Ultrasensitive HPV Blood Test Predicts Early Recurrence in Head and Neck Cancer

Postoperative decisions about adjuvant therapy are generally based on clinicopathologic risk factors that may not accurately reflect an individual’s likelihood of relapse. More sensitive blood-based approaches capable of detecting tumor-derived HPV DNA could improve risk assessment and surveillance. New findings demonstrate that an ultrasensitive liquid biopsy can detect residual disease and predict recurrence months before conventional methods.

Investigators at Mass General Brigham Cancer Institute (Boston, MA, USA) developed HPV-DeepSeek, an ultrasensitive blood assay designed to stratify postoperative risk in HPV-associated head and neck cancer. The technology detects tiny fragments of viral DNA released into the bloodstream as tumor cells grow and die. By directly measuring circulating tumor HPV DNA, the assay aims to identify patients with persistent disease after surgery who may benefit most from additional treatment.

The Clear-HPVca study followed 103 people with newly diagnosed, untreated HPV-associated head and neck cancer scheduled for surgery at Mass General Brigham Cancer Institute between August 2020 and March 2024. Researchers collected and analyzed 560 blood samples before surgery, after surgery, and during surveillance, tracking patients for more than two years. Overall, 73% received adjuvant radiation or chemoradiation after surgery.

At diagnosis, HPV-DeepSeek detected circulating tumor HPV DNA in 98.1% of patients. In the key postoperative window, 23% had detectable circulating tumor HPV DNA, and positivity was linked to worse outcomes: two-year disease-free survival was 60% among positive patients versus 100% among negative patients, and overall survival was 73% versus 98% by the end of the trial. Compared with an existing HPV blood test, HPV-DeepSeek was substantially more sensitive and identified recurrence a median of about seven months before clinical detection, with lead times up to 17.5 months and nearly double the advance warning of existing tests.

The study was observational and conducted within a single health care system, with a limited number of relapse or death events. The authors stated that larger, multi-site clinical trials are planned to assess treatment selection after surgery. The findings were published in Science Translational Medicine on June 10, 2026.

“What excites me most about this approach is the possibility of moving beyond generalized risk estimates and toward truly personalized care. If we can accurately identify residual disease at the molecular level, we may eventually be able to make treatment decisions based on the biology of an individual patient’s cancer rather than broad clinical categories,” said Daniel Faden, MD, Director of the Head and Neck Cancer Genomics and Liquid Biopsy Program at Mass General Brigham Cancer Institute.

Related Links
Mass General Brigham Cancer Institute