In front crashes, belted drivers of vehicles with good crash test ratings sustain serious injuries to the thorax at much higher rates than to other non-extremity body regions. The full width US NCAP and moderate overlap IIHS evaluation programs both assess driver injury risk using the Hybrid III 50 th percentile male dummy. The current study evaluated the ability of peak sternum deflection and other dummy and vehicle metrics to predict NASS-CDS thoracic injury outcomes for drivers restrained by a seat belt and airbag while controlling for delta-V and driver age. Delta-V estimates were based on adjustments derived from event data recorder output for specific front crash configurations. Results showed that two vehicle metrics, bumper-to-firewall distance and shoulder belt force, each were better able to predict outcomes across the range of driver ages and thoracic injury types than any single dummy metric. Some dummy metrics did improve injury prediction for drivers aged 50 and older, including sternum deflection, deflection rate, and the ratio of sternum deflection to thoracic acceleration. However, none of these findings was significant at the p = 0.05 level for both soft- and hard-tissue injury, and reductions in the predictive ability of most metrics were associated with procedural changes made to each test program.