Journal Article

High-risk human papillomavirus testing in first-void urine as a novel and non-invasive cervical cancer screening modality—a Danish diagnostic test accuracy study

Binderup, Karen Omann, Jensen, Jørgen Skov, Vorsters, Alex, Hammer, Anne, Blach, Christina, Tranberg, Mette, Bor, Pinar, Nørgaard, Pia, Van Keer, Severien, Gustafson, Line Winther

2025

Overview

Background First-void urine (FVU) collection for high-risk human papillomavirus (hrHPV) testing has game-changing potential to improve cervical cancer prevention among under-screened women who remain unreached by clinician-based cervical cancer screening and vaginal self-sampling. Yet, the wide variation in the clinical accuracy of hrHPV testing in urine for detecting high-grade cervical intraepithelial neoplasia (CIN2 + /CIN3 +) across studies and clinical settings highlights the importance of local piloting and validation. This study determined the relative clinical accuracy of hrHPV testing in FVU versus clinician-collected cervical samples to detect CIN2 + /CIN3 + in a Danish referral population. Methods In a diagnostic test accuracy study, paired FVU (10 mL Colli-Pee device; index test) and cervical samples (Cervex Combi brush; comparator test) were obtained from 325 women aged 23–64 years (median age 36.0 years (IQR 29–46) who were either referred for colposcopy and biopsy taking or a cervical excision (reference test; available for all participants). Samples were tested using Allplex HR HPV DNA extended genotyping assay. Same absolute cut-off for hrHPV positivity applied for cervical samples was used for FVU. Of the 325 women, 145 (44.6%), 180 (55.4%), and 138 (42.5%) were diagnosed with < CIN2, CIN2 + , and CIN3 + , respectively. Results Sensitivity to detect CIN2 + (ratio 0.97, 95% CI 0.92–1.02, p MCN  = 0.33) and CIN3 + (ratio 0.95, 95% CI 0.90–1.00, p MCN  = 0.09) using hrHPV testing in FVU samples was not significantly different to hrHPV testing in cervical samples, whereas specificity for < CIN2 (ratio 0.67, 95% CI 0.46–0.96, p MCN  = 0.04) was significantly lower in FVU than on cervical samples. Moderate to excellent hrHPV test agreements between paired samples were demonstrated (Cohen’s kappa = 0.44 to 0.88). Conclusions This is the first study proving similar CIN2 + /CIN3 + sensitivity for FVU-hrHPV testing using the 10-mL Colli-Pee device and Allplex HR HPV assay compared to testing in cervical samples. From an implementation perspective, further research is needed to gather additional clinical accuracy and acceptability data on hrHPV testing of FVU-device collection in under-screened populations to support its broader integration into screening programmes. Trial registration Clinicaltrials.gov: NCT05065853.

• Publisher: BioMed Central,Springer Nature B.V,BMC
• Subject: Accuracy / Adult / Biomedicine / Biopsy / Cancer / Cancer screening / Cellular biology / Cervical cancer / Cervical cancer screening / Clinics / Colposcopy / Denmark / Diagnostic tests / Early Detection of Cancer - methods / Female / Genotyping / HrHPV DNA testing / Human papillomavirus / Human Papillomavirus Viruses / Humans / Laboratories / Mass screening / Medical equipment / Medical screening / Medicine / Medicine & Public Health / Middle Aged / Papillomaviridae - genetics / Papillomaviridae - isolation & purification / Papillomavirus Infections - diagnosis / Papillomavirus Infections - urine / Papillomavirus Infections - virology / Pathology / Samples / Sensitivity / Sensitivity analysis / Sensitivity and Specificity / Urinary hrHPV testing / Urine / Uterine cervical dysplasia / Uterine Cervical Dysplasia - diagnosis / Uterine Cervical Dysplasia - urine / Uterine Cervical Dysplasia - virology / Uterine Cervical Neoplasms - diagnosis / Uterine Cervical Neoplasms - urine / Uterine Cervical Neoplasms - virology / Vagina / Womens health / Young Adult