We reported that induction gemcitabine, carboplatin and paclitaxel (GCP) failed to improve overall, disease free and distant metastasis free survival in patients with advanced nasopharyngeal carcinoma (NPC) treated with concurrent weekly cisplatin and IMRT. We next examined if pre-treatment plasma neutrophil-lymphocyte ratio (NLR) and EBV DNA predicted for patients who benefit from induction GCP.
Patients with biopsy-proven UICC/AJCC stage III, IVA/B NPC were eligible. Patients received IMRT (69.96 and 60 Gy to GTV and CTV, respectively) with weekly cisplatin (40 mg/m2). Induction GCP comprised of gemcitabine, 1 gm/m2, carboplatin, AUC 2.5, paclitaxel, 70 mg/m2, given on days 1 and 8 3-weekly for 3 cycles. Plasma EBV DNA was quantified using real-time quantitative PCR. NLR was computed as a ratio of absolute neutrophil to lymphocyte counts.
108 of 172 patients in the randomised trial had EBV DNA measurements at baseline. 53 and 55 patients were randomised to GCP and control arms, respectively. Median follow-up duration was 2.3 years (range = 0.11-3.92). 103 were alive at the time of analysis, 24 (22%) experienced disease progression, 15 (14%) had distant metastases and 12 (11%) had local relapses. Uppermost thirtile of EBV DNA titre was 6000 copies/ml and median NLR was 3.0. Patients with EBV DNA ≥6000 copies/ml (EBVhigh) and NLR ≥3.0 (NLRhigh) had a higher likelihood of disease progression and distant metastasis (HR 1.6, 95% CI = 0.7-3.83, DFS; 2.2, 0.77-6.11, DMFS). Likelihood of disease free and distant metastasis free survival was improved with induction GCP regardless of EBV DNA threshold (HR 0.40, EBVhigh vs 0.50, DFS; 0.30, EBVhigh vs 0.50, DMFS). Induction GCP led to significantly improved DFS and DMFS in the NLRhigh subgroup (HR 0.30, 0.07-0.92, p = 0.02, DFS; 0.10, 0.01-0.92, p = 0.01, DMFS), while outcomes did not differ between treatment arms in patients with NLR <3.0 (HR 0.90, 0.25-3.00, DFS; 1.7, 0.31-9.26, DMFS).
Elevated EBV DNA and NLR portend for an unfavourable prognosis in patients with advanced NPC. While the role of induction chemotherapy in an unselected patient cohort is uncertain, NLR may be a useful biomarker to select patients a priori for this treatment.
Clinical trial identification
All authors have declared no conflicts of interest.