Objective To investigate the effect of split-bolus spectral computed tomography(CT) on the portal venography and radiation dose. Methods The prospective study was adopted. The clinical data of 119 patients who underwent spectral CT at China-Japan Union Hopital from September 2014 to March 2015 were collected. Patients were randomly divided into the portal venography with split-bolus spectral CT single-phase enhanced scan group and portal vein multi-phase scan group by random sequence method. In the portal venography with split-bolus spectral CT single-phase enhanced scan group, the spectral CT was used with the method of split-bolus single phase imaging, and in the portal vein multi-phase scan group, standard spiral CT was used to perform three-phase scan. Two observers evaluated CT portal venography subjectively and objectively, measured CT values, contrast to noise ratio (CNR), signal noise ratio (SNR), and calculated radiation dose. Observed indices included (1) choice of optimal monochromatic images. (2) CT values of portal veins, measurement of CNR and SNR. (3) Subjective scoring of portal venography quality. (4) Comparison of radiation dose. Measurement data with normal distribution were presented as±s, comparison between groups was analyzed by independent sample t test. Comparison of count data was analyzed by chi-square test. Results The 113 patients were screened for eligibility, including 59 in the portal venography with split-bolus spectral CT single-phase enhanced scan group and 54 in the portal vein multi-phase scan group. (1) Choice of optimal monochromatic images: optimal monochromatic images were abstracted at 60 keV from spectral CT portal venography. (2) CT values of portal veins and measurement of CNR and SNR: the CT values of intrahepatic portal vein, extrahepatic portal vein and branches of portal vein were (319±44)HU, (328±53)HU, (294±45)HU in the reconstructed images at the energy level of 60 keV in the portal venography with split-bolus spectral CT single-phase enhanced scan group and (213±41)HU, (228±49)HU, (210±41)HU in the portal vein multi-phase scan group, with significant differences between the 2 groups(t=8.04, 6.34, 6.82, P 0.05). The SNR of intrahepatic portal vein, extrahepatic portal vein and branches of portal vein were 31±6, 29±6, 27±6 in the portal venography with split-bolus spectral CT single-phase enhanced scan group and 29±7, 28±9, 26±6 in the portal vein multi-phase scan group, respectively, with no significant differences (t=-0.688, 0.615, 0.600, P>0.05). (3) Subjective scoring of portal venography quality: the subjective score of image quality of portal venography was 14.3±1.0 in the portal venography with split-bolus spectral CT single-phase enhanced scan group and 12.5±1.8 in the portal vein multi-phase scan group, with a significant difference (t=12.43, P<0.05). (4) Comparison of radiation dose: the radiation dose was (8.1±1.1)mSv of patients in the portal venography with split-bolus spectral CT single-phase enhanced scan group and (17.4±7.5)mSv in the portal vein multi-phase scan group, with a significant difference (t=24.14, P<0.05). Conclusion Spectral CT portal vein imaging combined with split-bolus protocol can achieve better manifestations of portal vein and its branches, and reduce radiation dose in the scanning process. Key words: Liver diseases; Portal vein; Spilt-bolus; Tomography, X-ray computed; Spectral imaging; Radiation dose