CT-free attenuation correction for cardiac-dedicated CZT SPECT: first results of the CASCTEC study

Aim/Introduction The diagnostic accuracy of myocardial perfusion scintigraphy (MPS) for detecting coronary artery diseases (CAD) is improved when using attenuation-corrected (AC) images. However, for economic reasons, most clinical cardiac-dedicated CZT cameras are marketed without a CT. Therefore, the CASCTEC clinical study aims at developing an approach to generate AC images only from emission scans on a pinhole CZT camera, without the need for CT. Materials & Methods Up to now 25 patients referred for MPS have been enrolled in the CASCTEC single-center prospective study. Two-day stress/rest myocardial SPECT imaging was performed on a pinhole CZT camera. A standard of care (SOC) 5-10 minutes acquisition was immediately followed by an additional 5 minutes acquisition after radially retracting the detectors (FAR) without moving the patient. Subsequently all patients underwent SPECT/CT imaging to be used as reference. The FAR acquisition allowed to delineate the patient body contour, and together with the SOC acquisition, enabled the segmentation of soft and lung tissues to produce an attenuation map. Non-corrected (NC) and AC images were compared together and with respect to the SPECT/CT images using the AHA 17-segment model polar maps and a polar sub-segmentation of these maps (104 segments). All (sub-)segments were scored according to their relative perfusion uptake, and combined to generate stress and rest summed scores. Results The tracer distribution was found more homogeneous for CASCTEC-AC patient stress and rest perfusion images compared to NC images, and similar to the reference. All perfusion defects observed on the SPECT/CT-AC images were also visible on CASCTEC-AC images. The CASCTEC mean relative uptake increased by up to 40% in the inferior segments compared to the NC image. Changes by less than 10% were observed for all other segments, except for the apical segment that decreased by up to 25%. The differences of the changes in the (17-segment) right coronary artery territories (RCA) summed scores from NC to AC between CASCTEC and SPECT/CT images were 0.2±3.0 (p=0.41) and 0.6±2.2 (p=0.21), respectively for stress and rest scans. Similar results were obtained for the differences of the changes in the (104-segment) RCA summed scores, i.e. 0.1±2.7 (p=0.20) and 0.1±2.4 (p=0.14), respectively for stress and rest scans. Conclusions The CASCTEC CT-free patient-specific attenuation map enhances NC perfusion in the inferior wall similarly to CT-based AC images from a SPECT/CT acquisition. This method looks promising to improve the diagnostic accuracy of current cardiac pinhole CZT systems for detecting obstructive CAD but requires further studies.60 years old man – BMI = 30 73 years old man – BMI = 38.5