This retrospective study evaluated the diagnostic benefit of a deep learning image reconstruction (DLIR) algorithm for detecting and characterizing vertebral compression fractures (VCFs) using dual-energy computed tomography (DECT). The authors compared DLIR with conventional iterative reconstruction in 80 patients who underwent DECT for suspected VCFs. Using material decomposition images to assess bone marrow edema, the DLIR algorithm achieved superior image quality, reduced noise, and improved signal-to-noise and contrast-to-noise ratios. Radiologists rated fracture delineation and bone marrow visualization significantly higher with DLIR images. These improvements enhanced the detection of acute versus chronic VCFs without increasing radiation dose.