Would it be not attractive to have a scalable, imaging-based method at hand to predict pancreatic cancer before the patient becomes symptomatic and before a lesion becomes visible to the human eye�by �seeing the unseen�?