Digitalisierung und künstliche Intelligenz in der Radioonkologie

Neuronal networks enable flexible and adaptable information processing. This renders them particularly suitable for complex tasks such as image recognition or prognostic modeling. Via application of these concepts, artificial intelligence (AI) can contribute to more precise diagnoses, creation of individualized treatment plans, and improvement of therapeutic efficiency. The use of AI in diagnostic imaging to detect lung nodules is a promising approach, as is its application in lung cancer screening. Alongside prediction of the probability of adverse events, AI is also of interest for prediction of the response to radiotherapy and during follow-up. For example, during individualized treatment follow-up of patients with lung cancer, use of a web-based application with an integrated risk evaluation algorithm based on weekly recording of symptoms by the patients was shown to be superior to non-individualized follow-up, as longer overall survival was achieved due to early detection of recurrence and the overall clinical condition of patients was improved.