Interobserver agreement with MRI vs. CT for pancreatic tumour size measurement before and after neoadjuvant therapy

To assess inter-observer agreement of pancreatic ductal adenocarcinoma (PDAC) tumour size measurement using CT and MRI, before and after neoadjuvant therapy (NAT) MATERIALS AND METHODS: In this single-centre retrospective observational cohort study, we included consecutive patients with histologically documented PDAC diagnosed in 2010-2022 and imaged by CT and MRI before and after NAT. Three blinded independent radiologists evaluated the largest tumour length on two CT acquisitions and six MRI sequences. Inter-observer agreement was assessed by the intra-class correlation coefficient (ICC) and LOAM graphs (Bland & Altmann extension for multiple observers).

We included 50 patients (26 females, median age 68 [61-74] y). By CT, inter-observer agreement was excellent before NAT (ICC, 0.83 [0.73-0.90] at the arterial phase and 0.84 [0.74-0.90] at the portal-venous phase) and good after NAT (ICC, 0.66 [0.52-0.78] at the arterial phase and 0.65 [0.51-0.77] at the portal-venous phase). By MRI, inter-observer agreement was moderate-to-good before NAT (T1 at the arterial phase being the best sequence; ICC, 0.67 [0.53-0.79]) and only moderate for all sequences after NAT (T1 at the late phase being the best sequence; ICC, 0.55 [0.37-0.71]).

Inter-observer agreement for PDAC size measurement was better by CT than MRI and was better, with both methods, before vs. after NAT.