Publikation
Lung Cancer Survival Estimation Using Data from Seven German Cancer Registries
Sebastian Germer; Christiane Rudolph; Alexander Katalinic; Natalie Rath; Katharina Rausch; Heinz Handels
In: Elisavet Andrikopoulou; Parisis Gallos; Theodoros N. Arvanitis; Rosalynn Austin; Arriel Benis; Ronald Cornet; Panagiotis Chatzistergos; Alexander Dejaco; Linda Dusseljee-Peute; Alaa Mohasseb; Pantelis Natsiavas; Haythem Nakkas; Philip Scott (Hrsg.). Intelligent Health Systems – From Technology to Data and Knowledge. Medical Informatics Europe Congress (MIE-2025), Intelligent Health Systems – From Technology to Data and Knowledge, May 19-21, Glasgow, United Kingdom, Pages 457-461, Studies in Health Technology and Informatics, Vol. 327, IOS Press, 5/2025.
Zusammenfassung
Predicting the survival of cancer patients is of high importance for the medical community, e.g. for evaluating therapy strategies. This study is based on lung cancer data retrieved from seven German cancer registries according to the German basic oncology dataset. After data integration and preprocessing, we predicted the survival for 6, 12, 18 and 24 months respectively using a gradient boosting algorithm. To gain insight into the decision process of the models, we identified the features that have a high impact on patient survival using permutation feature importance scores as explainability metric. They show that age at diagnosis as well as the presence of distant metastases are key factors for long-term survival. The found factors can be used in a next step for multi-variate survival analysis.