Introduction

Hereditary transthyretin (TTR) amyloidosis is a systemic disease due to a mutation of the gene responsible for the synthesis of TTR [1]. This results in synthesis of an unstable form of TTR, which unfolds and results in monomer, hexamer, and amyloid fiber accumulation in the peripheral nervous system, the heart, the kidneys, and the eyes, with a preferential organ affinity depending on the mutation. More than 100 mutations have been described. Liver transplantation [2, 3] is the reference treatment for the neuropathy, but new treatments such as TTR stabilizers, antisense oligonucleotides, or small interfering RNA are under development [4,5,6,7]. As amyloid deposits result in nearly irreversible end-organ damage, future patient care will require both an efficient treatment and a very early diagnosis of organ involvement, before severe dysfunction. This is particularly true with regard to cardiac involvement [8], combining infiltrative cardiomyopathy (myocardial thickening by amyloid deposits, restrictive heart failure with preserved ejection fraction), conductive abnormalities, and cardiac sympathetic and parasympathetic denervation, which is of major prognostic value [9]. The diagnosis of cardiac amyloidosis (CA) among patients with genetically proven TTR mutation relies either on a positive cardiac biopsy, or on the association of a positive peripheral biopsy and evidence of a cardiac infiltration using multimodality imaging. Therefore, there is a great need to develop and evaluate very sensitive non-invasive techniques for the early diagnosis of cardiac TTR amyloidosis [5, 6, 10]. Both morphological and functional imaging modalities, such as echocardiography with strain imaging and cardiac magnetic resonance imaging, are used, along with scintigraphy, in a multimodality imaging process to achieve an early and accurate diagnosis [11,12,13]. Most studies have evaluated the role of diphosphonate scintigraphy, considering its high diagnostic performance and its prognostic value [14,15,16,17,18]. MIBG scintigraphy was only recognized as a major prognostic factor [19, 20], while its diagnostic value was not considered.

The aim of our study was to evaluate the respective contribution of 123I-MIBG (MIBG) and 99mTc-diphosphonates (DPD) scintigraphies in the diagnosis of cardiac TTR amyloidosis in consecutive patients with genetically proven TTR mutation, either spontaneously or after domino liver transplant (with histological and genetic TTR mutation proof in the donor).

Material and methods

Population

We prospectively included 75 consecutive patients undergoing assessment of cardiac involvement of amyloidosis in the French Reference Center for Amyloidosis (CRMR-NNERF) between October 2011 and June 2015, which underwent both scintigraphies in the nuclear medicine department. Inclusion criteria were as follows: documented genetic TTR mutation or previous domino liver transplantation (DLT) [21], clinically symptomatic or asymptomatic carrier with regard to cardiac, neurologic, or ophthalmic involvement. Patients referred for cardiac evaluation of suspected wild-type (senile) or other types of amyloidosis (AL, AA, ApoA1) were not included. All patients underwent clinical evaluation of their neurological and cardiac condition, including ECG, echocardiography [interventricular septum (IVS) thickness, left ventricular ejection fraction (LVEF), left ventricular filling pressure estimated by the E/e’ ratio and global longitudinal strain (GLS)], and evaluation of BNP (normal ≤ 80 pg/l) and troponin T (normal ≤ 0.05 μg/l) serum levels. History of disease-modifying therapies such as liver transplantation, transthyretin stabilizers (tafamidis) or inclusion in prospective studies with patisiran or inotersen were collected.

This study was in compliance with the ethical principles formulated in the Declaration of Helsinki, and all patients granted their written, informed consent to participate in the registry (Commission Nationale de l’Informatique et des Libertés n°1,470,960).

Nuclear imaging procedures

MIBG scintigraphy imaging, definition of normal and threshold values

Drugs suspected of interfering with MIBG uptake were withheld for an adequate period of time before imaging. As previously described [20], MIBG scintigraphy consisted of planar chest anterior and posterior view acquisitions (10-min acquisition on a Symbia T2, Siemens, Erlangen Germany; low-energy high-resolution collimators; matrix size 128 × 128; energy window of 20% centred on the 159-keV photopeak of 123I) 20 min and 4 h after intravenous injection of 3 MBq/kg of 123I-MIBG (after thyroid blockade) and tomoscintigraphic acquisition at 4 h on a CZT camera (D-SPECT, Spectrum Dynamics, Palo Alto, CA, USA; 12 min acquisition). The scan was associated with a myocardial perfusion tracer injection (201thallium) at rest with tomoscintigraphic acquisition on a CZT camera to assess innervation/perfusion mismatch.

