Clinical validation of an HPV whole genome sequencing assay for molecular residual disease detection in HPV-associated head and neck cancer patients treated with surgery

BackgroundSurgery is a common treatment for early-stage HPV-associated head and neck squamous cell carcinoma (HPV+HNC). Selection of patients who require adjuvant treatment is based on clinicopathologic risk factors, which have poor individualized prognostic capacity. Circulating tumor HPV DNA (ctHPVDNA) is a highly sensitive and specific biomarker for HPV+HNC at diagnosis, but current clinically available assays lack the necessary sensitivity for accurate minimal residual disease (MRD) detection after surgery. Here, we applied a significantly more sensitive HPV whole genome sequencing (WGS) assay to determine the prognostic value of ctHPVDNA-based MRD detection and compare this head-to-head with existing approaches and clinical standard of care. Patients and methods103 patients with AJCC 8 Stage I-IV HPV+HNC treated with definitive surgery were prospectively enrolled. Blood was collected before surgery, after surgery, and in surveillance and analyzed by clinically validated HPV WGS and droplet digital (dd)PCR assays. The primary hypothesis tested was that patients with MRD detection after surgery would have inferior disease-free survival (DFS) and overall survival (OS). ResultsWith a median follow up of 27 months, patients with ctHPVDNA detected after surgery had significantly worse 2-year DFS and OS compared to those without ctHPVDNA (DFS 60%, 95% CI:31-80% vs 100%, p<0.001; OS 73%, 95% CI:43-89% vs 98%, 95% CI:88-100, p=0.002). Patients with ctHPVDNA detected following multi-modality treatment completion also had significantly worse 2-year DFS and OS compared to patients without ctHPVDNA (DFS 0% vs 100%, p<0.001) (OS 50%, 95% CI:11-80% vs 100%, p<0.001). MRD status was a stronger predictor of DFS than standard clinicopathologic criteria (HR 25.2; p = 0.003). Lead time from molecular detection of recurrence to clinical detection of recurrence was up to 17.5 months and nearly twice as long compared to ddPCR (7.1 vs 4.1 months). ConclusionApplying an ultrasensitive HPV WGS liquid biopsy, HPV+HNC patients with ctHPVDNA detected after surgery and following treatment completion had significantly worse DFS and OS, highlighting the potential for MRD status for personalized adjuvant treatment decision-making. HighlightsO_LIApplying an ultrasensitive HPV WGS liquid biopsy, HPV+HNC patients with MRD after surgery and following treatment completion had significantly worse DFS and OS C_LIO_LIPatients with MRD had a strong benefit from adjuvant treatment in decreasing recurrence and death while patients without MRD showed no benefit C_LIO_LIMRD status after surgery outperformed conventional clinicopathologic features in predicting recurrence C_LIO_LIMedian lead time from molecular detection of recurrence to clinical detection was >7 months, and up to 17.5 months, double existing liquid biopsy approaches C_LI