Hyderabad Researchers Pioneer AI System for Mammogram Breast Cancer Detection
Hyderabad scientists create an AI model achieving 95% accuracy in identifying breast cancer from mammograms, offering efficient results with minimal computing resources.
Researchers at the Koneru Lakshmaiah Education Foundation in Hyderabad have unveiled an artificial intelligence-based system that significantly enhances the detection of breast cancer through mammograms. The study, authored by Kumari Jelli and Pavan Kumar Pagadala, was published in the Scientific Reports of Nature Portfolio.
Innovative AI Model with High Accuracy
The AI model, known as Fuzzy RS-Net, demonstrated a remarkable 94.9% accuracy, 95.8% sensitivity, and 93.8% specificity when tested on the Curated Breast Imaging Subset of the Digital Database for Screening Mammography. These metrics indicate the model's proficiency in accurately identifying both cancerous and non-cancerous cases.
Designed to operate with reduced computing power and memory, the Fuzzy RS-Net model employs a fuzzy logic approach to handle uncertainties in mammogram images, which often present challenges in distinguishing subtle abnormalities.
Performance Across Multiple Datasets
The researchers extended their tests to other public mammography datasets, maintaining the model's high performance across different datasets. This adaptability underscores its potential applicability in diverse clinical settings.
The model's architecture includes mechanisms for noise reduction in mammogram images, spotlighting areas of interest for further analysis. The system's components were individually assessed, each contributing to its overall efficacy, with statistical tests confirming significant improvements over existing models.
Future Validation and Clinical Trials
Despite promising results, the researchers emphasized that the system has only been validated on publicly available databases. They advocate for further testing with larger, more varied patient datasets to ensure the model's robustness in real-world scenarios.
Moreover, they recommended hospital-based trials to evaluate its clinical utility and suggested the integration of explainable AI tools to enhance interpretability for healthcare professionals. This step is crucial for gaining trust and facilitating adoption in medical environments.
According to the World Health Organization, over 2.3 million women were diagnosed with breast cancer in 2020, resulting in more than 685,000 deaths. The researchers highlighted the importance of early detection, as manual examination of mammograms is both time-consuming and prone to oversight of subtle indicators.