The aim of this study was to assess the diagnostic accuracy of PET-MRI to predict resectability of pancreatic adenocarcinoma after neoadjuvant chemotherapy ± radiation therapy.

Background: Pancreatic cancer is the 8th more common cancer in the world. At diagnosis, majority of patients present with unresectable locally advanced disease. Standard of care therapy for locally advanced pancreatic cancer includes chemotherapy ± radiation therapy. It is published that computed tomography underestimate the effectiveness of neoajuvant treatment and there is a lack of criteria allowing identifying the responders. The misinterpretation of scans may be linked to the large desmoplatic reaction, present in pancreatic cancer, which would not be expected to regress. PET-MR is an imaging technique that associates PET and MR imaging, performed during the same examination. The main hypothesis is that PET-MR imaging could accurately identify resectable and no resectable pancreatic adenocarcinoma after neoadjuvant chemotherapy ± radiation therapy.

Primary aim Assess the diagnostic accuracy of PET-MRI to predict resectability of pancreatic adenocarcinoma after neoadjuvant chemotherapy ± radiation therapy

Secondary aims Assess the accuracy of quantitative PET-MRI parameters to predict resectability and response of pancreatic adenocarcinoma after neoadjuvant chemotherapy ± radiation therapy Compare accuracy of PET-MRI and CT to predict resectability of pancreatic adenocarcinoma after neoadjuvant chemotherapy ± radiation therapy.

Assess inter and intra observer reproducibility of PET-MRI reading CT to predict resectability of pancreatic adenocarcinoma after neoadjuvant chemotherapy ± radiation therapy.

Number of subjects 125 Number of centers 8

Design 2 PET-MRI examination will be performed, one before the beginning of the neoadjuvant/induction treatment, and the second one after the neoadjuvant/induction treatment and less than 30 days before the surgery. The PET-MRI examinations will include whole body and organ specific imaging.

The whole body workflow will include

• [18F]-2-fluoro-2-deoxy-D-glucose PET acquisition
• T1-mDIXON imaging (for attenuation correction calculation)
• diffusion-weighted imaging
• T1-DIXON imaging post gadolinium chelate injection. The organ specific workflow will be focused on the abdominal area, including the liver and the pancreas, and will include
• [18F]-2-fluoro-2-deoxy-D-glucose PET acquisition,
• T2-weighted imaging with and without fat saturation,
• T1-DIXON imaging before and after dynamic injection of gadolinium chelate,
• diffusion-weighted imaging,
• IVIM-diffusion weighted imaging acquisition covering the pancreatic lesion. Qualitative analysis of PET-MRI using a Likert score will be compared to pathological results in order to obtain the accuracy of PET-MRI for resectability assessement.

Inclusion Criteria:

• Age > 18yo
• Locally advanced or borderline resectable pancreatic adenocarcinoma, according to the NCCN classification
• Eligible for neoadjuvant chemotherapy ± radiation therapy, according to the local tumor board
• With contraception if fertile woman
• With informed consent obtained
• Affiliated to French health care system

Exclusion Criteria:

• Previous treatment for the pancreatic adenocarcinoma
• Metastases
• Contra-indication to MRI acquisition (pace maker, metallic device, ..)
• Contra-indication to PET acquisition (non controlled diabetes with glycaemia > 11 mmol/L)
• Pregnancy or breast feeding
• Patient unable to give his consent