PD-L1 hotspot in tumor-infiltrating lymphocytes of radically treated esophageal squamous cell carcinoma: Pattern of recurrence and long-term clinical outcomes

Abstract

Introduction: Most cases of esophageal squamous cell carcinoma (ESCC) recurred and eventually succumbed to the disease despite radical treatment. Immune checkpoint inhibitor showed early promise in metastatic disease and preliminary evidence has suggested PD-L1 staining on tumor cells as a predictive marker for treatment outcomes. On the other hand, the role of PD-L1 on tumor-infiltrating lymphocytes (TILs) has not been fully elucidated. We retrospectively analysed PD-L1 immunohistochemical (IHC) staining on both tumor cells and TILs. The pattern of recurrence and long-term clinical outcomes are presented.

Methods: Consecutive cases of ESCC from 2006-2010 were screened in the prospectively-maintained database of the Department of Clinical Oncology and Department of Surgery, Queen Mary Hospital, Hong Kong. Patients were included for subsequent analysis if 1) no evidence of distant metastasis; 2) radical esophagectomy performed; 3) adequate tissue available for IHC staining. Patients were allowed to receive neoadjuvant and/or adjuvant treatment but the specimen for PD-L1 IHC must be obtained before any anti-cancer treatment. Tumors cells were positive for PD-L1 if at least 1% of cells showed positive staining. TILs were positive for PD-L1 (hotspot) if more than 10% of TILs showed positive staining. Demographics were compared with Chi-square test or Fisher's Exact test. Time-to-event endpoints were analyzed with the Kaplan-Meier method. A p value of < 0.05 was regarded as statistically significant. All statistical analyses were carried out by R statistics software. The study was supported by the Small Project Fund, The University of Hong Kong.

Results: Overall 187 patients were screened and 156 patients were eligible for subsequent analysis. Most patients were male (82.1%) and smokers (68.6%). Thoracic tumors were the commonest (89.7%) while multiple tumors were seen in 12.2%; 64.1% of tumor were T2 or above and 38.5% were node-positive. Over 60% of patients received neoadjuvant and/or adjuvant treatment.

After a median follow-up of 8.0 years (95% CI 7.27, 8.34) of surviving patients, there were 66 recurrences and 105 deaths. The median overall survival (mOS) was 3.9 years (95% CI 2.63, 5.28) for the whole population. PD-L1 was positive in 50.6% of tumor cells while PD-L1 hotspot was positive in 30.8% of TILs. Among those with recurrence, patients who were positive PD-L1 hotspot on TILs had significantly lower risk of distant recurrence (42.1% vs 72.3%; p = 0.042) but similar rate of local and regional recurrence as compared to those without PD-L1 hotspot. They also had significantly lower rate of recurrence-free survival with difference at 1- to 4-year by 18-21% (p-values ranged from 0.010 to 0.038) but not at 5-year (35% vs 51%; p = 0.057). OS was significantly better in those with PD-L1 hotspot (mOS 6.26 vs 2.63 years, p = 0.009). On the other hand, PD-L1 on tumor cells was not associated with pattern of recurrence or survival outcomes.

Conclusion: In this study with over 8 years of follow-up, PD-L1 hotspot on TILs was associated with lower rate of distant recurrence and improved clinical outcomes in radically-treated ESCC. Its predictive value for immune checkpoint inhibitor warrant further study.