Machine Learning is currently being used in many fields and it is therefore important to investigate the model predictions in terms of stability and quality. The still relatively new topic of distribution shifts is playing an increasingly important role. The work deals with exactly these shifts on tumor images originating from the Camelyon17 Challenge organized from the Diagnostic Image Analysis Group (DIAG) and the Department of Pathology of the Radboud University Medical Center (Radboudumc). Examined are different shifts and their impact on the accuracy and performance of a Machine Learning model in terms of performance drops and informative value.