Study design: A clinical study in human spinal cord. Objectives: To evaluate changes in spinal cord blood perfusion in patients with thoracic spinal stenosis using contrast-enhanced ultrasonography and to semiquantitatively analyze blood perfusion changes in compressed spinal cord before and after ventral decompression. Setting: Ultrasound department of a university hospital. Methods: Twelve patients with confirmed thoracic spinal stenosis who needed decompression surgery participated. They underwent an intravenous injection of a contrast agent before and after ventral decompression. Quantitative analysis software (Philips Healthcare, Bothell, WA, USA) was used to perform time-intensity curve (TIC) analysis. The enhanced intensity (ΔI), rise time (ΔT) and slope of the TIC (β) were separately calculated; t-tests of the independent samples were performed on the indicators. Results: The TICs showed no significant differences between compressed spinal cord and normal spinal cord in ΔT, enhanced ΔI and β (P = 0.46, P = 0.23 and P = 0.16, respectively). After ventral decompression, ΔI of the originally compressed spinal cord increased substantially (P = 0.04) compared with ΔI of the normal spinal cord; however, the ΔT and β showed no significant differences (P = 0.18 and P = 0.09, respectively). Comparison of the blood perfusion parameters (that is, ΔT and ΔI) of the compressed spinal cords before and after ventral decompression showed no significant differences (P = 0.14 and P = 0.12, respectively), but β showed significant difference (P = 0.02). Conclusion: Contrast-enhanced ultrasonography can dynamically display spinal cord blood perfusion. The characteristics of blood perfusion can be semiquantitatively analyzed using a software technique.