Scientists have trained neural networks to recognize almost 200 types of cancer.

German scientists from the Scharite medical clinic in Berlin have presented a new system for diagnosing brain tumors that accurately determines the type of tumor based on epigenetic markers. This scientific breakthrough will make diagnosis more accessible and effective, as well as allow for faster treatment initiation. The crossNN model successfully identifies over 170 types of tumors from various organs with an accuracy of 97.8%, opening new opportunities for combating cancer and other oncological diseases. The next step for the scientists will be to conduct large-scale clinical trials of the technology in oncological centers in Germany. Conclusion: the created artificial intelligence system opens new perspectives in the diagnosis and treatment of oncological diseases, allowing for quicker and more effective responses to them.