Post-mortem human specimen (PMHS) experiments have correlated torso biomechanics (e.g., deflection and VC) to injury risk using the chestband, an instrumented flexible steel belt. From the chestband, time dependent contours are derived representing torso cross-sectional shape throughout the impact event. Biomechanical injury metrics are typically computed from uniaxial measurements between two contour points in order to develop metrics for uniaxial transducers within anthropomorphic test devices (ATDs). Yet, the relationship between uniaxial deformation measurements and the underlying tissue response in humans is not well understood. Prior PMHS experiments have suggested a causative relationship between localized posterolateral thorax loading and regional visceral injury, but sample size has precluded development of injury metrics specific to this loading direction. Therefore, a plane strain viscoelastic finite element model was employed to correlate posterolateral uniaxial deformations, i.e., deflection and VCmax, with visceral (spleen) strain and strain energy density responses. Model visceral geometry was also varied to simulate small, median, and large visceral volumes across the American population. The model was exercised by applying normalized subject-specific chestband deformations to the model periphery;​ material responses throughout the spleen were determined. Considering all simulations, Coefficients of Determination (R²) and the Predicted Sum of Squares (PRESS) were computed to quantify correlations between visceral responses (i.e., parenchymal strain, capsular strain, and strain energy density) and external metrics (i.e., deflection and VCmax) measured between 90° and 140° with respect to the anterior direction. PRESS and R² statistics were determined for 3 visceral anatomic variants, 11 angles, 2 metrics, and 3 visceral response parameters, totaling 198 analyses. Correlations at 140° were not significant for any metric or anatomic variant. Generally deflection was a better predictor of visceral response than VCmax;​ peak R² values occurred at 125° (capsular strain), 115° (parenchymal strain), and 90° (strain energy density). Because the loading set was composed of only ten experiments, the PRESS statistic determined the degree to which analyses were generalizable to a larger population. PRESS results indicated that deflection was a more generalizable metric than VCmax and that angles 115 – 125° were associated with greatest correlations. These results suggested that, to measure visceral injury risk, an ATD transducer would be appropriately located between 115o and 125o within the thorax. This model formulation is useful for parametric examination of visceral response to torso deformations measured experimentally using the chestband device.