Work

Deep learning based analysis of factors associated with breast and cervical cancer using matched case control data

Year: 2026
Type: article
Study design: Observational Study
Abstract: Breast and cervical cancers remain leading causes of morbidity among women worldwide, with a disproportionate burden in low- and middle-income countries. Although numerous environmental, behavioral, and reproductive risk factors have been identified, their complex interrelationships are insufficiently explored using conventional statistical approaches. Advanced analytical methods may better capture multidimensional exposure patterns relevant to public health prevention. We conducted a case–control study including 632 women (confirmed breast or cervical cancer cases and age-matched controls). Standardized data collection assessed environmental exposures, lifestyle behaviors, and reproductive history. To address high-dimensional data and non-linear correlations, we implemented a deep learning framework using an autoencoder neural network for latent feature extraction and dimensionality reduction. This approach enabled the identification of complex interaction patterns beyond traditional regression models. Established environmental and behavioral factors were significantly associated with cancer status. The autoencoder-based model improved discrimination between cases and controls and revealed synergistic interaction effects between selected lifestyle and environmental exposures. These combined effects substantially strengthened predictive performance compared with models based solely on individual risk factors. Integrating Deep Learning techniques into epidemiological research provides added value for understanding the multifactorial etiology of women’s cancers. Identifying latent interaction patterns supports more refined risk stratification and informs targeted prevention strategies. This approach contributes to advancing precision public health initiatives aimed at improving women’s health outcomes and optimizing resource allocation.
Language: English
Subfield: Oncology
Field: Medicine
Open Access status: diamond
OpenAlex ID: W7237218975
FWCI:
Cites: 16
Cited by:
Related to: 10