RESUMEN
ObjectiveBoth kilovolt (KV) cone-beam computed tomography (CBCT) and KV radiography can determine set-up errors for patients with head-and-neck cancer.This study is to compare their performance.Methods16 patients with head and neck cancer were enrolled in this study.There were 160 sets of CBCT and corresponding orthogonal radiography images.Through registration of CBCT images with the planning CT images,and registration of radiography images with the digitally reconstructed radiographs,translational set-up errors were determined along left-right (x),super-inferior (y),and anterior-posterior (z) directions.Pearson correlation analysis was performed to evaluate the correlation of the set up errors determined by the two methods,and Bland-Altman plot analysis was used to assess the coincidence of these two methods.ResultsThe Pearson coefficient of correlation along all three directions was less than 0.01,and R2 was 0.95,0.84,0.81 on x,y,z,respectively.That means high correlation for two methods.The Bland-Altman plot analysis showed that the 95％ agreement limits of agreement were within preset 2 mm tolerance (x[ 1.3mm,- 1.2mm],y[ 1.6mm,- 1.1mm],z[0.8mm,-1.4 mm] ),which indicates an agreement exists for two methods.ConclusionsFor determination of set-up errors for patients with head and neck cancer,KV radiography is equivalent to CBCT.Considering CBCT delivers higher dose than KV radiography,but provides more soft tissue information.We suggest to use these two methods combinative in clinic.