Imaging prediction of islet yield and post-operative insulin requirement in children undergoing total pancreatectomy with islet autotransplantation.

Background: Predicting post-operative glycemic control in children undergoing total pancreatectomy with islet autotransplantation (TPIAT) remains difficult. The purpose of our study was to explore preoperative imaging as a marker for islet yield and insulin need in pediatric patients undergoing TPIAT.

Methods: This was a retrospective study of children (≤18 years) who had undergone TPIAT between April 2015 and December 2018 and had 6 or more months of post-TPIAT follow-up. Patient specific factors (height, weight, body mass index [BMI], body surface area [BSA]) and pancreas volume segmented from the most recent pre-operative cross-sectional imaging were explored as predictors of islet yield (total islet counts [TIC], total islet equivalents [TIE], islet equivalents per kilogram body weight [IEQ/kg]) and glycemic control (total daily dose of insulin per kilogram body weight [TDD/kg], insulin independence) using Pearson correlation and univariate and multiple regression.

Results: Thirty-three patients, median age 13 years (IQR: 10-15 years), 64% female (21/33) met inclusion criteria. Nine patients (27%) achieved insulin independence at six months. Median TIE isolated was 310,000 (IQR: 200,000-460,000). Segmented pancreas volume was moderately associated with TIE (coefficient estimate = 0.34, p = 0.034). On multiple regression analysis, there was no significant predictor of insulin independence but number of attacks of pancreatitis (estimate = 0.024; p = 0.018) and segmented pancreas volume by body weight (estimate = -0.71; p < 0.001) were significant predictors of insulin TDD/kg.

Conclusion: Pancreas volume segmented from pre-TPIAT imaging has predictive performance for post-TPIAT insulin need in children.