PT - JOURNAL ARTICLE AU - Sven Prevrhal AU - Kevn Brown TI - Removal of CT metal artifacts improves SPECT images in SPECT/CT imaging DP - 2010 May 01 TA - Journal of Nuclear Medicine PG - 1358--1358 VI - 51 IP - supplement 2 4099 - http://jnm.snmjournals.org/content/51/supplement_2/1358.short 4100 - http://jnm.snmjournals.org/content/51/supplement_2/1358.full SO - J Nucl Med2010 May 01; 51 AB - 1358 Objectives Metallic objects such as orthopedic hardware, pacemakers and dental implants can create streak artifacts that propagate to CT-based SPECT attenuation correction (AC) maps and result in artifacts in AC-corrected SPECT image data. We hypothesized that application of a novel image-based CT artifact reduction (MAR) algorithm could significantly reduce such artifacts. Methods The novel algorithm is image-based and does not require use of CT projection data. It can therefore be used retrospectively. Scans of a phantom simulating a prosthetic hip with the implant centered in a cylindrical water tank acquired on a Brightview XCT (Philips) SPECT/CT unit and archive patient SPECT/CT scans covering a variety of metallic objects were subjected to MAR for quantitative and qualitative assessment. Artifact-reduced AC maps and SPECT reconstructions were compared to native data. Results Clear improvement for both CT and SPECT images was seen for artifacts induced by large high-density objects such as prostheses and dental fillings. In phantoms with homogeneous background activity, homogeneity in the vicinity of the attenuator was significantly improved. Patient images also revealed marked differences. Artifacts of smaller objects such as pacemaker lead wires could not be reduced as effectively. Conclusions CT metal artifacts indeed propagate to SPECT images and bear significant risk of masking important findings, however their negative impact could be substantially mitigated with the proposed approach. In addition, image-based CT metal artifact reduction algorithms have higher prospects for quick adoption than previously proposed projection-data based algorithms. Research Support Internally funded by Philips Healthcare