A Clinical Evaluation Of Setup Errors Of IMRT For Head And Neck Cancers Using Integrated KV Cone-beam CT On IGRT Unit

Objective: Using KV-cone beam CT (CBCT) on IGRT to study the translational and rotational setup error in patients with carcinoma of head and neck treated by intensity modulated radiation therapy (IMRT),and to analyze the effect on target coverage and sparing of organs at risk.Methods: This study includes three stages: 1)12 patients with carcinoma of head and neck treated by IMRT were enrolled into the study of real-time correction of setup errors using KV-CBCT for a total of 217 times.XVI were acquired before radiation,between radiation(after correct the setup error),after radiation and compared to planned CT images with calculate the setup error of X-axis,Y-axis,Z-axis both of the translational and rotational errors.Then analyze the changes of setup errors after the correction and between the radiation treatment.2)Compare the setup errors detected by CBCT and EPID,and analyze the differences between them.3)Scan the MED-TEC phantom to get XVI images by the CBCT,then use the computer simulation to investigate the effect of translational errors alone,rotational errors alone,and combined translational and rotational errors on target coverage and sparing of organs at risk.Results: All 12 patients received CBCT scan at each treatment fraction before radiation delivery.A total of 217 CBCT scans were obtained.The systemic±random errors on translational X,Y,Z directions were (-0.04±2.63),(0.07±1.69),(-1.15±1.33)mm,on rotational X,Y,Z directions were (0.10±1.14),(0.16±1.41),(-0.06±1.22)°. All the setup errors were corrected,and after the correction both the systematic errors and random errors were obviously lower than those before correction.If we only compare the result of the translational errors <1mm and the rotational errors<1°,the errors were not obviously diminished,a part of them increased inversely,so we define the 1mm and 1°as the correction threshold.Then the setup errors after treatment were compared with that after correction but before treatment,we found that the differents between them were not significant.2)The estimated margins required for CTV-PTV were calculated according to the detected results by CBCT and the van Herk formula,which decreased 4.3mm in X-axis,4.4mm in Y-axis,1.7mm in Z-axis,compared with the CTV margins without CBCT.And the margins were smaller than that we calculated formally by EPID.3)Both translational and rotational errors resulted in decreased coverage of the target and increased doses to the organs at risk,when they coexisted the effect were larger,even the small translational and rotational error stack can change the dose distribution.Conclusions: This study shows as follows:1)Real-time correction of setup errors in patients with carcinoma of head and neck by CBCT can reduce the systematic error and random error.We define the 1mm and 1°as the correction threshold.And the setup error between the treatment is small.2)CBCT can detect the translational error much exactly than EPID,and less impacted by subjective factor,furthermore CBCT can detect rotational errors in three directions.So CBCT have more advantages than EPID to guide radiation more precisely.3)Rotational error can raise the effect of translational error on target coverage and sparing of organs at risk,and correction of the translational errors alone can not recover the dose well,so correctting rotational errors has important clinical meaning.