Summary: Researchers in the UCLA Department of Electrical and Computer Engineering have developed a cost-effective digital scanning microscope to improve pathological screening.
Background: High quality pathological imaging systems are essential to carefully assess tissue sectioning samples and produce accurate diagnoses for patients. Digitization of these systems has greatly improved throughput of pathology laboratories, and the development of new machine learning algorithms has achieved diagnostic accuracies similar to that of trained specialists. However, limitations in state-of-the-art digital pathology systems persist. The speed of conventional pathology microscopes remains limited by mechanical and optical complexities. Further, these systems can be prohibitively expensive (>$200k) for resource-limited clinics. To improve clinical and diagnostic outcomes, there is an urgent need to develop robust, inexpensive pathology equipment that can provide continuous operation and unencumbered diagnostic accuracy.
Innovation: Researchers led by Aydogan Ozcan have developed an AI-powered, cost-effective, and compact digital scanning optical microscope, termed BlurryScope. BlurryScope is a cost-effective alternative to traditional pathology scanners. The system continuously scans through tissue sections at 5,000 µm/s, algorithmically stitches the images back together, then interprets the recombined image. The research team has demonstrated that this platform can be used to efficiently classify HER2 scores for breast cancer prognosis on tissue samples with 86.2% accuracy using neural networks., This device is highly cost-effective and compact compared to industry standard platforms, costing less than $500 and weighing less than 3 kg., By leveraging AI to automate image analysis, BlurryScope offers a powerful, cost-effective, and portable tool for rapid and accurate tissue diagnosis, with the potential to revolutionize clinical care, especially in resource-limited settings.
Potential Applications: • Resource limited diagnostics • Inexpensive educational tools • Cancer screening • Digital pathology
Advantages: • Continuous imaging/high speed • Reduced cost • Small footprint and lightweight • Reduced motion artifact noise • Leverages deep learning models
Development-To-Date: The researchers have built Blurryscope and demonstrated nearly 90% accuracy using patient tissue sectioning samples.
Publications: BlurryScope: a cost-effective and compact scanning microscope for automated HER2 scoring using deep learning on blurry image data
Reference: UCLA Case No. 2025-070
Lead Inventor: Professor Aydogan Ozcan, UCLA Department of Electrical and Computer Engineering and Department of Bioengineering