Association of Plasma Circulating Tumor HPV DNA With HPV-Related Oropharynx Cancer

Abstract

This cross-sectional analysis examines whether circulating tumor human papillomavirus DNA is detectable in the plasma of healthy participants at risk for oral human papillomavirus infection and the agreement with oral rinse and serum biomarkers.

Early detection of human papillomavirus (HPV)-related oropharyngeal squamous cell carcinoma (HPV-OPSCC) is currently limited owing in part to a lack of effective screening tools. Recent research suggests possible utility for circulating tumor HPV DNA (ctHPVDNA) in oral rinses and plasma.^1,2 However, it is unclear whether plasma ctHPVDNA is detectable in a healthy population and if ctHPVDNA is concordant with other detectable HPV biomarkers including oral rinse and serum antibodies.

Hence, we aimed to understand whether ctHPVDNA is detectable in plasma of healthy participants at risk for oral HPV infection and the agreement with oral rinse and serum biomarkers.

Methods

This cross-sectional analysis included healthy people without HPV-OPSCC, but at increased risk of oral HPV infection. Participants were eligible if they reported 2 or more lifetime oral sex partners, history of anogenital dysplasia/cancer, or were a partner of someone with an HPV-related cancer. Participants enrolled through Johns Hopkins Hospital (Baltimore, Maryland, and other Maryland sites) and Mount Sinai Hospital (New York, New York) and provided plasma, serum, and oral rinse samples. The study was approved by the institutional review board of the participating centers, and written infomed consent was obtained from all participants.

Participant DNA was extracted using the NucliSENS easyMAG (input 500 μL plasma or 750 μL of 2.5 × diluted oral rinse sample) and tested for HPV DNA by polymerase chain reaction (PCR) amplification using the SPF10 system (version 1). All samples with positive results for HPV DNA were genotyped using SPF10 LiPA for 25 HPV types including oncogenic types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 66. Serum samples were tested for E6 antibodies to HPV 16, 18, 31, 33, 35, 45, 52, and 58 by the German Cancer Research Center (DKFZ) classification system using multiplex serology based on glutathione S-transferase capture immunoassay.³

Prevalence estimates were calculated. Concordance across biomarkers was explored by cross-tabulation and calculation of κ.

Results

Participants (n = 408; Table 1) had a median (IQR) age of 56 (46-64) years and were predominantly White (72%) men (90%). Participants were considered at risk for oral HPV based on being a man aged 30 years or older with 2 or more oral sex partners⁴ (85%), having a partner with HPV-OPSCC (7%), having a partner with history of anogenital cancer or dysplasia (2%), or themselves having a history of anogenital cancer or dysplasia (6%).

Table 1. Participant Characteristics at Baseline.

Characteristic	No. (%)
No.	408
Sex
Male	360 (90)
Female	40 (10)
Age, y
29-49	125 (31)
50-64	176 (44)
65-82	99 (24)
Race
Black	87 (21)
White	293 (72)
Othera	28 (7)
Study site
Johns Hopkins	371 (91)
Mount Sinai	37 (9)
Participant risk factors
History of anogenital cancer/dysplasia	23 (6)
Partners of someone with HPV-OPSCC	30 (7)
Partners of someone with anogenital cancer/dysplasia	10 (2)
Male aged >30 y and ≥2 oral sex partners	345 (85)

Abbreviations: HPV, human papillomavirus; OPSCC, oropharyngeal cancer.

^a Other races and ethnicities reported were Asian, American Indian or Alaska Native, Native Hawaiian or Pacific Islander, multirace, or any other.

Oncogenic ctHPVDNA was undetectable in plasma of all 408 participants. Although 20 participants had oncogenic oral ctHPVDNA and 43 were E6 seropositive to an oncogenic type, none of these people had ctHPVDNA detected in their plasma (Table 2). Paired comparisons of ctHPVDNA in plasma and oral rinse, as well as plasma ctHPVDNA and serum E6 antibodies yielded a κ statistic of 0. This implies agreement of biomarkers was no better than chance alone. In addition, oral ctHPVDNA and E6 antibodies had poor agreement (κ = 0.03), only 3 people had positive results for both.

Table 2. Prevalence of Circulating Oncogenic HPV DNA in Plasma, Among Participants With and Without Other HPV Biomarkers.

	No. positive (N = 408)	ctHPVDNA detected in plasma
Oral HPV DNA
Any oncogenic HPV DNA	20	0
Oral HPV16 DNA	5	0
E6 serum antibodies
Any oncogenic HPV E6 Ab	43	0
Serum HPV16 E6 Ab	8	0

Abbreviations: HPV, human papillomavirus; ctHPVDNA, circulating tumor HPV DNA.

Discussion

A good biomarker for a malignant disease should not be detectable in individuals without the disease, a characteristic that makes it specific. Previous specificity estimates for ctHPVDNA in plasma of patients with HPV-OPSCC were discrepant and derived from small groups of controls.^2,5,6 Our analysis focused on healthy participants enrolled based on eligibility criteria that have been shown to classify them as high-risk for oral HPV infection.⁴ In a large cohort of healthy people without cancer, plasma ctHPVDNA had 100% specificity. This analysis provides clarity for whether detectable ctHPVDNA is present in the plasma of healthy controls without cancer, with or at risk for oral HPV. Moreover, it supports ctHPVDNA in plasma as a potential biomarker in screening or surveillance given better specificity compared with other validated OPSCC biomarkers (oral ctHPVDNA and E6 antibodies). Detection of plasma ctHPVDNA among HPV-OPSCC, including those with localized disease,² suggests high positive predictive value, but how long before HPV-OPSCC plasma ctHPVDNA is detectable is unknown and will inform potential screening benefit. Limitations of this study include use of PCR rather than ddPCR or next generation sequencing, which may have improved performance properties, and hospital-based enrollment, which may not be representative of the general population.