Novel screening and triage approach to detect cervical precancer with HPV extended genotyping and AI

Novel screening and triage approach to detect cervical precancer with HPV extended genotyping and AI

🎙 Epicentre - MSF 👥 737 📅 June 28, 2026 ⏱ 15 min 👁 28 📄 original study 🧭 2026-08-16
Available in: English (current) Français

Keywords

HPVcervical precancerscreeningtriageAI

Summary

The presentation by Epicentre - MSF describes a novel screening and triage approach for cervical precancer in low-resource settings, developed by the PAVE Consortium. The approach combines self-sampling for HPV testing with extended genotyping, which stratifies HPV types by cancer risk, and an AI-based evaluation of cervical images obtained via mobile colposcopy. The strategy aims to improve access to screening, as less than 10% of women in Africa have ever been screened. The project enrolled 50,000 women across nine countries, including 1,400 living with HIV. Results showed that combining HPV genotyping and AI image analysis can effectively stratify women into risk categories: immediate treatment, follow-up, or reassurance. The approach targets 98% of precancers and cancers, with only 2% missed. The presentation highlights the potential of this one-visit strategy to reduce cervical cancer burden in resource-limited settings, aligning with WHO’s elimination goals.

141 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value of the information is high, as it presents results from a large-scale validation study with robust methodology, including histology confirmation for all HPV-positive women. The argumentation is logical and well-structured, explaining the natural history of HPV and cervical cancer, the rationale for the screening approach, and the results. The speaker effectively demonstrates how the combination of HPV genotyping and AI image analysis improves risk stratification compared to HPV testing alone. The presentation is persuasive, with clear evidence of the approach’s potential to reduce overtreatment and focus resources on those most at risk.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is generally high, given the large sample size and multi-country design. However, the presentation lacks detailed methodology and does not cite specific peer-reviewed publications. The speaker mentions the PAVE Consortium and contributions from MSF, but no external sources are provided. The title accurately reflects the content. The talk is a conference presentation, so it is not a formal publication, but the data presented appear credible. The adequacy between title and content is excellent.

185 words

Title / Content Match

The title accurately reflects the content, which focuses on a novel screening and triage approach combining HPV extended genotyping and AI for cervical precancer detection.

Quality & Reliability

8/10

The presentation is based on a large-scale validation study (PAVE Consortium) with over 50,000 women across multiple countries, including histology confirmation for all HPV-positive women. The speaker is a researcher from Epicentre/MSF, providing credibility. However, the talk is a conference presentation without detailed methodology or peer-reviewed publication reference, and some data slides were outdated.

Key Moments

Cited Sources

  • PAVE Consortium — Mentioned as the consortium conducting the study.

Concurring Sources

Contribution & Novelties

The presentation provides a novel contribution by demonstrating the feasibility and effectiveness of combining HPV extended genotyping with AI-based image analysis for cervical precancer screening in low-resource settings. This approach allows for risk stratification and triage in a single visit, potentially reducing overtreatment and improving resource allocation. The study’s large scale and inclusion of HIV-positive women add to its significance.

Pour aller plus loin :

121 words

Radar Profile

The radar profile shows high scores in quantity and quality of information, with a moderate level of technical depth. The fiabilite_globale is also high, reflecting the robust study design. The overall profile indicates a scientifically sound presentation with strong evidence.

Reliability 8/10