Estimation of Dose Distributions with the Dental Artifact Correction in Radiation Treatment Plans for Head-And-Neck Cancers
M Lee1*, T Suh1, J Park1, J Lee2, H Park1,3, B Lee4, (1) The Catholic University of Korea, College of Medicine, Seoul (2) Department of Radiation Oncology, Konkuk University Medical Center, Seoul (3) Department of Radiation Oncology, Ajou University School of Medicine, Suwon, Gyeonggi-do (4) Department of Radiation Oncology, Sun Medical Center, DaejeonTU-E-141-6 Tuesday 2:00PM - 3:50PM Room: 141
Purpose: Metal artifacts generated by dental implants cause dose discrepancy in intensity-modulated radiation therapy (IMRT) of head-and-neck cancer cases due to the inaccurate predicted doses. To determine the effective metal artifact reduction (MAR) method presenting minimum dose errors with delivered dose, dosimetric impact of different MAR methods were evaluated in radiation treatment plans.
Methods: MAR was implemented using Euler elastic and curvature based sinogram inpainting (EECSI) algorithm, volume clustering with Hounsfield unit (HU) overriding (VCU), and multimodal image-based correction (MIC). When delivered doses were calculated on the corrected computed tomography (CT) images by employing the same plan parameters in the reference IMRT plan (Planref), dose errors were analyzed with dose-volume histograms (DVHs), monitor unit, and dose conformity of target volumes. Dosimetric impacts were also estimated using the biophysical evaluation parameters in three oropharyngeal patients.
Results: Optimized fluence in the Planref resulted in overdose up to 5% in the salivary glands and spinal cords on the corrected CT images, while the overdose in target volumes increased tumor control probability of 3-7%. Target doses in peripheral regions affected by the surrounded dark pixels were deformed and decreased conformity. The DVHs of organs at risk (OAR) and target volumes also showed the 5-8% volume gap in the prescription dose range. As the efficient MAR is employed into the head-and-neck IMRT plans, more accurate predicted doses of primary structures would be achieved, even if the dose differences of each MAR method are small less than 2-3%.
Conclusion: EEECSI and VCU were useful methods to reduce the metal artifacts and to predict more accurate delivered doses. Since dental artifacts cause dose variation in peripheral region of target volumes and OAR, dose calculation on corrected CT images using efficient MAR techniques and dose verification would be encouraged in head-and-neck IMRT plans.