Regions of interest were manually drawn on the left ventricle and the upper mediastinum areas, and cardiac MIBG uptake was calculated as the heart-to-mediastinum ratio (HMR). The normal HMR threshold assessed on planar images acquired 4 h after injection (referred to as HM4) has been previously determined in a cohort of 49 TTR mutation carriers (age: 45 ± 14 years, 56% females) without cardiac involvement at clinical, biological, and echocardiographic evaluation on initial presentation and during the 2 years following MIBG scintigraphy. Their mean HM4 was 2.15 ± 0.15, so cardiac denervation was defined by HM4 < 1.85 (e.g., 2.15 minus 2 SD).

We also determined a lower threshold suggestive of increased risk of adverse events. This threshold was set at 1.4 based on consistent results of two studies. In a TTR amyloidosis population in whom Val30Met mutation was the most frequent, Algalarrondo et al. [20] reported an increased death rate after transplantation in cases of HM4 < 1.4 (5-year survival rate of 64% if HM4 < 1.4 vs 93%, p < 0.0001). Similarly, Coutinho et al. [19] reported a dramatic increase of 5-year mortality rate (from < 20% to 50%) when HM4 was < 1.4 in an exclusively Val30Met population. Hence, three subgroups of patients were considered: subgroup A (HM4 ≥ 1.85) with no denervation, subgroup B (1.85 > HM4 ≥ 1.4) with moderate denervation, subgroup C (HM4 < 1.4) with severe denervation. In addition, a segmental analysis of MIBG uptake was performed on SPECT images, in order to determine the location and the extent of sympathetic denervation. As previously described [22], the number of denervated segments was visually graded on a 17-segment model using a 5-level scoring scale (0: normal, 4: no uptake). The total severity score index (SSI) was calculated by summing the scores of individual segments.

Diphosphonate scintigraphy

Diphosphonate scintigraphy consisted of planar thoracic anterior and posterior view acquisition (Symbia T2; Siemens, Erlangen Germany; low-energy high-resolution collimators; matrix size, 256 × 256; energy window of 20% centred on the 140-keV photopeak of 99mTc) at 3 h after tracer injection (9.5 MBq/kg of 99mTc-DPD [99mTechnetium-3,3,-diphosphono-1,2- propanodicarboxylic acid]), followed by a whole-body planar acquisition and a cardiac tomoscintigraphic acquisition.

Whole-body DPD uptake pattern was graded using the Perugini score [23], with a grading scale ranging from 0 to 3 (0: no cardiac uptake, scintigraphy considered as negative; grade 3: strong cardiac uptake with mild/absent bone uptake) [14]. Since non TTR-related amyloidosis had been ruled out before enrolment, grades 1 to 3 were considered as positive for cardiac amyloidosis deposits. Heart to lung ratio (H/L) was calculated as the ratio between 3D isocount volume of interest generated over the myocardium and a standard volume in right pulmonary base on SPECT images.

Statistical analysis

Continuous variables were expressed by their mean ± standard deviation and compared by use of Student’s t test or ANOVA followed by the Scheffé’s post-hoc test for comparison of more than two groups. Categorical variables were expressed by their percentage and compared by use of Fisher’s exact test. Independent predictors of MIBG and DPD cardiac uptake were determined using logistic regression with stepwise selection of variables. In the multivariate model, we did not include the parameter “amyloidosis targeted therapy”, since therapy initiation was considered upon a clinical appraisal of the global severity of the disease, dependent on the other patient’s characteristics. Linear regression and Pearson correlation coefficients were performed to determine the relation between continuous variables. Statistical analysis was performed using SPSS (IBM) and Medcalc software version 16.4.3 (Ostend, Belgium). A p value < 0.05 was considered as significant.

Results

Study population

A total of 75 consecutive patients were included in the study (Table 1). The most frequent mutation was Val30Met (32 patients, 43%), then Ser77Tyr and Ser77Phe accounted for seven patients each. Ten patients had a history of domino liver transplantation. Most patients were symptomatic (44 patients, 59%), predominantly with neurological symptoms.

Table 1 Baseline characteristics and results of MIBG and DPD scintigraphy
Full size table

Nuclear imaging findings

The delay between MIBG and DPD scans was 6 ± 12 days. In the overall population, MIBG was abnormal in 49 patients (65%), whereas DPD was positive in only 29 (39%) (Table 2). When MIBG was normal, DPD was negative except for two patients (Ala36Pro and Arg34Gly mutations) presenting with grade 3 pattern (other patients’ characteristics are detailed in Fig. 1). The relationship between quantitative parameters of MIBG and DPD uptake are presented in Fig. 2.

