The primary objective of the RACOON-PDAC project is to improve the prognosis for patients with pancreatic adenocarcinoma (PDAC) by identifying predictive biomarkers that enable individualised and thus more effective first-line chemotherapy for PDAC patients in the long term. It is hypothesised that the use of CT-based data in combination with clinically established parameters (tumour marker CA 19-9, age, gender) and preclinical parameters (molecular subtype) will enable the development of algorithms based on machine learning. These should be able to predict various clinical endpoints. These include the response to one of two standard chemotherapies, the time to failure of first-line chemotherapy and the overall survival of patients. Three cohorts will be analysed: resectable PDAC who have received neoadjuvant and/or adjuvant chemotherapy and metastatic, non-resectable PDAC with palliative chemotherapy. In total, data from approximately 5000 patients will be analysed.

The knowledge gained from the results of RACOON-PDAC forms an excellent starting point for future prospective studies on the treatment of PDAC patients, but also for other oncological diseases. The use of structured radiological findings also contributes to a short-term improvement in the routine clinical care of patients. In addition to the primary benefits for the treatment of pancreatic cancer, this project can and should provide valuable algorithms for clinical trials at all Network University Medicine (NUM) sites.