911P - Post-radiotherapy (RT) Epstein-Barr virus (EBV) DNA dynamics identifies optimal timepoint for stratification of high-risk nasopharyngeal carcinoma (NPC)

Background

Detectable plasma EBV DNA post-RT portends for a poor prognosis in NPC patients. Adjuvant chemotherapy (AC) trials however failed to yield a survival benefit in these high-risk patients. Studies have shown that post-RT EBV DNA levels and kinetics could further stratify NPC patients into risk-groups with disparate survival. There is thus a need to correctly identify patients who benefit from AC. Here, we investigated whether an early (0-2 w) or delayed (8-12 w) measure of plasma EBV DNA better predicts for disease relapse.

Methods

We curated a prospective single institution NPC registry for patients with a plasma EBV DNA result at 0-2 w post-RT and a minimum follow-up of 2 years. A subset of patients had a duplicate test done at 8-12 w post-RT. Plasma EBV DNA was quantified using the international harmonised BamH1 PCR-based assay. Primary clinical endpoint was disease-free survival (DFS).

Results

297 patients with an EBV DNA result at 0-2 w post-RT were analysed. Median follow up was 51 (IQR: 47-55) mo. 71% (210/297) had TNM-stage III-IVA NPC; and 41% (122/297) reported pre-RT EBV DNA of ≥4000 copies/mL. At 0-2 w post-RT, 82% (243/297) had 0 EBV DNA copy/mL, while 16% (14/297) and 2% (7/297) had 1-264 copies/mL and ≥265 copies/mL, respectively. Using different EBV DNA level cut-offs of 0, 1-499, and ≥500 copies/mL, we validated the RPA model by Hui et al. with the high-risk group (≥500 copies/mL) yielding the worst HR of 4.9 (95%CI: 1.9-12.5) relative to the low-risk group. Next, we observed substantial EBV DNA changes in 109 patients with a repeat EBV DNA at 8-12 w. Specifically, 65% (22/34) with a detectable EBV DNA at 0-2 w had 0 copy/mL subsequently. We found that a detectable EBV DNA at 8-12 w was a stronger predictor of relapse than a positive result at 0-2 w; AUCROC for DFS was 0.806 (95%CI: 0.727-0.886) for stage+EBV DNA_8-12w versus 0.679 (95% CI: 0.576-0.783) for stage+EBV DNA_0-2w.

Conclusions

Our findings revealed that post-RT EBV DNA levels are dynamic, and an EBV DNA test at 8-12 w more accurately identifies high-risk patients than a test at 0-2 w post-RT. This has potential implications on AC trials that adopt an early EBV DNA result for treatment stratification.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Nasopharyngeal Cancer- Large Collaborative Grant (NPC-LCG), National Medical Research Council (NMRC), National Cancer Centre (NCC) Cancer Fund, Goh and Kua Hong Pak Foundation.

Disclosure

Y.L. Soong: Financial Interests, Personal, Stocks/Shares: GSK. D.S.W. Tan: Financial Interests, Personal, Advisory Board: Amgen, Novartis, Boehringer Ingelheim, C4 Therapeutics, AstraZeneca, GSK, Takeda, Eisai, Guardant, Merck, Pfizer, Roche; Financial Interests, Institutional, Research Grant: AstraZeneca, Amgen, Pfizer, ACM Biolabs; Financial Interests, Personal, Steering Committee Member: Novartis. D.W. Lim: Financial Interests, Institutional, Invited Speaker: Boehringer-Ingelheim; Financial Interests, Institutional, Advisory Board: MSD, Roche, BeiGene; Financial Interests, Institutional, Trial Chair, Grant funding for investigator-sponsored study: Bristol Myers Squibb; Financial Interests, Institutional, Steering Committee Member: Novartis. M.L.K. Chua: Financial Interests, Institutional, Invited Speaker: Varian, AstraZeneca, Janssen, Pfizer, MSD; Financial Interests, Personal, Stocks/Shares: Digital Life Line; Financial Interests, Institutional, Other, Research agreement: Decipher Biosciences; Non-Financial Interests, Leadership Role: Head and Neck Cancer International Group. All other authors have declared no conflicts of interest.