Rofo 2004; 176 - PO134
DOI: 10.1055/s-2004-828127

Image quality of volume rendering of the bronchial tree: Kymogram-gated versus retrospectively ECG-gated and non-gated MDCT

HA Alkadhi 1, S Wildermuth 1, DA Sennst 1, T Schertler 1, M Kachelries 1, W Kalender 1, B Marincek 1, T Böhm 1
  • 1Institut für Radiologie, Departement für Medizinische Radiologie, Zürich, Schweiz

Ziele: To assess the image quality of different heart-cycle-synchronized MDCT reconstruction techniques for volume rendered 3D visualization of the bronchial tree. Methode: Non-ECG-gated and retrospectively ECG-gated MDCT was performed in 25 patients. The ECG-gated raw data were reconstructed using absolute reverse delays of 200 to 500 msec. The same raw data were reconstructed using kymogram-gated technique in 10 phases of the cardiac cycle. For both reconstructions the optimal artifact-free diastolic phase was determined. Volume rendered reconstructions of the bronchial tree were generated from these two data sets and from the non-gated data. Stair-step artifacts of the carina, artifacts in the lung parenchyma, and the extent of bronchial tree visualization were rated by two blinded readers. Ergebnis: Kymogram-gated reconstruction showed stair-step artifacts to the same extent as non-gated reconstruction, but less compared to ECG-gated reconstruction in right (P<0.01), left (P<0.05), and in both main bronchi (P<0.001). Artifacts in the lung parenchyma were similar with kymogram-gated and non-gated reconstruction, but less compared to ECG-gated reconstruction for the right (P<0.05), left (P<0.05), and both lungs (P<0.01). Kymogram-gated reconstruction showed no significant differences in visualization of segmental and subsegmental bronchi compared to ECG-gated reconstruction, but was inferior to non-gated reconstruction for the right (P<0.01), left (P<0.05), and both lungs (P<0.05). Schlussfolgerung: Kymogram-gated reconstruction results in significantly fewer artifacts when compared to ECG-gated reconstruction. Non-ECG-gated reconstruction shows the best results of bronchial tree visualization together with a low artifact level and remains the method of choice for acquiring data for 3D post-processing.