Utilization of Imaging in the Diagnosis and Surveillance of Pancreatic Ductal Adenocarcinoma (PDAC): Analysis of the TRICARE Data

Background It is unclear if adding MRI in the pre-treatment staging for pancreatic ductal adenocarcinoma (PDAC) is associated with changes in treatment compared to only obtaining a CT scan of the abdomen/pelvis. Similarly, there remains unclear benefit from frequent postoperative surveillance imaging after surgical resection for PDAC. Methodology Claims data from the Military Health Services Data Repository were used to identify all adult patients with PDAC who underwent surgical resection between 2017 and 2022. Patient demographics, medical and surgical interventions, radiographic studies, and survival data were sourced from the TRICARE database. The proportion of patients undergoing neoadjuvant chemotherapy compared to upfront surgical resection was compared between patients with and without a preoperative abdominal MRI, along with those who underwent one versus multiple preoperative CTs, after controlling for covariates of interest. Overall survival was compared between patients who underwent surveillance imaging every six months versus those with more frequent imaging. Results There was no significant difference in whether patients who had preoperative MRIs underwent surgery first (88.5% vs. 77.9%, p = 0.58). However, among patients who received multiple CT scans, there was a higher proportion who underwent surgical resection before chemotherapy (92.2% vs. 74.3%, p = 0.017). Decreased surveillance intervals were associated with a decreased hazard ratio (0.88, p = 0.007) when compared to the six-month standard surveillance frequency, although only after the 24-month mark. Conclusions Overall, preoperative MRIs were not associated with initial choice of treatment (i.e., surgery first or chemotherapy first). Frequency of surveillance imaging might differ before and after the 24-month mark, but disentangling the reason for imaging in the post-surgical surveillance period is difficult and introduces bias.