Real-world Performance of Multiparametric MRI with Diffusion-Weighted Imaging and MRCP for Pancreatic Cancer Surveillance

Purpose This study aimed to clarify the real-world performance of multiparametric MRI with diffusion-weighted imaging and magnetic resonance cholangiopancreatography for surveillance of patients affected by pancreatic cancer. Methods Findings for 169 high-risk patients who underwent surveillance from June 2017 to February 2023 were analyzed for this multicenter retrospective observational study analyzed. Those included patients who had high risk findings such as cystic lesions identified during screening or incidental imaging examinations, but not findings suggestive of malignancy shown by either EUS or multiparametric MRI. Surveillance was performed every six months using multiparametric MRI and tumor marker assessments. A variant was defined as either main pancreatic duct stenosis with tail-side dilatation, or an area with high signal intensity in diffusion-weighted images and low signal intensity in the apparent diffusion coefficient map. The primary outcome was variant incidence within a six-month period and secondary outcome was positive predictive value for variants. Results The median follow-up period was 30 months (range 6–78 months). The incidence of variants over six months ranged from 1.8–13%, with a median of 2.8%. A total of 414 person-years of follow-up were accrued, during which 19 variants were identified. The event rate for variants, calculated using person-years, was 0.046. Of the 19 variants, two were confirmed as cancer, for a positive predictive value of 11%. Conclusions Multiparametric MRI-based surveillance was shown to be minimally invasive and enabled detection of pancreatic cancer at a curable stage, suggesting its potential as a standard method for detecting high-risk pancreatic cancer.