AI Breakthrough: New Tool Detects Colorectal Cancer with 96.74% Accuracy
Jyväskylä, Finland: A groundbreaking AI-based tool for detecting colorectal cancer from tissue samples has been developed at the University of Jyväskylä, surpassing previous models in classification accuracy. The neural network model, achieving an impressive 96.74% accuracy, has been made freely available to advance cancer research and diagnostics.
In collaboration with the University of Turku’s Institute of Biomedicine, the University of Helsinki, and Nova Hospital of Central Finland, researchers have designed an artificial intelligence tool capable of automating the analysis of colorectal cancer tissue slides. The AI model outperformed all previous models in identifying tissue categories relevant to cancer detection.
“Our model accurately classifies all tissue types crucial for cancer identification, achieving a 96.74% accuracy rate,” said Fabi Prezja, a doctoral researcher at the University of Jyväskylä.
Traditionally, pathologists manually examine digital microscopy slides of tissue samples, marking cancerous and related areas. The AI tool streamlines this process by analyzing the slides and highlighting different tissue categories, potentially reducing histopathologists’ workload and speeding up diagnoses.
The research team has made the tool publicly available to encourage global collaboration in cancer research. “By providing free access, we hope to inspire scientists, developers, and researchers worldwide to enhance the tool and explore new applications,” Prezja added.
Despite its promising results, researchers emphasize that integrating AI tools into clinical practice requires a cautious, step-by-step approach. Before becoming a routine diagnostic tool, the AI model must undergo rigorous validation to ensure it meets clinical and regulatory standards.
This breakthrough marks a significant step in the fight against colorectal cancer, paving the way for faster and more efficient cancer detection using artificial intelligence.
