node status [14, 15, 25,26,27].

In BC, the prognostic impact of TILs and IS on survival is still an object of debate. Sharma et al. were the first to investigate the prognostic impact of the immune system in BC and concluded to a positive association between TILs densities and a prolonged OS and DFS on univariate and multivariate analyses [21]. Otto et al. reported a correlation between a high intra-tumoral TILs density and a prolonged OS [28]. However, based on their findings, Hulsen et al. stated that a superior OS is associated with a high peri-tumoral CD3+ density, but a high IT TILs density was associated with a deteriorated OS [9].

In the present study, no significant association between TILs densities IS and neither OS nor DFS was noted. Nevertheless, the OS and DFS medians were superior in highly T-cell-infiltrated tumors along with tumors with a high IS.

These discrepancies between the different studies regarding the prognostic value of TILs and IS on BC and its impact on survival may also be in part explained by the differences in the degree of TILs activation depending on the areas of their distribution. Indeed, it has been proven that intra-tumoral CD8+ TILs are mostly effector memory cells that are retained on the tumor site using the integrin CD103 [29]. The latter is also involved in the anti-tumoral cytotoxic activity of CD8+ TILs which could explain their association with a good prognosis. On the opposite, the TILs in the peri-tumoral area could contain a proportion of non-effector CD8+ TILs which could be related to T-cell exhaustion due to a dysfunction of the PD-1/programmed cell death ligand (PD-L1) pathway [15]. Hence, an increased rate of CD8+ PDL1 TILs in the peri-tumor areas is more likely to be associated with a bad outcome [15]. In this context, the analysis of TILs in association with the expression of PDL-1 could not only give further prognostic biomarkers but also predict the tumor response to immunotherapy.

Hence, the immune microenvironment of UC can be considered an interesting emerging biomarker for prognosis, or even as markers for bladder cancer therapy, which can be a focus of immunotherapy for bladder cancer.

However, further multicenter studies through standardized quantification methods of TILs and IS in BC are needed to definitively assess their prognostic value.