Leveraging Digital Pathology to Enhance Breast Cancer Diagnosis in Low-Resource Settings: A Cross-Sectional Study of Major Histopathology Laboratories in Uganda

ABSTRACT Background Breast cancer remains a leading cause of morbidity and mortality among Ugandan women, particularly in rural areas where pathology infrastructure and specialist access are limited. This study aimed to improve diagnostic efficiency and accuracy by digitizing breast cancer slides and implementing a web-based histopathology system to support remote diagnostics. Methods A total of 100 glass slides from core breast biopsy samples were digitized using an Aperio AT2 slide scanner. A web-based Breast Histology Information Management System (BHIMS) was developed to store and provide remote access to these digital slides. Two experienced pathologists independently evaluated the slides using both conventional microscopy and digital platforms. Assessments focused on key morphological features, including cancer diagnosis, type, grade, mitosis, lymphatic invasion, nuclear pleomorphism, and tubule formation. Inter- and intra-observer agreement was analyzed using kappa statistics. Results Digital pathology showed high inter- and intra-observer reliability, particularly for cancer diagnosis and type. Slightly lower agreement was noted for nuanced features such as mitosis and lymphatic invasion. Conventional microscopy yielded stronger observer agreement and confidence, especially for mitotic count and invasion. Conclusion Digital pathology offers a promising approach to improve breast cancer diagnosis in resource-limited settings. The BHIMS system enabled remote access and collaborative review of cases, supporting decentralized pathology services. However, broader adoption will require investment in high-resolution scanners, reliable internet, data storage, and training of pathologists. With these enhancements, digital pathology could significantly strengthen cancer care delivery in rural Uganda.