Plasma Circulating Tumor HPV DNA for the Surveillance of Cancer Recurrence in HPV-Associated Oropharyngeal Cancer

Abstract

Purpose: Plasma circulating tumor human papillomavirus DNA (ctHPVDNA) is a sensitive and specific biomarker of human papillomavirus (HPV)-associated oropharyngeal squamous cell carcinoma (OPSCC). We investigated whether longitudinal monitoring of ctHPVDNA during post-treatment surveillance could accurately detect clinical disease recurrence.

Methods and materials: A prospective biomarker clinical trial was conducted among patients with nonmetastatic HPV-associated (p16-positive) OPSCC. All patients were treated with curative-intent chemoradiotherapy (CRT). Patients underwent a 3-month post-CRT positron emission tomography/computed tomography scan and were thereafter clinically evaluated every 2-4 months (years 1-2), then every 6 months (years 3-5). Chest imaging was performed every 6 months. Blood specimens were collected every 6-9 months for analysis of plasma ctHPVDNA using a multianalyte digital polymerase chain reaction assay. The primary endpoint was to estimate the negative predictive value (NPV) and positive predictive value (PPV) of ctHPVDNA surveillance.

Results: One hundred fifteen patients were enrolled, and 1,006 blood samples were analyzed. After a median follow-up time of 23 months (range, 6.1-54.7 months), 15 patients (13%) developed disease recurrence. Eighty-seven patients had undetectable ctHPVDNA at all post-treatment time points, and none developed recurrence (NPV, 100%; 95% CI, 96% to 100%). Twenty-eight patients developed a positive ctHPVDNA during post-treatment surveillance, 15 of whom were diagnosed with biopsy-proven recurrence. Sixteen patients had 2 consecutively positive ctHPVDNA blood tests, 15 of whom developed biopsy-proven recurrence. Two consecutively positive ctHPVDNA blood tests had a PPV of 94% (95% CI, 70% to 99%). Median lead time between ctHPVDNA positivity and biopsy-proven recurrence was 3.9 months (range, 0.37-12.9 months).

Conclusion: Detection of ctHPVDNA in two consecutive plasma samples during post-treatment surveillance has high PPV and NPV for identifying disease recurrence in patients with HPV-associated oropharyngeal cancer and may facilitate earlier initiation of salvage therapy.

Trial registration: ClinicalTrials.gov NCT02281955 NCT03077243 NCT03161821.

Figures

FIG 1.

REMARK diagram of study population and patient subsets used for study analyses. ctHPVDNA, circulating tumor human papillomavirus DNA; NPV, negative predictive value; PPV, positive predictive value.

FIG 2.

Longitudinal ctHPVDNA surveillance identifies patients at high risk of disease recurrence. (A) Eighty-seven out of 115 patients (76%) had undetectable ctHPVDNA levels at all post-treatment surveillance time points (starting at 6 months after start of CRT). (B) Twenty-eight patients had a positive ctHPVDNA test result during post-treatment surveillance. Red lines indicate patients who have been diagnosed with biopsy-proven disease recurrence. The blue shaded region indicates the interval for initial treatment with CRT and assessment of response by positron emission tomography/computed tomography, which was not included in the post-treatment surveillance period. (C) Kaplan-Meier relapse-free survival of patients with undetectable ctHPVDNA at all surveillance time points versus patients with at least one abnormal ctHPVDNA blood test. P value was calculated using a two-tailed log-rank test. ctHPVDNA, circulating tumor human papillomavirus DNA; neg, negative; pos, positive.

FIG 3.

Patients with two consecutively abnormal ctHPVDNA surveillance tests have a higher risk for disease recurrence. ctHPVDNA levels at the first abnormal surveillance time point (T1) and a subsequent time point (T2) for patients with (A) biopsy-proven disease recurrence or (B) no clinically evident disease recurrence. Kaplan-Meier estimates for (C) relapse-free survival, (D) locoregional control, and (E) distant metastasis-free survival for patients who had two consecutively abnormal ctHPVDNA surveillance tests (ctHPVDNA positive) versus those without two consecutively abnormal ctHPVDNA surveillance tests (ctHPVDNA negative). P values were calculated using a two-tailed log-rank test. ctHPVDNA, circulating tumor human papillomavirus DNA; neg, negative; pos, positive.

FIG 4.

ctHPVDNA surveillance facilitates early detection of disease recurrence. (A) Modified swimmer plot illustrating positive and negative ctHPVDNA surveillance tests in 11 patients with recurrence who had a blood sample available for analysis within 6 months prior to biopsy-proven recurrence. Day 0 is day of biopsy proven recurrence. Radiographic detection of recurrence occurred approximately 2 weeks prior to day 0. Ten out of 11 patients had an abnormal ctHPVDNA test result prior to recurrence diagnosis (shaded in red). (B) ctHPVDNA kinetics and clinical surveillance imaging results for a study patient with a 14-month interval from the first abnormal ctHPVDNA blood test result and clinical diagnosis of disease recurrence. CRT, chemoradiotherapy; CT, computed tomography; ctHPVDNA, circulating tumor human papillomavirus DNA; mets, metastases; NED, no evidence of disease; PET, positron emission tomography; rCR, radiographic complete response.

FIG A1.

ctHPVDNA kinetics after salvage therapy in patients with recurrence. ctHPVDNA kinetics or 12 patients with clinical recurrence who had blood samples analyzed after initiation of salvage therapy (indicated by arrow). Current disease status is also shown for each plot, and corresponds to residual ctHPVDNA levels. AWD, alive with disease; ctHPVDNA, circulating tumor human papillomavirus DNA; NED, no evidence of disease.