Table 2 Scintigraphic results in the whole population (n = 75)
Full size table
Fig. 1
figure 1

Example of MIBG and DPD reverse pattern, in a 41-year-old woman with Ala36Pro mutation and ocular involvement. Biological markers were normal (BNP: 9 pg/ml; troponin: 0.04 μg/ml), echocardiography showed interventricular septum thickness of 8 mm, E/e’: 5.7, global longitudinal strain: −19, and left ventricular ejection fraction was 59%. MIBG scintigraphy was completely normal with HM4 > 1.85 and the absence of segmental defect, but DPD showed a marked uptake with grade 3. The second patient with a reverse pattern was a 60-year-old man with Arg34Gly mutation, had ocular involvement, left ventricular hypertrophy (interventricular septum thickness: 14 mm) but normal E/e’, global longitudinal strain, and left ventricular ejection fraction. Despite normal MIBG scintigraphy (HM4: 1.9, WOR: 34), DPD was grade 3 with H/L: 13.95

Full size image
Fig. 2
figure 2

Relationship between MIBG (HM4) and DPD uptake (H/L ratio) in the study population. There is an inverse relationship between HM4 and H/L (Pearson coefficient R = −0.41, p < 0.001)

Full size image

With regard to rest myocardial perfusion scintigraphy, no patient presenting denervation on MIBG had significant perfusion defect in the corresponding denervated segments, therefore confirming perfusion/innervation mismatch.

We performed a visual regional analysis of MIBG and DPD abnormalities and found different abnormality patterns: whereas MIBG demonstrated preferential abnormalities in the apex and the inferior wall, DPD uptake was most intense in the septum with a relative apical-sparing, as presented in Fig. 3 (and Fig. 1 Supplementary Material) and Table 1 Supplementary material.

Fig. 3
figure 3

Results of visual regional analysis of MIBG (a) and DPD (b) scintigraphic abnormalities using 5-point scales (a, for MIBG: 0, normal uptake; 1, mildly reduced uptake; 2, moderately reduced uptake; 3, severely reduced uptake; 4, no uptake; b, for DPD: 0, no uptake; 1, low intensity uptake; 2, low to moderately intense uptake; 3, moderate uptake; 4, intense uptake); with examples of diffuse MIBG (c) and DPD (d) abnormalities in a patient with severe denervation and intensely increased cardiac DPD uptake; and mismatched MIBG (e) and DPD (f) abnormalities in another patient with moderate denervation and moderate intensity DPD cardiac uptake

Full size image

Predictors of scintigraphic abnormalities

Table 1 shows that age, amyloidosis targeted therapy, the presence of symptoms, increased BNP level, IVS thickness, altered GLS and E/e’ were associated with both MIBG abnormalities and DPD positive uptake, whereas troponin T level was lower in patients with normal MIBG, and LVEF was significantly decreased in patients with positive DPD. Patients with normal MIBG had no or little evidence of cardiac amyloidosis, whereas patients with positive DPD presented markers of biological and functional cardiac dysfunction. On multivariate analysis (Table 3), in addition to patients’ age that was a common determinant of abnormal finding both with MIBG and DPD, E/e’ was an independent predictor of abnormal MIBG, and altered GLS was an independent predictor of DPD positive uptake.

Table 3 Independent predictors of abnormal findings on scintigraphy (multivariate analysis)
Full size table

Semi-quantitative evaluation of MIBG and DPD uptake intensity are provided in Figs. 4 and 5. As shown in Figs. 4 and 5, before the age of 50 only two patients presented DPD uptake, whereas ten had mild cardiac denervation on MIBG and none presented severe denervation.

Fig. 4
figure 4

Combined analysis of the semi-quantitative uptake intensity of MIBG (a) and DPD (b) according to age and E/e’. Severe denervation is absent in patients < 50 years old and preserved E/e’, while mild denervation may be present. There was no significant correlation between either E/e’ and HM4 or E/e’ and H/L (Pearson coefficient R respectively = −0.49 and −0.39, p < 0.0001 and = 0.0005)

