Expression of Programmed Death Ligand1 (PD-L1) in Gastric Carcinoma (Histopathological and Immunohistochemical Study)

Objectives: To evaluate immunohistochemical expression of PD-L1 in cases of gastric adenocarcinoma. To correlate PD-L1 immuno-histochemical expression with other available clinico-pathological parameters such as age, sex, grade, stage, lymph node (L.N) metastasis and others. Material and methods: The present retrospective study retrieved the data and archived paraffin blocks of 60 cases of Gastric carcinoma. Immunohistochemical evaluation was done to assess the expressions of PD-L1 in the tumor cells (TC), tumor infiltrated lymphocytes (TILs) and combined positive score (CPS). Results: TC PD-L1 expression was detected in 56.7% of cases, TILs PD-L1 expression was detected in 53.3 % of cases and CPS PD-L1 expression was detected in 63.3% of case, with no statistically significant correlation with clinico-pathological parameters except TILs PD-L1 expression showed statistically significant correlation with positive TILs (P value ˂0.019). Conclusion: Our findings supported the expression of PD-L1 by TC, TILs, and CPS in gastric cancer, with increased expression in a subpopulation of TILs rich in PD-L1 identifying them as potential targets for PD-1/PD-L1 therapy.

The inhibition of programmed death-1 (PD-1) and programmed death-ligand 1 is one of the most often used mechanisms underlying immunotherapy (PD-L1). Tumor immune evasion is caused by the PD-1/PD-L1 interaction, which functions as an inhibitory factor in the final stage of the cancer immunity cycle. It causes functional impairment of antigen-specific T cells (Pardoll, 2012) and (Chen and Mellman, 2013).

Study Design and Patients'data
The study included 60 cases of gastric carcinoma that underwent total or subtotal gastrectomy. The study's data were gathered as formalin-fixed, paraffin-embedded tissue blocks between January 2016 and September 2019 from the pathology department archives at Kasr AL-Ainy, the Faculty of medicine at Cairo University, and a private laboratory. Cases were included if they have confirmed diagnosis of gastric carcinoma. We excluded cases who had undergone prior preoperative radiation, chemotherapy or targeted therapy, and Cases with evidence of other types of malignant neoplasms or having an associated benign gastric lesion. The following data were collected from the records of eligible cases before our histopathological reevaluation and immunohistological staining: sex, age at presentation, site of tumor, histologic diagnosis of the tumor, World Health Organization (WHO) grading.

Histopathological Examination
The paraffin blocks were sectioned at 3-4 μm thick and stained with routine Hematoxylin and Eosin stain. The grading of the tumor and histopathological classification were done in concordance with the WHO classification of digestive system ,

PDL-I Immunohistochemical Staining and evaluation
Immunostaining was done using BenchMark XT (Ventana) autostainer. Human tonsil sections were used as a positive control for cases stained for PDL-1 because they show significant staining in some crypt epithelial regions and weak to moderate staining of the follicular macrophages in the germinal centres. The surface epithelium, fibroblasts, and endothelium should all express PD-L1, which is employed as an internal & negative control ("Interpretation Manual -Gastric," 2019). Staining intensities and percentages were scored in both the tumor cells (TC) and the TILs.

Scoring of PD-L1 in Tumor cells
For PD-L1 staining of tumor cells, membrane staining alone was taken as positive and further subjected to the following immunoreactivity scoring system (IRS) (Park et al.,2014).
B: staining intensity-0 (no immunostaining), 1 (weak staining/light yellow), 2 (moderate staining/brown), and 3 (strong staining/dark brown). The addition of category A and B resulted in a total score ranging from 0 and 7. A total score of ≤2 was considered negative and >2 was considered positive (Wei et al., 2018).
Scoring of PD-L1 in Tumor infiltrating lymphocytes For PD-L1 staining of immune cells, membrane or/ cytoplasmic staining was taken as positive and further subjected to the following immunoreactivity scoring system.

