The influence of low- dose CT attenuation correction on artefacts of myocardial SPECT images for nuclear medicine studies

Abstract

Introduction: Myocardial perfusion imaging (MPI) with single photon emission tomography (SPECT) imaging has improved the sensitivity of myocardial scintigraphy in coronary artery disease (CAD), but often at the expense of specificity. Artefacts in SPECT imaging cause poor image quality, which may present as false-positive results that may lead to unnecessary, expensive and invasive procedures. Myocardial perfusion studies performed with Tc-99m sestamibi, consist of stress and rest imaging with patients positioned in a supine or prone position which can be performed on the same day as a one day protocol or on different days as a two day protocol. Breast and diaphragm attenuation may cause anterior and/or inferior soft tissue artefacts during Tc-99m sestamibi imaging and may interfere with the visualisation of the perfusion defects of the myocardium. The application of attenuation correction (AC-A) to the breast tissue and diaphragm of the patient might improve the images and provide a clear view of the myocardium. Aim: The study aimed at evaluating the efficacy of x-ray- derived attenuation correction in myocardial perfusion scintigraphy. The primary objective of the study was to determine whether hybrid SPECT with computed tomography (CT), will improve the image quality by reducing the soft tissue artefacts with the application of attenuation correction maps in the stress and rest Tc-99m sestamibi myocardial SPECT perfusion studies. Materials and Methods: A retrospective study was performed on 100 image sets (an image set equals a stress and rest study) of patients with suspected ischaemic heart disease referred to the Nuclear Medicine department between January 2015 and December 2016 for myocardial perfusion scintigraphy. The results of the patients were quantitatively reviewed with a descriptive cross sectional design with a comparison group. The images were given a score of 0-4. A score of 0 implies normal study, 1 equals a defect with no change, 2 equals the defect change <50%, 3 equals the defect change >50% and 4 was given when an artefact was generated by the application of AC on the non- attenuation corrected (NAC) images of the stress and rest studies. An overall score of the results of the improvement (yes) versus no improvement (no) of AC-A to stress and rest studies was also given. The NAC and the attenuation corrected (AC) stress and rest images are the dependent variables of the study. The independent variables which are the relevant factors such as gender, age, weight, height and position of the patients during imaging are also discussed for their influence on the attenuation correction on the stress and rest images. The critical assessment areas were the inferior and anterior wall defects. The image analysis was performed by the two NM physicians, including the co-supervisor of this study. For statistical analysis an inferential multivariate technique, called a Generalized Linear Model, was applied to determined relationships between the dependent and independent variables. Frequency tables for the results are demonstrated in the form of tables and bar charts. Results: The result of the stress as opposed to rest inferior images (with >50% change of the defect improvement) revealed a score of 46%. The effect of the independent variable on the dependent variable shows with the p-value being >0.05, non-significance. The scores showed AC-A improved 49% of the image sets and 51% of the image sets did not improve with the overall outcome. However, this score of 51% with no improvement, excludes 18 images of stress studies and nine images of rest studies which showed improvement (scores 2-3) in the inferior wall of the myocardium. Discussion: The results demonstrated that with anterior stress and rest SPECT images, the most artefacts were created with AC-A. In the stress studies, the anterior wall showed 45 artefacts and the rest study 55 artefacts. Improvement for the stress inferior images is 61% of which 20 images are less than 50% improvement and 41 images are more than 50% improvement. The stress, as opposed to rest inferior images (with >50% change of the defect) showed an improvement with a score of 46%. AC-A to the stress and rest AC and NAC images scored an improvement of 49% and 51% of no improvement for the overall outcome. In the total amount of stress studies reviewed, it was found that 18 stress studies showed improvement and nine rest studies showed improvement (scores 2-3) in the inferior wall of the myocardium but with no overall improvement. For nine image sets of stress/rest image sets, an improvement was observed in both stress and rest studies in the inferior myocardial walls but AC-A did not improve the results. The relative factors indicate with a p-value >0.05, that these factors have no influence on the outcome with the application of AC-A on the images. Conclusion: There is a significant difference in the AC-A scores of the images in the inferior wall of the myocardium. Therefore, the AC versus NAC with MPI shows a significant influence on the outcome of the study for the inferior wall of the myocardium.