Automated Visual Evaluation (AVE) to improve cervical cancer screening: A prospective observational study in Rwanda, Senegal, and Malawi

Principal investigator: Francois UWINKINDI
Check PI profile

Investigators

Francois Uwinkindi, Henry Phiri, Mamadou Diop, Marc Hagenimana, Jean Paul Balinda, Twambilire Phiri, James Kachingwe, Malick Anne, Ibou Thiam, Marie Edouard Faye Diémé, Philippe Moreira, Alfousseyni Gaye, Chérif Dial, Halimatou Diop, Fatma Guenoune, Sylvie Gaju, Blandine Bourgoin, Frehiwot Birhanu, Caroline Barrett

Study information

  • Research Area: Cancer
  • Research type: This is a two-phase prospective, observational study to evaluate the performance, feasibility, and acceptability of AVE in a cervical cancer screening program in a resource limited setting, and to collect images to make future improvements to the algorithm. Phase 1 of the study aims to enroll 19,500 women across Rwanda, Senegal, and Malawi over a period of six months. All women will receive an HPV test, VIA, and AVE. This phase of the study will confirm the performance of the AVE algorithm compared to standard naked-eye VIA. Both tests will be judged against histopathology as the reference standard; all women who screen positive on any test (HPV, VIA, or AVE) will be referred for colposcopy and biopsy (and, if appropriate, treatment). Figure 1: Phase 1 participant flow Following confirmation of the statistically superior sensitivity of the AVE algorithm compared to VIA, Phase 2 of the study will begin. The aim of Phase 2 is to monitor the implementation and clinical impact of AVE as a decision-making aid, without the Phase 1 requirement for biopsies on all screen positive women, and to gather perspectives of providers and clients to inform scale up. While the machine learning validation process to prepare the algorithm will provide data as to the algorithm’s accuracy in identifying pre-cancer / cancer cases, the results from this study in Rwanda, Senegal, and Malawi will also be essential to informing widespread implementation and scale up of AVE across other geographies. The study will provide additional evidence and learnings related to performance of the AVE algorithm directly compared to VIA; factors associated with health workers’ attitudes towards and acceptance of AVE; and factors associated with clients’ attitudes towards and acceptance of AVE. Additionally, images collected during this study will be anonymized and stored, with client permission, for future analysis and improvements to the AVE algorithm (see Section VIII – Ethical and Regulatory Aspects for further details around image ownership and privacy protections). Women between 30 and 49 years old eligible for cervical cancer screening will be included.
  • Start date: Dec 01, 2021
  • End date: Jun 30, 2022

Study aim of the objectives

This is a two-phase prospective, observational study to evaluate the performance, feasibility, and acceptability of AVE in a cervical cancer screening program in a resource limited setting, and to collect images to make future improvements to the algorithm.

Study setting/Area (Location)

Rwanda, Senegal, and Malawi. In Rwanda, the study will be conducted in all health public health facilities (DHs and HCs) in Gasabo district, City of Kigali, Rwanda

Implementing institution

Rwanda Biomedical Centre, Clinton Health Access Initiative, Ministère de la Santé et de l’Action Sociale (MSAS), Senegal, Malawi Ministry of Health

Department/Division

Non Communicable Division

Impact on Policy

The study findings will inform guideline and policies for cervical cancer diagnosis

Related research

Login
Support