Evaluation of screening algorithms based on HPV testing with partial genotyping for the prevention of cervical cancer among HIV-infected women in resource-limited countries: results of the ANRS 12375 study

Title

Presenter

Pierre Debeaudrap

Authors

P. Debeaudrap * (1), F. Kabore (2), L. Setha (3), J. Tegbe (4), B. Doukoure (5), O. Segeral (6), A. Jaquet (7), P. Petignat (8), G. Clifford (9), A. Horo (4)

Institutions

(1) Institut de Recherche pour le Développement, CEPED, Paris, France, (2) Centre Muraz, Bobo Dioulasso, Burkina Faso, (3) Calmette Hospital, Phnom Penh, Cambodia, (4) Programme PACCI, Abidjan, Cote D''Ivoire, (5) Université Felix Houphouet Boigny, Abidjan, Cote D''Ivoire, (6) ANRS site in Cambodia, Phnom Penh, Cambodia, (7) ISPED, Bordeaux, France, (8) Hôpitaux Universitaire de Genève, Geneva, Switzerland, (9) International Agency of Research on Cancer, Lyon, France

BACKGROUND: In resource-limited countries, cervical cancer (CC) is the leading cause of cancer death among women living with HIV (WLHIV). The WHO recommended the use of HPV testing for primary CC screening because of its high sensitivity as compared to other screening methods. However, triage is desirable to identify HPV+ women having cervical intraepithelial neoplasia grade 2 or worse (CIN2+) and requiring treatment. In resource-limited contexts, Visual inspection with acetic acid and/or lugol (VIA/VILI) and partial genotyping are both feasible in a single-visit approach (screen-and-treat).
The ANRS-12375 study aimed to assess the performance (sensitivity, specificity), feasibility and benefit of different triaging options to detect CIN2+ lesions: partial genotyping (16/18/45), VIA/VILI alone and VIA/VILI combined with partial genotyping.
METHODS: From Nov 2019 to Dec 2021, 2,255 WLHIV aged between 30 and 49 recruited in Abidjan (Côte d''Ivoire [CI], n = 1500), Bobo-Dioulasso (Burkina Faso [BF], n = 422) and Phnom Penh (Cambodia [KH], n = 333) were primarily screened with an Xpert HPV test, followed by VIA/VILI when positive and treatment if required. Relaxed criteria (or ABCD-criteria for Acetowhiteness, Bleeding, Coloring, and Diameter) were used to identify CIN2+ lesion suring the VIA/VILI. The performance of the triage strategies to identify CIN2+ lesions were evaluated considering histology as standard reference.
RESULTS: High risk-HPV infection was reported in 881 (41%) participants (median age: XX) with significant differences between sites (CI: 47%, BF: 33%, KH: 21%; P<xxx). HPV 16 accounted for 18% of infections and HPV 16/18/45 for 31%.
Among the 533 (60%) HPV+ participants with available histology (all data will be available by July), 15% had CIN2+ lesions, 71% were VIA+ and 31% HPV16/18/45+.
VIA sensitivity , partial HPV 16/18/45 genotyping and their combination as triage were 88% [79-94], 53% [42-63] and 91% [83-96], respectively, while the specificities were 33% [37-28], 73% [77-69], and 25% [36-27] (inter-sites heterogeneity: p=0.04).
CONCLUSIONS: Visual inspection after a positive HPV result is a good triage option with high sensitivity to identify women needing treatment even if inter-sites heterogeneity suggests differences in practice. While partial genotyping showed insufficient sensitivity, an algorithm combining partial genotyping and VIA/VILI appears to be the most efficient.

Download the e-Poster (PDF)