诊断CT用于放疗模拟时的系统性旋转误差校正:一项基于图像后处理的临床验证研究OACHSSCD

Objective To investigate a systematic image rotational error discovered when using a diagnostic CT for radiotherapy simulation positioning of cancer patients,analyze its impact on three-dimensional(3D)translational setup accuracy,and validate the effectiveness of an image post-processing correction method.Methods Due to the breakdown of a dedicated CT simulator in the Department of Radiation Therapy,Tangdu Hospital,Air Force Medical University,twenty cancer patients underwent radiotherapy simulation on a diagnostic CT.A systematic clockwise roll angle error of approximately(1.0±0.2)° was identified in all resulting CT images.To correct this,a post-processing workflow was implemented using ImageJ software to apply an inverse transformation to the original DICOM images(unprocessed group),generating a new,geometrically corrected DICOM series(post-processed group).A retrospective registration analysis was then performed using the cone-beam CT(CBCT)images acquired at the first treatment for each patient.Each CBCT was rigidly registered to the planning CT of both the unprocessed and post-processed groups.The 3D translational setup errors in the left-right(X),superior-inferior(Y),and anterior-posterior(Z)directions were recorded.A paired-samples t-test was used to statistically evaluate the differences in errors between the two groups.Results Analysis of setup errors showed that for the unprocessed group,the mean errors in the Z and X directions were(0.38±0.10)cm and(0.38±0.08)cm,respectively.In the post-processed group,the mean error was significantly reduced to(0.19±0.08)cm in the Z direction and(0.27±0.06)cm in the X direction.These reductions were statistically significant in both directions(P<0.001).There was no statistically significant difference in the Y direction error between the two groups(P=0.89),which was consistent with the physical model of rotational error.Conclusion The use of a diagnostic CT for radiotherapy simulation positioning under non-standard conditions can introduce systematic image rotational errors due to inherent geometric deviations,which directly translate into clinical 3D setup errors and compromise treatment precision.A targeted image post-processing technique can accurately and effectively compensate for such systematic errors,significantly improving setup accuracy in critical directions.This method serves as a crucial,low-cost quality assurance measure to ensure the quality and safety of radiotherapy,particularly in exceptional circumstances.