Combined Positive Score PD-L1 expression
PD-L1 staining cells include tumor cells, lymphocytes, and macrophages. The combined positive score (CPS) is calculated by multiplying of PD-L1 staining cells by the total number of viable tumor cells (Liu et al., 2020). CPS considered negative if <1, and considered positive if>1. CPS is defined accordingly:

Statistical Analysis
The histopathological and immunohistochemical data were then transferred to the SPSS Software program, version 24 to be statistically analysed. Simple descriptive statistics (arithmetic mean and standard deviation) were used for summary of quantitative data and frequencies were used for qualitative data. Estimation of the association between categorical variables was performed using the chi-square test. P value < 0.05 is considered as statistically significant.

Results
study including 60 cases of gastric carcinoma. Most cases ≤55 years (26 cases) representing 52%. The mean age of cases was 55.56 years, with standard deviation12.477. Range (25-80); median 55 years (data of 10 cases regarding the age could not be identified in the records). Thirty Two cases (53.3%) were males and twenty eight cases (46.7%) were females, with male to female ratio was 1:0.9. Regarding the histological type, most cases were tubular moderately differentiated adenocarcinoma (22 cases, 36.7%). Regarding the histological grade: Among the collected cases, 37 cases (61.7%) were high grade and 23 cases (38.3%) were low grade . Regarding the T stage, most cases were T3 (46.6%), Regarding the lymph nodal metastasis of the studied cases, 10 cases were negative for lymph node metastasis (N0; 16.7%), while 50 cases (83.3%) were positive for metastasis. Among positive cases; 9 cases were N1 (15%), 16 cases were N2 (26.6%) and 25 cases were N3 (41.7%). Regarding tumor size; ≥5 cm were representing most cases (71.7%) .Regarding tumor site; 37 cases were present in the body of the stomach representing 68.5% .out of 54 cases of the total 60 studied cases (N.B. specific site of six cases could  The pathological characteristics of the studied cases are summarized in (Table 1).
Regarding the relation between the age and gender of the studied cases on one side and Tumor cell(TC) PD-L1 expression on the other side, cases of both age groups showed 50%positive PD-L1expression (P value = 0.768, insignificant) . Females showed more expressionrate52.9% As for the lymph node metastasis cases with positive metastasis showed highest TC, TILs and CPS PD-L1 expression (82.4%, P value =0.816, insignificant), (84.4%, P value =0.817, insignificant) and (84.2%, P value =0.811, insignificant) respectively .

