Bc-Deeds

bladder cancer has a high recurrence and progression rate. in order to reduce recurrence rate, knowledge of the tumor grade is essential to perform resection to the correct depth of the layers in the urinary bladder during primary resection of the tumor. however, visual estimation by the surgeon regarding tumor grade is very poorly correlated to the true tumor grade. this is the main reason for the high recurrence rate and delay of relevant adjuvant treatment. using deep-learning, we can predict tumor grade during the initial removal of the tumor, offering personalized treatment and thereby reduce recurrence rates. the project is funded by the innovation fund denmark under the innoexplorer program. based on our initial proof of concept cnn-model, the grant covers further development of a more complex cnn-model and the development of a prototype for clinical testing.