17 Impact of a CT artifact reduction algorithm in radiotherapy: preliminary qualitative and quantitative aspects. (December 2018)

• Record Type: Journal Article
• Authors: Villeneuve, R.; Biteghe, S.; Allegrini, O.; Amblard, R.; Garnier, N.; Serrano, B.
• Is Part Of: Physica medica. Volume 56(2018)Supplement 1
• Page Start: 48
• Publication Date: 2018-12
• Subjects: Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2018.09.099
• Languages: English
• ISSNs: 1120-1797

Abstract

Abstract : Introduction: In radiotherapy, CT images have two main objectives: allow volumes delineation and obtain electron densities for dose calculation. The presence of dense structures such as hip prostheses or dental implants create artifacts that can significantly impair both of these goals. Iterative Metal Artifact Reduction (IMAR) is a metal artifact reduction algorithm developed by Siemens, which greatly facilitates the delimitation of volumes. We studied its impact on the modification of Hounsfield units (HU) as well as the dosimetric consequences for pelvis and head&neck (H&N) regions. Methods: All CT acquisitions were made on a Somatom AS 20 Definition (Siemens) at 120 kV, 400 mAs and 2 mm slice thickness. For the pelvic region, we used a phantom composed of a water equivalent material in which we inserted two hip prostheses. For the H&N sphere, we used an anthropomorphic phantom (CIRS Atom 701) with two pieces of lead inserted in the teeth. CT acquisitions were performed with inserts (prostheses or lead), with IMAR reconstruction (imarCT), and without (artefactCT). We used the protocol HIP implants for pelvis and Dental fillings for H&N. A CT acquisitions without insert were also done (baselineCT). Using the Matlab software (V2014), we calculated, in significant regions (between metal inserts), the HU differences between baselineCT and respectively imarCT and artefactCT. In order to estimate the dosimetric impact of these HU variations we performed a treatment planification on both phantoms and on patients with pelvic and dental implants with the treatment planning system V13.6 (Varian). In pelvic we use 4 static fields of 15 MV and in the H&N 2 RapidArc beams of 6 MV. The results are given in terms of mean, standard deviation ( σ ) and most likely value (p). Results: For the pelvic area, the HU deviations from the baseline CT are 84 HU ( σ = 53 HU, p = 100 HU) and −106 HU ( σ = 487 HU, p = 200 HU) respectively for imarCT and artefactCT. For the H&N the HU deviations are 15 HU ( σ = 34 HU, p = 0 HU) and 44 HU ( σ = 413 HU, p = 200 HU) respectively for imarCT and artefactCT. We note in both cases a spread of HU without IMAR and a slight overestimation of HU with IMAR. The dosimetric impact, based on the comparison between the planification made with imarCT and artefactCT showed for the pelvic area, in the phantom, a dose variation from −11% to 1% with an average of −4.6% ( σ = 3%) while in the patient, the variation is between −10.8% and +0.4% with an average of−3.4% ( σ = 2.3%). For the H&N area, in the phantom, the difference in dose varies from −10% to 7.4% with an average of −0.6% ( σ = 2%), whereas in the patient, the difference in dose varies from −3% to 24% with an average of 0.6% ( σ = 2%). Conclusions: In addition to being an undeniable aid to delineation, it seems in this preliminary study that IMAR is a significant tool to issue metal artefact problem in calculating dose in patient with metal implants close to treated areas.