Purpose To evaluate statistically significant differences of perfusion imaging for differentiating benign from malignant solitary pulmonary nodules (SPNs) by DSCT perfusion.Materials and methods Twenty-two patients met the inclusion criteria of a solitary less than 30mm pulmonary nodule without calcification or fat at computed tomography.Twenty patients were examined successfully; 17 received a histologic diagnosis, and three received a diagnosis by means of observation over 2years.Of these patients,there are 15 malignant nodule and 5 benign.Perfusion CT images were acquired 4seconds postrior to injecting contrast agent (via the antecubital vein at a rate of 4 ml/s) for a total of 50 secongs.Diagnostic characteristics for differentiation were examined by using time-density curve and other parameters for blood volume, time to peak and permeability.Receiver operating characteristic curves were calculated to test the usefulness of these parameters.The diagnostic performance of a combination of curve profiles and morphologic contrast material distribution were tested by using a decision tree.Results Frequency of malignancy was 75% (15 of 20 nodules).Malignant nodules showed stronger permeability with a higher time to peak and a faster slope (P <.001).Significant permeability was found only in malignant lesions (13 of 15 lesions).Sensitivity, specificity, and accuracy were 95.1%, 87.3%, and 94.4%,respectively, for time to peak; 91.4%, 73%, and 82.5% for blood volume; and 54%, 90%, and 72% for blood volume.When combined perfusion imaging and morphologic enhancement patterns, sensitivity increased to100%.Conclusions The application of perfusion imaging and its CT perfusion software by DSCT perfusion can be helpful in the differential diagnosis of between malignant and benign solitary pulmonary nodules.permeability seems to be highly specific for malignancy.