Assessment of individual tumor biology and response to systemic therapy in pancreatic ductal adenocarcinoma (PDAC) remains a clinical challenge. The significance of anthropometric (body composition) changes during chemotherapy as a surrogate for tumor biology in the setting of localized PDAC is unknown. A retrospective, single-institution analysis of patients with PDAC who received neoadjuvant therapy (NAT) and pancreatectomy from 2017 to 2021 was performed. Radiologic anthropometric analysis used artificial intelligence-driven software to segment and compute total and sub-compartment muscle area, adipose tissue area, and attenuation values at the level of the L3 vertebra. Kaplan-Meier survival estimates, log-rank tests, and multivariable Cox regression models were used in survival analyses. The inclusion criteria were met by 138 patients. Although decreases in muscle and adipose tissue areas during NAT were predominant, a subset of patients experienced an increase in these compartments. Increases in muscle greater than 5% (hazard ratio [HR], 0.352; 95% confidence interval [CI] 0.135-0.918; p=0.033) and increases in adipose tissue greater than 15% (HR, 0.375; 95% CI 0.144-0.978; p=0.045), were significantly associated with improved survival, whereas loss of visceral fat greater than 15% was detrimental (HR 1.853; CI 1.099-3.124; p=0.021). No significant associations with single time-point anthropometrics were observed. Gains in total muscle and adipose mass were associated with improved pathologic response to systemic therapy and less advanced pathologic tumor stage. Dynamic anthropometric analysis during NAT for PDAC is a stronger prognostic indicator than measurements taken at a single point in time. Repeated anthropometric analysis during preoperative chemotherapy may serve as a biomarker for individual tumor biology and response to therapy.