CT radiogenomic characterization of the alternative lengthening of telomeres phenotype in pancreatic neuroendocrine tumors

OBJECTIVE. The objective of this study was to identify imaging characteristics in patients with known pancreatic neuroendocrine tumors (PanNETs) that predict the alternative lengthening of telomeres (ALT) phenotype by blinded retrospective review of preoperative multiphasic CT scans. MATERIALS AND METHODS. For this retrospective study of 121 preoperative CT examinations of patients with resected PanNETs, two radiologists independently reviewed the CT examinations for tumor characteristics including size, shape, cystic or necrotic elements, calcifications, invasion of adjacent organs and vessels, biliary duct dilatation, pancreatic duct dilatation, and hepatic metastases. Univariate analysis of association of CT characteristics with ALT phenotype was performed with Fisher exact tests or t tests, and multivariate analysis was assessed by multiple logistic regression. RESULTS. Univariate analysis showed that the following CT features were significantly associated with the ALT phenotype: Lobulated or irregular tumor shape (p = 0.001), tumor necrosis (p = 0.002), vascular invasion (p < 0.001), pancreatic duct dilatation (p < 0.001), and hepatic metastasis (p < 0.001). Multivariate analysis found that the combination of pancreatic duct dilatation, hepatic metastasis, and size of tumor ≥ 3 cm was a strong predictor of ALT phenotype (odds ratio = 20.3; 95% CI = 2.3-176.3; AUC = 0.58; p = 0.006). CONCLUSION. This study showed that several preoperative CT features of PanNETs are associated with the ALT phenotype, which is known to predict poor prognosis. Additionally, CT findings of intratumoral calcifications and metastases predicted poor survival independent of the ALT status.