Discussion
Gastric carcinoma is one of the most common and fatal malignancy worldwide. The prognosis of gastric carcinoma remains poor despite the use of multidisciplinary treatments. So there was an urgent need to develop new therapeutic modalities in gastric carcinoma such as immunotherapy and targeted therapy. And Anti-PD-L1 is one of the; recently described therapies. PDL-1 abolishes the antitumor immune response through induction of apoptosis of cytotoxic T cells (Soliman and Ibrahim, (2020). In this study, we recruited 60 tumor sections from Gastrectomy specimens collected from the pathology department at the Kasr el Aini hospital and some private laboratories in the period between January 2016 and September 2019.
Regarding the relation between the age and gender of the studied cases and TILs PD-L1, age group >55 years showed (52%) higher positive PD-L1 expression and showed no difference in positive PD-L1 expression between males and females (P value = 0.571 and 0.580 respectively, insignificant) .
Regarding the relation between the age & gender of the studied cases and CPS PD-L1 expression, age group ≤55years showed higher positive PD-L1 expression (53.3%) ,and females showed more expression rate 52.6% (P value = 0.817 and 0.224 respectively, insignificant).
The various antibodies utilised in each investigation, as well as the various methodologies (manual vs. automated, entire tumour blocks, or endoscopic biopsy) and interpretations, can be used to explain the significant differences in the expression pattern of PDL-1 (different cutoff value of positivity).
Regarding the age of patient, in our study age ≤55 showed the highest rates of PD-L1 expression in TC,TIL and CPS and CPS, but there was no significant association between them. This agreed with the results of other studies that showed high rate of PD-L1 expression in age group <59 years with no significant correlation between them Fang et al., (2017) and Soliman and Ibrahim,( 2020). On the other hand, Wang et al., (2018) showed highest rate of PD-L1 expression in old age >60 with significant statistical correlation between them.
Regarding the gender of patient, in our study female gender showed higher rate of PD-L1 expression in TC, TILs and CPS than male. But there was no significant association between them.This agree with results of studies done by Wang et al., (2018) and Rhee et al., (2020) with insignificant correlation, But Soliman and Ibrahim, (2020) showed significant correlation between TILs and CPS PD-L1 expression with gender.
In the present study, regarding the histological subtypes, tubular moderately differentiated and tubular poorly differentiated adenocarcinoma (intestinal type) showed the highest rates of PD-L1 expression in TC, TILs and CPS with insignificant correlation between them. This was similar to the results of Wang et al., (2018) and Soliman and Ibrahim, (2020) which showed that the highest rates of PD-L1 expression was in intestinal type, with significant correlation.
In our study, high grade cases showed higher rates of PD-L1 expression in TC, TILs and CPS than low grade cases, with insignificant correlation. This was consistent with most of the studies did by Ju et al., (2017); Fang et al., (2017);Xing et al., (2018), and Soliman and Ibrahim, (2020). In the other hand Rhee et al., (2020)showed significant correlation between them.
Concerning the T stage, our study showed that T3 exhibited the highest rate of PD-L1 expression on TC, TILs and CPS, with insignificant correlation. Wang et al., (2018 reported similar results. Though, Soliman and Ibrahim, (2020) showed significant correlation.
In this study, cases positive for LN metastasis showed higher rates of PD-L1 expression in TC, TILs and CPS  Table 3. Continued than LN negative cases, with insignificant correlation. This agreed with study of Wang et al., (2018).
Among the LN positive cases in our study, the rate of PD-L1 expression increased with the increase in the N stage; e.g. PD-L1 expression in TC N1 (17.6%), N2 (23.5%) and N3 (41.2%). This was consistent with the results of Rhee et al., (2020). In the present study, cases with tumor size ≥5 cm showed higher rates of PD-L1 expression on TC, TILs and CPS than tumor size <5 cm, with insignificant correlation between them. This disagree with studies of Wang et al., (2018) and Soliman and Ibrahim (2020).
In the present study, cases with tumor located in body region showed higher rates of PD-L1 expression on TC, TILs and CPS than tumor located in pyloric region, with insignificant correlation between them. This was consistent with the results of Wang et al., (2018). Although, Soliman and Ibrahim (2020) showed high expression in antrum region. In this study, cases positive for LV and peineural invasion showed higher rates of PD-L1 expression on TC, TILs and CPS than LV and peineural negative invasion cases, with insignificant correlation between them. This disagreed with study done by Soliman and Ibrahim, (2020) Which showed high expression of PD-L1 in cases with negative LV and perineural invasion.
In our study, cases with positive TILs (≥ 25%) showed higher rate of PD-L1 expression in TC, TILs and CPS than cases with negative TILs (<25%). Additionally, there was significant correlation between TILs with PD-L1 expression in TILs (P value = 0.019).This was consistent with the results of Ju et al., (2017) and Fang et al., (2017). This result led us to speculate that PD-L1 expression on TILs, rather than tumor cells, could be more relevant to the inhibitory effects of anti PD-L1therapy, and support the use of PD-L1 expression on TILs as a predictive biomarker in GC for immunotherapies.
The different sample size, different variants, different antibodies used to detect PD-L1, absence of universal standard for PD-L1 expression and different methodology enrolled in the studies might explain the contradictory results regarding correlation between PDL-1 expression in gastric carcinoma and other clinico-pathological parameters. Further studies are needed to reveal the precise mechanism of PD-L1 up-regulation in GC.