Increasing survival by early detection of pancreatic cancer
Pancreatic cancer is the 10th most common cancer in the United Kingdom. The cancer is often diagnosed when it has already spread outside the pancreas preventing curative surgery. This limits treatment options for patients, with few people surviving for more than one year.
The study to be carried out by Dr Susanna Kallioinen at University Hospital Coventry and Warwickshire will use CT images from pancreatic cancer patients that were obtained at least 180 days prior to their diagnosis of pancreatic cancer. These images will be compared to images from patients with no disease, that were obtained at a similar time and are of a similar age to the pancreatic cancer patients. An advanced computer technique called deep learning to identify high risk features can be used to predict who may develop pancreatic cancer. Dr Kallioinen will also use deep learning techniques on the CT images obtained at diagnosis to investigate whether it is possible to predict what type of treatment they will get and whether the treatment will be effective or not. If it is possible to predict who will develop pancreatic cancer, hospital doctors may be able to offer curative treatment more frequently. This should increase the number of people who survive longer after being diagnosed with pancreatic cancer.