Download Full Text (5.5 MB)

Presentation date


College, Institute, or Department

Radiation Oncology

Faculty Mentor

Dinah Clark

Research Mentor

Dr. Dandan Zheng


Radiomics is a process that mines quantitative data from imaging techniques, including MRIs, CTs, and PET scans. In this study, radiomics is used to find associations between clinical factors such as the number of metastases in pancreatic ductal adenocarcinoma (PDAC) patients from their preoperative CT scans. PDAC patients have a five-year survival rate only barely above 10%. By the time of diagnosis over half of pancreatic cancers are metastasized, and when that is the case the five-year survival rate is only 3%. It is hypothesized that radiomics can help predict clinical factors such as the number of metastases in pancreatic patients. These predictions can lead to improved treatment plans by risk stratification. Patients with a higher risk of metastases would adopt more aggressive treatment than patients with lower risk. This study included 87 patients diagnosed with PDAC who consented to a Rapid Autopsy Program. Although results have not yet been processed, future work would include finishing developing a model to use to predict PDAC metastases.


Radiomics, pancreatic cancer, image analysis

Radiomics: Medical Imaging Can Predict Pancreatic  Cancer Patient Outcome