Relative and normalized iodine concentrations derived from photon counting computed tomography and their correlation with tumor grade and Ki67 in pancreatic neuroendocrine neoplasia: A pilot study

Neuroendocrine neoplasms are a rare and complex tumor entity, among which pancreatic neuroendocrine neoplasms generally display a more aggressive behavior. Despite a notable stage migration towards lower stages at initial diagnosis, the incidence of pancreatic neuroendocrine neoplasms is rising. In recent publications, iodine concentration derived from dual energy computed tomography was explored as a potential biomarker for pancreatic neuroendocrine neoplasia tumor grade and Ki67. However, methodologies exhibited significant variability, and reported outcomes were ambiguous, ranging from weak correlations to strong predictive performance in complex multivariate analyses. With the advent of photon counting computed tomography and improved technical capabilities, this study revisits the topic and aims to provide evidence for tumor characterization using photon counting computed tomography-derived iodine concentration in pancreatic neuroendocrine neoplasms. Iodine concentration in neuroendocrine pancreatic primaries was analyzed in the portal venous phase regarding correlation with histopathological tumor grade and Ki67. Iodine concentration was normalized to aortic iodine concentration (normalized iodine concentration), as well as calculated relative to unaffected pancreatic tissue (relative iodine concentration). Correlations were analyzed using Spearman's rank correlation, and mean concentrations were analyzed using Mann-Whitney U test. Eighteen cases with pancreatic neuroendocrine neoplasms were included. Relative Iodine concentration exhibited a strong and statistically significant correlation with tumor grade (ρ = 0.54, p = 0.02) and Ki67 (ρ = 0.53, p = 0.02). Mean relative iodine concentration was higher in high-grade tumors (p = 0.02). Normalized iodine concentration showed weak, non-significant correlations with tumor grade (ρ = 0.33, p = 0.18) and Ki67 (ρ = 0.30, p = 0.22). Mean normalized iodine concentration did not differ significantly between low-grade and high-grade pancreatic NEN. Our preliminary results show that photon counting computed tomography derived iodine concentration and especially relative iodine concentration is a potential biomarker for tumor grade and Ki67 prediction with strong, statistically significant correlations in untreated pancreatic neuroendocrine neoplasms. The method is non-invasive, requires little to no additional resources and may support early, evidence-based therapeutic decisions even before primary tumor biopsy.