Abstract
Human papillomavirus–associated (HPV+) oropharyngeal carcinoma is associated with excellent survival, yet treatment drives substantial toxicity. Improved biomarkers are needed to select patients for de-escalated treatment. Circulating tumor HPV DNA (ctHPV-DNA) represents a promising noninvasive biomarker to gauge treatment response and surveil for disease recurrence.
A prospective biomarker clinical trial of response-stratified de-escalation was conducted. Eligible patients with non-metastatic HPV+ oropharyngeal carcinoma received neoadjuvant chemotherapy, followed by risk/response-stratified de-escalation with transoral robotic surgery, de-escalated radiation with or without chemotherapy to 50 Gy, or standard chemoradiation to 70 Gy. Deep response (≥50% tumor shrinkage per RECIST v1.1) qualified patients for de-escalation. ctHPV-DNA was measured using HPV-SEQ in plasma at baseline, during neoadjuvant chemotherapy, radiation, and following treatment. The primary endpoint was the correlation of ctHPV-DNA kinetics and radiographic response.
Forty-six eligible patients were enrolled, and 488 ctHPV-DNA samples were analyzed (median 11 per patient). The median follow-up was 30 months, and five recurrences were observed (10.9%). Baseline ctHPV-DNA was detected in 95% of evaluable patients. Rapid early ctHPV-DNA clearance after one cycle of neoadjuvant therapy (≥95% reduction) predicted radiographic deep response (P = 0.04). Detection of ctHPV-DNA 3 months or later after treatment was associated with worse progression-free and overall survival (P < 0.001). Sensitivity, specificity, and positive and negative predictive values of longitudinal ctHPV-DNA were 100%. The longest lead time from positive ctHPV-DNA to detection of recurrent disease was 25 months.
Rapid early clearance of ctHPV-DNA during neoadjuvant therapy demonstrates utility in predicting response to treatment. Detectable ctHPV-DNA following treatment is predictive of both disease recurrence and worse survival.