Full size image
Fig. 5
figure 5

Selected parameters illustrating cardiac dysfunction according to semi-quantitative uptake intensity of MIBG (column a) and DPD (column b). * p < 0.05 vs the two other groups. # p < 0.05 vs the two other groups (except Perugini 1). There were significant differences between the MIBG subgroups (column a) with regard to BNP and E/e’ (subgroups HM4 ≥ 1.85, 1.85 > HM4 ≥ 1.4 and HM4 < 1.4 respectively; BNP = 28 ± 43 vs 102 ± 120 vs 208 ± 165, p = 0.0001; and E/e’ = 7.2 ± 4.2 vs 11.6 ± 6.8 vs 14.2 ± 5.6, p< 0.0001) but no differences with regard to LVEF (65 ± 7 vs 63 ± 9 vs 60 ± 10, p = 0.2)

Full size image

Subgroup analysis

Asymptomatic patients

In the subgroup of asymptomatic patients (n = 31), MIBG was abnormal in 15 (48%). DPD was always negative when MIBG was normal, while DPD was positive in eight out of 15 patients with abnormal MIBG (Table 4). Although asymptomatic patients with abnormal MIBG scintigraphy were older than asymptomatic patients with normal MIBG, patients’ age did not make it possible to differentiate those at higher risk of abnormal scan because of an important overlap. The same held true for biological markers and echocardiography findings (Table 4). Regional analysis revealed significant differences of MIBG denervation between asymptomatic and symptomatic patients with more pronounced denervation among symptomatic patients, with a preferential infero-latero-apical denervation pattern in both groups (Table 2 Supplementary Material).

Table 4 Asymptomatic patients (n = 31)
Full size table

Domino liver transplanted patients

Although domino liver transplanted patients represent a specific subgroup with a shorter exposure time to mutated TTR than carriers of an hereditary mutation, we compared our results in this subgroup and the rest of the population and found no significant differences with regard to scintigraphic results, except for HM20, which was significantly greater in this subgroup (Table 3 Supplementary Material).

MIBG subgroups

The three subgroups defined by the HM4 thresholds presented significant differences with regard to age, symptoms, BNP, echographic findings, positivity of diphosphonates scintigraphy, and H/L: as the HM4 decreased, clinical, biological and echographic parameters were more severely or more frequently impaired (Table 4 Supplementary Material).

Discussion

The present study shows that in TTR mutation carriers, cardiac sympathetic denervation evidenced by decreased MIBG uptake is detected earlier than amyloid burden evidenced by DPD. At the presymptomatic stage, carriers with a normal MIBG scan seem to have a very low likelihood of cardiac involvement, with only a few patients with mild morphological or functional cardiac alterations.

In our study population, a positive DPD uptake was associated with a rather advanced stage of CA. First, age was a major determinant of DPD uptake, with only two patients presenting a detectable uptake below 55 years old. Increased LV filling pressure (E/e’) was a predictor of abnormal MIBG, while DPD uptake was associated with markers of mild LV systolic dysfunction such as BNP increase, GLS alteration, and LVEF impairment. Diagnostic performances of diphosphonate scintigraphy for TTR amyloidosis have already been demonstrated [14,15,16], but only a few studies have evaluated its value in asymptomatic patients [18, 24]. Many studies considered a septum thickness ≥ 12 mm as a diagnostic criterion for CA, while this cutoff value does not take into account the existence of a “gray zone”. Accordingly, the sensitivity of DPD in cases of cardiac amyloid deposits with septum thickness < 12 mm is not well elucidated. Only rare false-negative diphosphonate exams have been reported, mainly attributed to the type of involved fibrils or to the tracer used [25, 26]. Several scores using semi-quantitative ratios have been described, such as heart/whole body ratio [27,27,28,30], heart/skull ratio [27, 31] and two studies ing heart to lung ratios [16, 32], with one [16] finding a correlation between 5-year all-cause mortality rate and heart to lung ratio. The value of a semi-quantitative ratio such as H/L will need to be evaluated in the future.

The prognostic value of MIBG scintigraphy in TTR amyloidosis is well established [19, 20]. However, few studies [19, 33] have evaluated cardiac denervation in asymptomatic carriers or in TTR patients with only echographic abnormalities. In our study population, a mild decrease of MIBG uptake was associated with very early signs of CA, including in the subgroup of asymptomatic carriers. The prognostic significance of a mild decrease of MIBG uptake, without DPD abnormality will need to be assessed in further studies evaluating the occurrence of cardiac events and the potential progression of functional abnormalities. When abnormal, the HM4 decreased with more advanced disease, in keeping with the well-established poor prognosis of patients with severe denervation.

