DL-Based Banding Classification with Ikaros

With this advance notice, MetaSystems proudly announces the launch of its first product utilizing Deep Neural Networks (DNN) for image analysis. In one of the next versions of the outstanding karyotyping system Ikaros, artificial intelligence (AI) will be used to assign chromosomes automatically to their respective classes. The new algorithms have been trained and tested based on thousands of pre-classified G-banded chromosomes which have been kindly provided by the cytogenetics experts from MLL (Münchner Leukämielabor GmbH, Munich, Germany).

“Our lab had the opportunity to be the first user to test the beta-version of MetaSystems’ new AI-based karyotyping software Ikaros. After fully automated metaphase finding and image acquisition - established many years ago - AI-based chromosome classification is the next quantum leap in optimizing chromosome analysis. We experienced a time gain of up to 50% in the karyotype analysis of bone marrow metaphases. This enormous gain in efficiency allows us to keep pace with the ever-increasing workload in times of shortage of personnel resources.”

(Prof. Claudia Haferlach, MD from MLL)

Ikaros with its new, improved automated chromosome classification based on Deep Learning will only be the first out of many AI-based products in MetaSystems’ pipeline.

“The enormous potential of the new DNN technology is obvious”, Andreas Plesch, Managing Director of MetaSystems said. “To fully leverage this potential the integration of AI-based analysis with high-throughput imaging is crucial. And this is exactly what our Deep Learning-enhanced slide scanning platform Metafer will provide to our customers: user friendly and highly efficient workflows that achieve maximum results with a minimum of operator interaction.”