Transforming Breast and Cervical Cancer Screening Data into the OMOP CDM: Integrating Clinical and Genomic Insights from Senegal

Background: Breast and cervical cancers remain major health burdens in low- and middle-income countries (LMICs), where screening coverage is limited and late-stage diagnoses are common. In Senegal, approximately 2,000 new cervical cancer cases and nearly 1,800 breast cancer cases were reported in 2022. Although community-based screening campaigns generate large volumes of clinical data, these are rarely exploited for research due to a lack of standardization. The OMOP Common Data Model (CDM), developed by the OHDSI community, provides a framework to harmonize such data and enable reproducible, collaborative research. Methods: This study analyzed data from 491 women who participated in a community-based breast and cervical cancer screening campaign jointly organized by IRESSEF and Diamniadio Children's Hospital (Dakar, Senegal). The dataset included demographics, clinical outcomes, HPV PCR results, lesion status, comorbidities, and metagenomic profiles of the vaginal microbiome and virome. Data transformation followed the OHDSI ETL framework using White Rabbit (profiling), Rabbit-in-a-Hat (ETL design), and Usagi/Athena (concept mapping), with implementation in PostgreSQL. Variables originally collected in French were translated into English for OMOP compatibility, and data were mapped to core OMOP domains (person, observation, condition_occurrence, measurement, procedure_occurrence). Microbial and viral abundances were captured as quantitative measurements, representing one of the first attempts to integrate multi-omic data into OMOP CDM in West Africa. Results: The transformation achieved approximately 80% overall mapping to OMOP domains, with full coverage of demographic variables, 85% of clinical conditions, and 75% of laboratory and measurement data. Some specific variables, such as PCR-specific tests and genomic data, required manual curation due to missing OMOP equivalents. The study population ranged in age from 14 to 71 years (mean 34), with frequent comorbidities including diabetes and hypertension. Integration of microbiome and genomic data revealed distinct patterns: Gardnerella vaginalis, G. piotii, and G. swidsinskii were enriched among women with precancerous lesions and positive HPV tests, while Lactobacillus variants were more common in women without lesions, suggesting a potential protective role. Oncogenic HPV types (e.g., Alphapapillomavirus 7) and co-infections such as Human gammaherpesvirus 4 were also detected, demonstrating OMOP CDM's capacity to support integrative multi-omic analyses. Conclusions: This work represents the first implementation in West Africa that integrates clinical, demographic, and genomic cancer screening data into the OMOP CDM. By aligning local data with international standards, it lays the foundation for reproducible and comparable analyses across settings and fosters collaborative research. Early microbiome insights point to potential biomarkers warranting further investigation. By sharing ETL scripts and mappings, this work serves as a reusable template for other African institutions adopting OMOP CDM.