Early change of plasma Epstein-Barr virus DNA load and the viral lytic genome level could positively predict clinical outcome in recurrent or metastatic nasopharyngeal carcinoma receiving anti-programmed cell death 1 monotherapy

Abstract

Purpose: Patients with recurrent or metastatic nasopharyngeal carcinoma (RM-NPC) have proven benefit from anti-programmed cell death 1 (anti-PD-1) monotherapy. Here, we retrospectively analyze the association of plasma Epstein-Barr virus (EBV) DNA load and tumor viral lytic genome with clinical outcome from 2 registered phase I trials.

Methods: Patients with RM-NPC from Checkmate 077 (nivolumab phase I trial in China) and Camrelizumab phase I trial between March 2016 and January 2018 were enrolled. Baseline EBV DNA titers were tested in 68 patients and EBV assessment was performed in 60 patients who had at least 3 post-baseline timepoints of EBV data and at least 1 post-baseline timepoint of radiographic assessment. We defined "EBV response" as 3 consecutive timepoints of load below 50% of baseline, and "EBV progression" as 3 consecutive timepoints of load above 150% of baseline. Whole-exome sequencing was performed in 60 patients with available tumor samples.

Results: We found that the baseline EBV DNA load was positively correlated with tumor size (spearman p < 0.001). Both partial response (PR) and stable disease (SD) patients had significantly lower EBV load than progression disease (PD) patients. EBV assessment was highly consistent with radiographic evaluation. Patients with EBV response had significantly improved overall survival (OS) than patients with EBV progression (log-rank p = 0.004, HR = 0.351 [95% CI: 0.171-0.720], median 22.5 vs. 11.9 months). The median time to initial EBV response and progression were 25 and 36 days prior to initial radiographic response and progression, respectively. Patients with high levels of EBV lytic genomes at baseline, including BKRF2, BKRF3 and BKRF4, had better progression-free survival (PFS) and OS.

Conclusion: In summary, early clearance of plasma EBV DNA load and high levels of lytic EBV genes were associated with better clinical outcome in patients with RM-NPC receiving anti-PD-1 monotherapy.

Keywords: Biomarkers; EBV; Immunotherapy; NPC.

Fig. 1

Baseline Epstein-Barr virus (EBV) load and clinical outcome. A. Spearman correlation between tumor size and plasma EBV copy number at baseline. B. Boxplot of log10 baseline viral load by Response Evaluation Criteria in Solid Tumors group. ORR, objective response rate; PD, progression disease; PR, partial remission; SD, stable disease

Fig. 2

Change of EBV load during treatment and response to immunotherapy. A. Waterfall plot of change from baseline in tumor size for patients with nasopharyngeal carcinoma. B. Change from baseline in EBV viral load for patients with PD and PR. C. OS curve of patients with EBV response vs. those with EBV progression. D. Swimmer plot that shows time to initial EBV response/progression and RECIST response/progression in PR and PD patients. RECIST, Response Evaluation Criteria in Solid Tumors; PD, progression disease; PR, partial remission; SD, stable disease; EBV, Epstein-Barr virus; OS, overall survival

Fig. 3

EBV and RECIST assessment of 1 PR and 1 PD patients. A. Tumor size and plasma EBV load change over time in a PD patient. EBV load was on log10 scale. B. Tumor size and plasma EBV load change over time in a PR patient. EBV load was on log10 scale. EBV: Epstein-Barr virus; RECIST: Response Evaluation Criteria in Solid Tumors; PD: progression disease; PR: partial remission

Fig. 4

Levels of BKRF2, BKRF3 and BKRF4 in patients with durable clinical benefit (DCB) and non-durable clinical benefit (NDB)

Fig. 5

The effect of levels of BKRF2, BKRF3 and BKRF4 on patients’ progression-free survival (PFS) and overall survival (OS). PFS (A) and OS (B) curve stratified by median levels of BKRF2. PFS (C) and OS (D) curve stratified by median levels of BKRF3. PFS (E) and OS (F) curve stratified by median levels of BKRF4