HPV/Cervical Cancer Screening: Comparative Effectiveness of Primary HPV-Based Testing versus Cytology

Abstract

Background: Cervical cancer is a major preventable malignancy and in 2022, there were about 662,301 new cases and 348,874 deaths due to
cervical cancer worldwide. In more than 99% of cases, the necessary cause is high-risk human papillomavirus (hrHPV). Cytology (Pap smear)
screening has long been used and has proven effective, but its sensitivity has been identified as a limitation. HPV DNA-based primary screening
has become a molecularly superior tool; however, synthesized evidence from randomized trials based on detection of CIN2+ data is limited
over the last 6 years (2022-2026).
Objectives: To systematically review and metaanalyse comparative performance of HPV-based primary screening compared to cervical cytology
for detection of cervical intraepithelial neoplasia grade 2 or worse (CIN2+); to appraise HPV vaccination efficacy data; to evaluate self-sampling
strategies for under-screened populations; to evaluate emerging technologies such as AI-assisted colposcopy.
Methods: A systematic review and meta-analysis of peer-reviewed randomized controlled trials, systematic reviews and meta-analyses were carried
out between January 2022 and June 2026, in accordance with the guidelines of the PRISMA 2020 and the Cochrane guidelines. Databases
explored were PubMed/MEDLINE, Embase, CENTRAL, Scopus and Web of Science. Primary outcome: detection rate for CIN2+ (RR, 95%
CI). Secondary outcomes were HPV vaccination effectiveness, self-sampling uptake, test sensitivity/specificity, and cancer burden measures at
the population level.
Results: Across 8 RCTs encompassing 414,846 participants, HPV DNA-based screening detected CIN2+ lesions at a rate 61% higher than
cytology (RR 1.61; 95% CI: 1.30–1.98; p < 0.00001). HPV mRNA testing was found to be equally sensitive (93.2% for CIN2+) but was more
specific (84.0%) than HPV DNA testing (80.8%). Of the 145 RCTs included in the meta-analysis, HPV vaccination was found to be effective
in decreasing the risk of developing CIN I by 85%, CIN II by 80%, and persistent HPV16/18 infection by 84%. The self-sampling strategies
resulted in 2.1–3.1-fold higher screening participation rates than standard care. The risk of cervical cancer was 2.78-fold higher for vulnerable
populations (RR 2.78; 95% CI: 2.32–3.32). The deep learning colposcopy models had an AUC-ROC of 95.3%.
Conclusion: HPV-based primary screening is shown to be very clearly superior in terms of sensitivity for CIN2+ detection to cytology. The
WHO 90-70-90 elimination framework requires integrated vaccination, risk-stratified genotyping triage and self-sampling. AI-powered tools are
a game-changer in resource-constrained environments. There is an urgent need to address the disparities between high and low-HDI countries.