19 Assessment of Smart MAR (Metal Artifact Reduction) algorithm for metallic artifact correction in external radiotherapy planning

Abstract

Introduction When metal prosthesis was present in CT acquisition, like hip or spinal prostheses, image quality was deteriorated and the Hounsfield Units (HU) became corrupted by metal artifacts. They may obscure visualization of anatomical structures and affect dose calculation accuracy. Usually, these artifacts are manually corrected by delineation and by assigning electron density. This correction was time consuming and operator-dependent. The aim of this study was to evaluate the capability of MAR (Metal Artifact Reduction) algorithm1 [1] to restore real HU and avoid manual correction. Methods The phantom study was conducted using a CT electron density phantom CIRS (model 062). The phantom contained inserts made of various tissue-simulating materials. The phantom was scanned on a Optima 580 CT scanner (GE Healthcare) (120 kv, mAs auto) without metal (reference CT scan) and with titanium hip prosthesis (4.2 g.cm−3) and steel hip prosthesis (8 g.cm−3). These scans were reconstructed with and without MAR algorithm. In the presence of the metal prosthesis, the mean HU and corresponding standard deviations were determined for each rod with and without MAR correction and compare with the reference CT scan. Results Without MAR correction, HU were significantly impacted by metal artefact. HU errors were close to 20% for titanium hip prosthesis and 50% for steel hip prosthesis. With MAR correction, the mean HU were in good agreement with reference CT image (less 2%). However, residual artifacts were still visually observable. Conclusions Using MAR correction algorithm can potentially improve the global efficiency of RT treatment in the presence of metallic implants. In further works, the dosimetric impact of MAR will be investigated for various clinical ballistics.