The Expression Of PD-L1 And Its Relationship With Prognosis In Different Subtypes Of Esophageal Cancer

Objective:To analyze the relationship between the expression of programmed death ligand-1(PD-L1)and clinicopathological features and prognosis in esophageal cancer.To compare the consistency of four antibodies(SP263,22C3,E1L3N and SP142)in esophageal cancer.Methods:968 patients with radical resection of esophageal cancer from the year 2003 to 2017 in Hebei Medical University were selected,then made tissue microarrays,included all esophageal rare subtypes:59 cases of Sarcomatoid carcinoma(SC),65 cases of basal-like squamous cell carcinoma(BSCC),18 cases of large cell neuroendocrine carcinoma(LCNEC),55 cases of undifferentiated carcinoma,122 cases of adenocarcinoma(AC),96 cases of adenosquamous carcinoma,26 cases of adenoid cystic carcinoma(ACC),160cases of small cell carcinoma,39 cases of mucoepidermoid carcinoma(MEC),85 cases of mixed neuroendocrine carcinoma.There also had 243 cases of squamous cell carcinoma(SCC)that were randomly selected in the corresponding time period.Four kinds of PD-L1 antibodies were used for immunohistochemical staining to detect the expression of PD-L1 in esophageal carcinoma tissues,and the positive was calculated according to tumor proportion score(TPS)and combined positive score(CPS).Statistical software SPSS 25.00 for analysis and processing,χ~2 test for correlation and consistency analysis;Kaplan-Meier method and Log-rank test for single factor survival analysis;Cox regression model for multivariate analysis,P<0.05 for differences with statistics learning meaning.Results:1 Expression of PD-L1 in different subtypes. 1.1 Expression of PD-L1 antibody SP263:When using TPS for interpretation,low expression group(1%≤TPS<50%):SCC was 37.4%(91/243),BSCC was29.2%(19/65),and adenosquamous carcinoma was 28.1%(27/96);high expression group(TPS≥50%):SC was 22.0%(13/59),SCC was 9.9%(24/243),undifferentiated carcinoma was 9.1%(5/55).When using CPS for interpretation(CPS≥10),SCC was 60.1%(146/243),SC was 42.4%(25/59),and AC was 41.8%(51/122). 1.2 Expression of PD-L1 antibody 22C3:When TPS was the standard of interpretation,Low expression group(1%≤TPS<50%):SCC was34.6%(84/243),adenosquamous carcinoma was 27.1%(26/96),SC was23.7%(14/59);High expression group(TPS≥50%):SC was 16.9%(10/59),SCC was 8.6%(21/243),BSCC was 6.2%(4/65);when CPS(CPS≥10)was the interpretation standard,SCC was 57.2%(139/243),SC was 39.0%(23/59),AC was 37.7%(46/122). 1.3 Expression of PD-L1 antibody E1L3N:When using TPS for interpretation,low expression group(1%≤TPS<50%):SCC was 36.6%(89/243),SC was27.1%(16/59),AC was 27%(33/122);High expression group(TPS≥50%):SC was 18.6%(11/59),SCC was 9.5%(23/243),BSCC was 7.7%(17/65);when using CPS for interpretation(CPS≥10),The SCC was 59.7%(145/243),SC was 39.0%(23/65),and both AC and adenosquamous carcinoma were 38.5%(47/122,37/96). 1.4 The expression of PD-L1 antibody SP142 in various types of esophageal cancer:low expression group(1%≤TPS<50%):SCC was 32.5%(79/243),adenosquamous carcinoma was 25.0%(24/96),SC was 23.7%(14/59);high expression group(TPS≥50%):SC was 13.6%(8/59),BSCC was 6.2%(4/65),undifferentiated carcinoma was 5.5%(3/55);when CPS(CPS≥10)was the interpretation standard,SCC was 37.0%(90/243),SC was 32.2%(19/59),BSCC was 27.7%(18/65). In conclusion,the positive expression of PD-L1 in LCNEC was 0;when using TPS or CPS,the positive rate of undifferentiated carcinoma,adenocarcinoma and squamous cell carcinoma changed greatly;when using TPS for interpretation,there were fewer patients with high expression,so The TPS threshold is defined as 1%and the threshold of CPS is 10. 2 With TPS or CPS as the interpretation standard,the consistency of SP263,22C3 and E1L3N was higher(Kappa>0.88),and the consistency with SP142was lower(Kappa minimum is 0.700).Therefore,antibody 22C3 and antibody SP142 were selected for analysis. 