Quantification and determination of normal 123I-Meta iodobenzylguanidine heart-to-Mediastinum ratio (HMR) from cardiac SPECT/CT and correlation with planar HMR

Abstract

Assessment of cardiac 123I-meta iodobenzylguanidine (123I-mIBG) uptake relies on the heart-to-mediastinum ratio (HMR) derived from planar images. We have developed novel semiautomated quantitative methodologies for assessing HMR from SPECT images using a dedicated cardiac multipinhole SPECT/CT system and determined the lower limit of normal (LLN) SPECT-derived HMR and the correlation to planar-derived HMR. Methods: Twenty-one healthy volunteers were injected with 123I-mIBG and imaged using 2 different cameras. Planar images were acquired using a conventional SPECT camera equipped with parallel hole collimators, and hybrid SPECT/ CT images were acquired using a dedicated cardiac SPECT system with 19 pinhole collimators interfaced with 64-slice CT. Planar HMR was calculated as per standard guidelines (manual traditional method) and elliptic region-of-interest (Elip-ROI) and region-growing (RG-ROI) techniques. SPECT HMR was quantified using a new method that incorporates various cardiac and mediastinal segmentation schemes in which upper and lower limits of the heart were determined from CT and the left ventricular ROI, and mean counts were calculated using Elip-ROI and RG-ROI techniques. Mean counts in mediastinal ROI were computed from a fixed volume in 3 different regions: upper mediastinum (UM), lower mediastinum (LM), and contralateral lung (CL). HMRs were processed by 2 observers, and reproducibility was assessed by intraclass correlation coefficient and Bland–Altman analysis. Results: Planar HMR calculated using the RG-ROI method showed highest intra- and interobserver levels of agreement compared with Elip-ROI and manual traditional methods. SPECT HMR calculated on the basis of UM, LM, and CL background regions showed excellent intra- and interobserver agreement. SPECT HMR with UM resulted in highest correlation (R 5 0.91) with planar HMR compared with that with LM (R 5 0.74) and CL (R 5 0.73). The LLN of SPECT HMR with UM and that of planar HMR was calculated as 5.5 and 1.6, respectively. The normal values of SPECT-derived HMR and planar-derived HMR were correlated linearly. Conclusion: We reconfirmed the previous planar HMR threshold and determined SPECT LLN HMR for SPECT. Planar HMR can be estimated from SPECT HMR via a simple linear regression equation, allowing use of the new cardiac-dedicated SPECT camera for 123I-mIBG imaging.