Anant Madabhushi, the F. Alex Nason professor II of biomedical engineering and director of the Center for Computational Imaging and Personalized Diagnostics, was awarded a patent for capturing cancer architecture in digital pathology images.
His patent is titled “Histogram of hosoya index (HoH) features for quantitative histomorphometry.”
The invention relates to a method for classifying a region of cancerous tissue using the Hosoya descriptor, which is used for graph complexity. The specific invention relates to the use of the Hosoya index for describing graph complexity associated with nuclear graphs from routine H&E pathology images. The higher the Hosoya index, the higher the graph complexity and hence the more complex the nuclear architecture and potentially the worse the disease prognosis. The Madabhushi team has shown that the Hosoya index can help discriminate between more and less aggressive early stage, estrogen receptor positive breast cancers using digital pathology images.
Co-inventors include Ajay Basavanhally, scientist at Inspirata Inc., and Sahirzeeshan Ali, a graduate student.