Dear Sir,

Bellevre et al. used D-SPECT (Spectrum Dynamics, Israel) with a cadmium zinc telluride-based camera for the determination of the heart-to-mediastinum ratio (HMR) of myocardial 123I-metaiodobenzylguanidine (MIBG) uptake [1]. Since some institutions have changed their SPECT equipment from Anger SPECT to D-SPECT for cardiac studies, conventional planar images are not available in the routine clinical setting. Although the European Association of Nuclear Medicine (EANM) Cardiovascular Committee and the European Council of Nuclear Cardiology have recommended the use of medium-energy (ME) collimators for 123I-MIBG studies [2], a number of hospitals still use low-energy (LE) collimators, and SPECT MIBG HMRs have been added recently.

While the 123I-MIBG study has been widely used to patients with chronic heart failure and Lewy body disease, most studies have used planar HMR [3–5]. Whether we can still use the previous numerous stock of data and how we can connect the follow-up data in the same series of patients are important issues. Considering the exchangeability of data between the Anger-type camera and D-SPECT, the authors of this study successfully used an anterior view equivalent planogram and introduced a correction formula to standardize the data between the two systems [1].

We have proposed the use of standardized HMR to the condition obtained with the most common ME general purpose collimators for all 123I-MIBG studies irrespective of the collimator used in an individual hospital [6, 7]. The conversion coefficient (CC) for the mathematical reference value using the most popular ME collimators was CC = 0.88 [7]. Therefore, an institutional HMR (HMRi) can be converted to the standard HMR (HMRstd) using the following equation:HMRstd = 0.88/CCi × (HMRi − 1) + 1, where CCi is the CC of the institutional camera/collimator system. The CC ranged from 0.5 to 1.0 depending on the collimator type used [1].

However, to extend the standardization of planar HMR to SPECT HMR, conversion between SPECT and planar studies again becomes important. We therefore recalculated the data of Bellevre et al. [1], assuming a LE high-resolution collimator with a CC of 0.55 (average CC of LEHR collimators) [7]. The relationships between D-SPECT HMR, the corrected HMR using the correction formula of Bellevre et al. and the standardized HMR based on the phantom method using the above equation are shown in Fig. 1. As found by Bellevre et al. [1], the regression line comparing the Anger camera HMR and D-SPECT HMR shows a slope of 1.64. After correction using the correction equation of Bellevre et al. the slope becomes 0.96 [1]. Based on our proposed conversion to standardized HMR, the slope becomes 1.02 with an intercept of −0.13 (Fig. 1c).

Fig. 1
figure 1

a, b Linear regression analysis of the relationship between Anger camera HMR and D-SPECT HMR (a) and between D-SPECT HMR corrected and Anger camera HMR (b) using the data of Bellevre et al [1]. c, d Linear regression analysis of the relationship between standardized HMR (CC 0.88) and D-SPECT HMR (c) and between D-SPECT HMR + 0.1 (d). The dotted lines indicate HMR 1.6 and 1.9. The shaded areas denote confidence limits for the regression lines

Full size image

Using the correction D-SPECT HMR +0.1 gives a standardized HMR nearly equal to the HMR obtained using the ME collimator (CC 0.88). Since the HMR using the LEHR collimator of 1.6, which was used in the ADMIRE-HF study [3], can be converted to 1.9 using ME collimators, the corresponding dotted lines are shown in Fig. 1d, and indicate good concordance between conventional Anger camera and D-SPECT. Before generalizing the relationship for HMRs between D-SPECT and Anger camera, we need to know the CC of individual institutions and the reproducibility of planogram-based D-SPECT HMR among institutions, which might partly depend on preferences as to acquisition and processing conditions. However, calibrations between D-SPECT and Anger cameras are promising, and all the HMRs, from either planar imaging or SPECT, can be converted to comparable ME values as recommended by the EANM Cardiovascular Committee and the European Council of Nuclear Cardiology [2]. A larger amount of comparative data are required for validation of this method. The phantom-based calibration of HMR has already been performed in Japan in approximately 500 hospitals for standardization, and has just been started in some European countries.

Standardization is also important for identifying appropriate thresholds to differentiate good and bad prognoses [3–5], as well as for incorporating 123I-MIBG HMR into cardiac mortality risk models [8]. This standardization would also help create large databases in Europe, North America and Japan.