3 The relationship between the expression of PD-L1 and pathological features in various types of esophageal cancer: 3.1 SC:22C3(TPS≥1%,CPS≥10),SP142(CPS≥10)were associated with T stage,TNM stage,lymph node metastasis(P<0.05);SP142(TPS≥1%)was also associated with gross type(P<0.05). 3.2 SCC:22C3(TPS≥1%)were associated with tumor length,T stage,lymph node metastasis(P<0.05);22C3(CPS≥10),SP142(TPS≥1%,CPS≥10)was only related to tumor length(P<0.05). 3.3 Undifferentiated carcinoma:22C3(TPS≥1%)was associated with TNM stage(P<0.05);SP142(CPS≥10)was associated with tumor location(P<0.05). 3.4 AC:22C3(TPS≥1%)was associated with TNM stage(P<0.05);22C3(CPS≥10)were associated with TNM stage and lymph node metastasis(P<0.05);SP142(TPS≥1%)was associated with tumor location(P<0.05). 3.5 MEC:SP142(CPS≥10)was associated with age and TNM stage(P<0.05). 3.6 Mixed neuroendocrine carcinoma:22C3(TPS≥1%)was associated with age(P<0.05). 3.7 SCC:22C3(TPS≥1%)were associated with tumor length,TNM stage and lymph node metastasis(P<0.05);22C3(CPS≥10)were associated with TNM stage and lymph node metastasis(P<0.05);SP142(TPS≥1%,CPS≥10)were associated with tumor length and lymph node metastasis(P<0.05). 3.8 The relationship between ACC,small cell carcinoma and PD-L1expression was not founded.(P>0.05). 4 Kaplan-Meier univariate analysis of prognostic factors for esophageal cancer: 4.1 Prognosis of SC were associated with tumor length,TNM stage,lymph node metastasis,22C3(TPS≥1%,CPS≥10)and SP142(TPS≥1%,CPS≥10);4.2 BSCC were associated with tumor length,T stage,TNM stage,lymph node metastasis,22C3(TPS≥1%,CPS≥10)and SP142(TPS≥1%,CPS≥10);4.3 LCNEC wre associated with age,TNM stage and lymph node metastasis(P<0.05); 4.4 Undifferentiated cancer were associated with T stage,TNM stage,and lymph node metastasis(P<0.05); 4.5 AC were associated with TNM stage,lymph node metastasis,22C3(TPS≥1%,CPS≥10)and SP142(TPS≥1%,CPS≥10)(P<0.05); 4.6 Adenosquamous carcinoma were associated with gender,tumor location,gross type,tumor length,TNM stage,lymph node metastasis,22C3(TPS≥1%,CPS≥10)and SP142(TPS≥1%,CPS≥10)(P<0.05); 4.7 ACC was associated with tumor length,T stage,TNM stage and lymph node metastasis(P<0.05); 4.8 Small cell carcinoma were associated with age,T stage,TNM stage,lymph node metastasis and treatment(P<0.05); 4.9 MEC was associated with lymph node metastasis(P<0.05); 4.10 Mixed neuroendocrine carcinoma were associated with tumor length,T stage,TNM stage,lymph node metastasis and 22C3(TPS≥1%)(P<0.05); 4.11 SCC were associated with T stage,TNM stage,lymph node metastasis,treatment,22C3(TPS≥1%,CPS≥10)and SP142(TPS≥1%,CPS≥10)(P<0.05). 5 COX multivariate analysis of independent risk factors for esophageal cancer prognosis: 5.1 SC:Tumor length,lymph node metastasis and SP142(CPS≥10)had significant effects on OS(P<0.05); 5.2 BSCC:TNM stage and 22C3(TPS≥1%)had significant effects on OS(P<0.05); 5.3 LCNEC:Age and TNM stage had significant effects on OS(P<0.05); 5.4 Undifferentiated carcinoma:TNM stage had a significant effect on OS(P<0.05); 5.5 AC:TNM stage and 22C3(CPS≥10)had significant effects on OS(P<0.05); 5.6 Adenosquamous carcinoma:Gender,tumor length and lymph node metastasis had significant effects on OS(P<0.05); 5.7 ACC:TNM stage had a significant effect on OS(P<0.05); 5.8 Small cell carcinoma:Age,TNM stage,and treatment had significant effects on OS(P<0.05); 5.9 MEC:Lymph node metastasis had a significant effect on OS(P<0.05); 5.10 Mixed neuroendocrine carcinoma:TNM stage had a significant effect on OS(P<0.05); 5.11 SCC:Lymph node metastasis,treatment and 22C3(CPS≥10)had significant effects on OS(P<0.05).Conclusion:1.Three PD-L1 antibodies SP263,22C3,and E1L3N have high consistency in esophageal cancer and are less consistent with SP142. 2.When PD-L1(22C3,SP142)TPS threshold is 1%and PD-L1(22C3,SP142)CPS threshold is 10,it can affect the prognosis of SC,AC,adenosquamous carcinoma and SCC. 3.PD-L1(22C3)CPS threshold of 10 is an independent prognostic factor for AC and SCC. 4.The PD-L1(SP142)TPS threshold of 1%is an independent prognostic factor for SC. 5.The PD-L1(22C3)TPS threshold of 1%is an independent prognostic factor for BSCC.