[Scan acquisition parameter optimization for the treatment of moving tumors in radiotherapy]

In the case of lung tumor treatment, to adjust 3D helical computed tomography (CT) acquisition parameters using a dynamic phantom and compare to the theory the volumes of a moving object.

Three helical CT acquisitions were compared using a Big Bore CT scan : an "initial" 3D CT scan (constructor parameters), an "optimized" 3D CT scan which parameters are chosen to obtain an axial slow scan like acquisition and a 4D CT scan. We used a phantom composed by a ball filled with water set on a dynamic platform moving in the antero-posterior or cranio-caudal direction with a 14 mm amplitude and a 4s period. For each acquisition and modality (static and dynamic), we quantified the ball volume by automatic contouring and we estimated relative errors.

For an antero-posterior displacement, the volume of the moving ball is under estimated by 14.1 % with the "initial" scan, by 0.2 % with the "optimized" scan and over estimated by 0.8 % with the averaged 4D scan. For a cranio-caudal displacement, it is under estimated by about 22 % with the "initial" scan and by about 1 % with the "optimized" scan and the averaged 4D scan.

Volume measurements performed with the dynamic phantom allowed us to validate the "optimized" 3D CT scan parameters because it accurately reflects the volume of a moving object. Radiotherapy departments without 4D CT should adapt scan parameters for internal target volume definition.