Ted Nilsson, Pawel Rasinski, Ernesto Sparrelid, Antonios Tzortzakakis, Thuy A. Tran, Örjan Smedby, Rimma Axelsson, Mark Lubberink, Maria Holstensson
Abstract Purpose Fibroblast activation protein (FAP)–targeted PET using [⁶⁸Ga]Ga-FAPI-46 visualizes fibroblasts abundant in pancreatic cancer (PC) but also present in pancreatitis, complicating interpretation of static images. Dynamic imaging and kinetic modeling may provide additional insight, but their diagnostic value remains unclear. This study evaluated whether kinetic parameters from dynamic [⁶⁸Ga]Ga-FAPI-46 PET can differentiate PC from pancreatitis and their relationship with standardized uptake value (SUV) and tumor-to-blood ratio (TBR). Methods Sixty-one patients with suspected pancreaticobiliary cancer underwent a 45-min dynamic [⁶⁸Ga]Ga-FAPI-46 PET scan, followed by static scans at 60 and 180 min. Time–activity curves were generated for 51 malignant and 53 benign lesions. Compartment models and Logan analysis yielded kinetic parameters (K 1 , k 2 , k 3 , k 4 , V T , V NS , V S ). SUV and TBR were correlated with V T , and group comparisons and ROC analyses assessed discriminatory performance. Results Reversible models best described the tracer kinetics. V T and V S were significantly higher in PC than pancreatitis, and k 2 and k 4 were lower, indicating higher [⁶⁸Ga]Ga-FAPI-46 binding respectively slower washout in malignant lesions. SUV correlated strongly with V T ( r ≥ 0.784), and TBR showed very strong correlations ( r ≥ 0.902) for the 0–60 min interval, with strong correlations observed across all models and time points. ROC analyses demonstrated comparable differentiation between V T , SUV max , and TBR max . Conclusions Kinetic parameters showed strong correlations with simplified methods and similar ability to differentiate PC from pancreatitis. SUV and TBR measures thus represent practical alternatives to kinetic modelling for lesion characterization. ClinicalTrials.gov ID: NCT05172310