A team from Northeastern University has developed a groundbreaking AI system utilizing network technology, achieving an impressive detection accuracy of 99.72% in breast cancer diagnosis. Their findings are detailed in the journal "Cancer."
This innovative system features an integrated deep learning model that enhances diagnostic precision by combining various methodologies to reduce errors. The model was trained on publicly available databases, including the Breast Cancer Histopathology Database (BCHD), which comprises images of both malignant and benign breast tissues.
The AI system efficiently analyzes high-resolution images and leverages historical data to accurately identify and diagnose cancer. Notably, it demonstrates near-perfect tumor detection in biopsy samples and consistently delivers reliable results across multiple diagnoses without experiencing fatigue. This advancement aims to enable healthcare professionals to identify patients more quickly and accurately while laying the groundwork for new AI models tailored to diagnosing rare cancers that currently lack sufficient patient data. Earlier this year, the team also introduced a fast and precise AI tool for prostate cancer detection.
Ultimately, the researchers aspire to create a comprehensive system that empowers healthcare practitioners to leverage cutting-edge AI technologies for a broad spectrum of cancer diagnoses, potentially revolutionizing the field of digital pathology.