Angiogenesis and Artificial Intelligence, a long overdue match

Recognizing patterns – a key feature of human intelligence and imperative for our evolution – now perfected by computers. 

Over the last decades plenty of imaging techniques emerged that allow researchers to investigate microstructures more thoroughly on smaller scales, over time or in multiple layers. These imaging techniques were often supported by specific staining techniques to highlight the microscopic regions of interest. Quantifying morphological features, however, was and still is often achieved only semi-quantitatively by analyzing contrast or fluorescence units of an image and remained somewhat elusive and time-consuming. The lack of reproducible morphological quantification techniques may also have led researchers to execute excessive staining and the use of advanced imaging techniques where phase contrast microscopy might have been a sufficient and inexpensive alternative.

A prime example of complex morphological structures are vascular networks found in mammals. These are not only inherent to healthy but also cancerous tissue, and need to be accurately quantified in order to fully understand the process of angiogenesis. To facilitate this venture, various in-vitro assays exist to model such networks in a laboratory setting in efforts to study endothelial cell biology and ultimately combat cancer.

By employing the latest Deep Learning applications researchers now are also able to easily and accurately quantify such morphological structures and are freed up to focus on advancing science.

Do you want to know why this is only possible now?
Read more in KML Vision’s recent blog