BACKGROUND AND PURPOSE: Ventilation measured on 4D cone-beam computed tomography (CBCT) using deformable image registration (DIR) may predict specific radiation sensitivity, but the measurement is affected by the current image quality. With 4D computed tomography (CT) measured ventilation acting as a gold standard the current study investigates if image improvements increase the accuracy of 4D-CBCT measured ventilation.
MATERIAL AND METHODS: The study consists of 4D-CBCT and 4D-CT scans of 20 non-small-cell lung cancer patients. Raw CBCT projections were subjected to a standard or an improved projection correction and reconstructed by the common FDK-algorithm or the more advanced SART-algorithm. Ventilation was measured as Jacobians calculated from DIR and the comparison between CBCT and CT was done by Spearman correlation.
RESULTS: A significant increase in the mean correlation was observed when combining improved projection correction and SART reconstruction (0.34) compared to the clinical standard (0.21). The correlation further increased when averaging ventilation measured from three successive CBCT scans (0.38).
CONCLUSION: The study showed that the combination of improved projection correction and the SART reconstruction increased the accuracy of CBCT ventilation and this result can be a stepping stone to extract dynamic changes in respiration pattern of patients during radiotherapy.