U.S. Patent 10,614,285 addresses techniques for efficient learning on millions of features from whole slide images
Proscia, a leading provider of AI-enabled digital pathology software, has been granted U.S. Patent 10,614,285 by the United States Patent and Treasury Office (USPTO). The patent relates to techniques that are essential for effectively developing AI in pathology by enabling the scalable storage, management, and retrieval of analysis results derived from whole slide images.
The standard of care for diagnosing cancer is based on the pathologist’s assessment of patterns in tissue biopsies under a microscope. This 150-year-old manual and subjective practice cannot keep pace with the rising cancer burden, especially as the pathologist population continues to decline. Laboratories are quickly adopting digital pathology, which captures high-resolution whole slide images of tissue specimen, to overcome many of these systemic challenges. They are also increasingly implementing AI to realize even more value from their digital data, driving productivity and quality gains and expanding the breadth of critical diagnostic information available to improve patient outcomes.
Developing AI in pathology requires massive amounts of data to account for the complexity of tissue specimen as well as the high level of variability seen in practice. To effectively identify patterns, AI applications capture and learn from the millions of features and statistics that are extracted from each whole slide image. Large volumes of data must be stored, managed, and efficiently retrieved in this process, often requiring significant investments in infrastructure that can impede AI research and delay its adoption in clinical settings.
U.S. Patent 10,614,285 covers techniques related to a computer-implemented system of analyzing whole slide images and storing features and statistics derived from that analysis in a database. Systems as disclosed by the patent can make use of cloud-based object storage services, such as Amazon S3, in a novel manner to minimize infrastructure costs while maximizing performance for large repositories of data. Such systems enable scalable and efficient data collection, management, and retrieval, which are essential for advancing the research and development of AI in pathology.
The patent reflects Proscia’s commitment to transforming the way the world practice pathology. The company’s Concentriq digital pathology platform, which serves as a launchpad for AI, incorporates the techniques covered by the patent for training Proscia’s pipeline of speciality-specific AI applications. These applications can drive meaningful efficiency, productivity, and quality gains.
U.S. Patent No. 10,614,285 is the fourth granted to Proscia in the past year. The company was granted U.S. Patent 10,496,742, which covers a computer-implemented method of evaluating medical images, in January 2020. In October 2019, Proscia was issued U.S. Patent No. 10,460,150, which relates to AI-powered dermatopathology. The company also received U.S. Patent No. 10,346,980 addressing techniques for processing and analyzing medical images critical to the advancement of digital pathology in September 2019.