A preliminary study combining MIBG and bone scintigraphy has been performed in 12 patients with symptomatic TTR amyloidosis [34]. Although the MIBG uptake was assessed only qualitatively, the authors reported severe global or segmental denervation in all patients, while bone scintigraphy was positive in only four of them. Interestingly, ten out of 12 patients had had endomyocardial biopsy showing histologic evidence of amyloid deposition, suggesting the lack of sensitivity of bone scintigraphy in this population. Our study confirms the presence of a significant number of patients (nearly one third of the study population) with abnormal MIBG and negative DPD in a greater population sample, and demonstrates that this pattern is even more prominent at the presymptomatic stage, highlighting the potential role of MIBG as an early diagnostic marker of cardiac amyloidosis. In addition, the increase of MIBG uptake associated with regression of autonomic dysfunction after treatment by small interfering RNAs has been reported in a small sample study [35]. Taken together, these results suggest that MIBG could be used in decision-making for early initiation and monitoring of disease-modifying therapies.

Regional analysis of denervation on MIBG and of DPD uptake revealed different patterns, with a predominant denervation of the apex and inferior segments, and a relative apical-sparing pattern on DPD as previously described [36]. Those findings support the hypothesis that two different pathophysiological mechanisms are observed, cardiac denervation with MIBG versus myocardial amyloid deposition, with different locations. The regional DPD uptake is consistent with apical sparing of longitudinal strain observed with echocardiography. The regional MIBG results can be explained on the basis of anatomical and histological studies, demonstrating that sympathetic nerve distribution is heterogeneous among the left ventricular walls, with a preferential distribution among basal segments versus apical segments [37,37,38,39,40,42]. A study evaluating myocardial muscarinic receptor density with PET in TTR amyloidosis patients [43] found that FAP patients are also affected by parasympathetic myocardial denervation in a heterogeneous manner, as compared to parasympathetic tone decrease in heart-failure patients. Therefore, we can hypothesize that denervation will occur earlier in the lesser innervated regions, as observed in our study. Those findings suggest that not only do those two pathological mechanisms occur at different stages of the disease, but that they evolve independently.

On regional analysis, we found that two asymptomatic patients presented with segmental denervation on tomographic images, whereas HM4, HM20, and MIBG wash-out rate were within normal ranges. Those patients were considered as true negative of MIBG scintigraphy based on planar findings; however, further longitudinal studies evaluating the evolution of these patients could reveal that tomographic acquisition is more sensitive to detect beginning denervation.

The correlation between age and IVS thickness, and between scintigraphic semi-quantitative parameters (HM20, HM4, MIBG wash-out rate, H/L) and IVS thickness, suggests the role of the exposure duration and amyloid burden in the positivity of scintigraphic examinations. Correlation between diphosphonate scintigraphy and IVS thickness had already been described [27, 28, 31, 44].

Interestingly, two patients presented with a reverse scintigraphic pattern consisting of normal MIBG and markedly positive DPD. Although one of them had mildly increased septum thickness, none had increased filling pressure or abnormal biological markers. Both had very rare mutations: Arg34Gly [45] and Ala36Pro [46,46,47,48,50] presenting with ophthalmic symptomatology, without cardiac nor neurologic symptoms.

Study limitations

Endomyocardial biopsy, which is the gold standard for cardiac amyloidosis, is an invasive and potentially harmful technique and is not performed on a routine basis in our center, and not ethically justified when patients are asymptomatic with normal multimodality imaging. However, the main purpose of our study was to compare the chronological relation between nuclear imaging approaches rather than to compare their diagnostic abilities which have already been reported [14, 15, 19, 20]. Also, MRI findings were not included in the characterization of CA due to substantial methodological and industrial changes throughout the study duration. In the future, advanced MRI sequences should be part of a comprehensive evaluation of CA.

We acknowledge that the results presented here may not apply in a different study population, particularly for TTR gene mutations associated with exclusively cardiac involvement such as Val122Ile. However, based on the results of the THAOS registry our study population matches with the reported distribution of mutations in patients investigated in centers outside the United States with a predominance of Val30Met [51].

Conclusion

The present study shows that in TTR mutation carriers, cardiac sympathetic denervation evidenced by decreased MIBG uptake is detected earlier than amyloid burden evidenced by DPD, although both are related to the patient’s age. MIBG and diphosphonate scintigraphy are complementary non-invasive explorations that allow for diagnosis and evaluation of cardiac involvement in hereditary TTR amyloidosis. These results raise the possibility of a diagnostic role for MIBG scintigraphy at an early stage of cardiac involvement in TTR mutated carriers, in addition to its well-established prognostic value. Conversely, DPD uptake, strongly related to patients’ age and associated with mild LV systolic dysfunction, seems to be a marker of disease severity.