Blood-based biomarkers of human papillomavirus-associated cancers: A systematic review and meta-analysis

Abstract

Background: Despite the significant societal burden of human papillomavirus (HPV)-associated cancers, clinical screening interventions for HPV-associated noncervical cancers are not available. Blood-based biomarkers may help close this gap in care.

Methods: Five databases were searched, 5687 articles were identified, and 3631 unique candidate titles and abstracts were independently reviewed by 2 authors; 702 articles underwent a full-text review. Eligibility criteria included the assessment of a blood-based biomarker within a cohort or case-control study.

Results: One hundred thirty-seven studies were included. Among all biomarkers assessed, HPV-16 E seropositivity and circulating HPV DNA were most significantly correlated with HPV-associated cancers in comparison with cancer-free controls. In most scenarios, HPV-16 E6 seropositivity varied nonsignificantly according to tumor type, specimen collection timing, and anatomic site (crude odds ratio [cOR] for p16+ or HPV+ oropharyngeal cancer [OPC], 133.10; 95% confidence interval [CI], 59.40-298.21; cOR for HPV-unspecified OPC, 25.41; 95% CI, 8.71-74.06; cOR for prediagnostic HPV-unspecified OPC, 59.00; 95% CI, 15.39-226.25; cOR for HPV-unspecified cervical cancer, 12.05; 95% CI, 3.23-44.97; cOR for HPV-unspecified anal cancer, 73.60; 95% CI, 19.68-275.33; cOR for HPV-unspecified penile cancer, 16.25; 95% CI, 2.83-93.48). Circulating HPV-16 DNA was a valid biomarker for cervical cancer (cOR, 15.72; 95% CI, 3.41-72.57). In 3 cervical cancer case-control studies, cases exhibited unique microRNA expression profiles in comparison with controls. Other assessed biomarker candidates were not valid.

Conclusions: HPV-16 E6 antibodies and circulating HPV-16 DNA are the most robustly analyzed and most promising blood-based biomarkers for HPV-associated cancers to date. Comparative validity analyses are warranted. Variations in tumor type-specific, high-risk HPV DNA prevalence according to anatomic site and world region highlight the need for biomarkers targeting more high-risk HPV types. Further investigation of blood-based microRNA expression profiling appears indicated.

Keywords: anal cancer; biomarker; blood biomarker; cancer prevention; cancer surveillance; cervical cancer; human papillomavirus (HPV); oropharyngeal cancer; penile cancer; screening.

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses diagram. CTL, cytotoxic T-lymphocyte; HLA indicates human leukocyte antigen; HPV, human papillomavirus; OPC, oropharyngeal cancer.

Figure 2.

Performance characteristics of HPV-16 E6 antibodies in patients with p16- or HPV-unspecified and specified cancer versus cancer-free controls. Patients with p16+ or HPV+ cancers had greater odds of exhibiting HPV-16 E6 antibodies compared to patients with p16- or HPV-unspecified cancers. CI indicates confidence interval; HPV, human papillomavirus.

Figure 3.

Performance characteristics ofHPV-16 E6 antibodies in serum collected up to 13.7years before diagnosis in HPV-unspecified patients versus cancer-free controls. E6 antibodies showed the strongest associations with anal and oropharyngeal cancers. CI indicates confidence interval; HPV, human papillomavirus.

Figure 4.

Performance characteristics of HPV-16 E antibodies in patients with HPV-unspecified oropharyngeal cancer versus cancer-free controls. HPV-16 E antibody cotesting showed nonsignificantly improved performance characteristics compared to HPV-16 E6 or E7 antibodies alone. The 2016 study by Anantharaman et al was not included in the pooled analysis because the study population overlapped with Anantharaman et al’s 2013 study. CI indicates confidence interval; HPV, human papillomavirus.

Figure 5.

Performance characteristics of circulating HPV-16 and HPV-18 DNA in patients with cervical cancer versus cancer-free controls. Circulating HPV-16 DNA showed nonsignificantly greater validity compared to HPV-18 DNA. CI indicates confidence interval; HPV, human papillomavirus.

Figure 6.

Performance characteristics of other candidate blood-based biomarkers in patients with cervical cancer versus cancer-free controls. None of the biomarkers listed here showed significant associations for cervical cancer. Four studies were not included in the pooled analyses because of overlapping patient populations or heterogenous data-reporting methods (eg, reporting medians or geometric means).^– CI indicates confidence interval; IFN-γ, interferon-γ; IGF-1, insulin-like growth factor 1; IGFBP-3, insulin-like growth factor binding protein 3; SD, standard deviation.