Myocardial uptake of 68 Ga-DOTATATE: correlation with cardiac disease and risk factors

Introduction

Low-grade neuroendocrine neoplasms (NENs) are rare, slow-growing tumors commonly originating from the aerodigestive tract. Their behavior can be unpredictable and sets them apart from other cancers. This includes the occasional production of vasoactive amines, causing diarrhea, flushing, and other effects commonly referred to as the carcinoid syndrome (1), as well as the expression of somatostatin receptors, which allows imaging with specific radioisotopes and treatments with somatostatin analogues.

Gallium 68 DOTA-0-Tyr3-Octreotate (⁶⁸Ga-DOTATATE) positron emission tomography/computed tomography (PET/CT) (is a recently approved imaging modality for neuroendocrine tumor staging and diagnosis. The radioisotope has a great affinity for somatostatin receptor (SSTR) subtype 2 and a low affinity for SSTR4 and SSTR5 (2). Its full spectrum of utility is still not well known; while appropriate use criteria have been published (3) for neuroendocrine tumor diagnosis and treatment monitoring, multiple groups have reported radiotracer concentration in other pathologies, including benign tumors and even normal tissue (4). Myocardial uptake of DOTATATE has been shown to be reproducible, and correlated with vascular uptake (5), suggesting a relationship to inflammatory large vessel changes. Other studies demonstrated DOTATATE uptake with activated macrophages in atherosclerosis (6,7), in myocardial inflammation after infarction (8) and in sarcoidosis (9). In our institutional experience, we have noticed myocardial uptake in a subset of patients who have had ⁶⁸Ga-DOTATATE PET/CT and cardiac disease. This is something not usually quantified or reported but might have potential clinical implications if the significance and correlations are better understood. The aim of the present study was to explore the relationship between myocardial uptake of DOTATATE and cardiac disease by: (i) developing a standard methodology to semi-quantify the myocardial uptake; (ii) internally validating this measure and assessing its reproducibility among readers; and (iii) correlating this measure with patients’ cardiac disease status, and patients’ risk for coronary artery disease as per the Framingham risk score.

Material and Methods

Myocardial ⁶⁸Ga-DOTATATE uptake score (MUS) calculation and reproducibility

For the included 145 examinations, 107 scans were performed on a Siemens Biograph mCT scanner and 27 scans were performed on a Siemens Biograph TruePoint scanner; the remaining 11 scans were from outside facilities. Images were acquired from the top of skull to the mid thighs after infusion of 0.054 mCi/kg of ⁶⁸Ga-DOTATATE to a maximum of 5.4 mCi as per institutional protocol.

The MUS was measured using the average SUV mean values of five circular regions of interest (ROIs), with a radius of 1 cm, divided by the SUV mean of the myocardial blood pool measured within the left ventricular lumen by an experienced dual American Board-certified physician (Radiology and Nuclear Medicine). We used ventricular wall to blood ratio as a method for internal standardization. We used measurements of the left ventricular wall because it is easier to see on CT scan than the thinner wall of the right ventricle. The ROIs were placed over the following specific areas of the left ventricular wall: lateral mid, lateral apex, anterior mid, septal mid, and septal base (Fig. 1). In order to evaluate the reproducibility of the MUS, 18 patients were randomly selected to be reviewed by a second experienced physician, also certified in Radiology and Nuclear Medicine. In addition, the readers used a visual score for subjective assessment of myocardial uptake of ⁶⁸Ga DOTATATE scored as negative or positive, compared to the intraventricular blood pool activity, with three levels of positivity (mild, moderate, or high uptake).

OPEN IN VIEWER

Fig. 1.

Method of calculating MUS. Axial fused PET CT image in a 61-year-old man with neuroendocrine tumor and carcinoid heart disease demonstrates ⁶⁸Ga-DOTATATE uptake throughout the myocardium, higher than that of the blood pool (yellow circle). Measurements were performed at specific locations in the myocardium, as detailed in the “Material and Methods” section, shown here at the mid-lateral wall (white circle). The patient also had bone metastasis in the sternum (arrowhead). MUS, myocardial uptake score.

Results

Study population and cardiac disease subgroups

A total of 145 patients (66 men, 79 women; mean age = 56.9 ± 13.7 years) had imaging in the included time frame, five of whom had two scans, of which the most recent was included. SUV mean LV blood pool was 0.46  ±  0.19 and for MUS was 1.89  ±  0.81 (Table 1). Out of 145 patients, patients with coronary calcifications, Framingham risk score of >10% putting them at high risk of coronary artery disease (CAD), and other cardiac disease were grouped into the cardiac disease positive group (n  =  83). Out of those 83 patients, five had CaHD, but three out of five had coronary artery calcifications as well. Patients with no coronary calcifications, risk score of <10% and no other cardiac disease constituted the cardiac disease negative group (n  =  23). The remaining patients with no known cardiac disease, no coronary calcifications, and no information available to calculate Framingham risk score were classified as unknown cardiac risk (n  =  39) (Fig. 2).

OPEN IN VIEWER

Fig. 2.

Flow chart of the study participants.

OPEN IN VIEWER

Table 1.

Demographics and MUS of different groups.

	All	Positive for cardiac disease	Unknown cardiac risk	No cardiac disease
Number	145	83	39	23
Age (years)	56.9 ± 13.7	62.0 ± 10.6	51.3 ± 14.41	47.6 ± 13.7
Sex (M:F)	66:79	48:35	15:24	3:20
Framingham score available	73/145	50/83	0/39	23/23
Framingham score	15.9 ± 12.4 (0.82–74.5)	21.2 ± 11.4 (4.56–74.5)	NA	4.39 ± 2.7 (0.82–8.64)
MUS	1.89 ± 0.80 (0.66–4.54)	2.03 ± 0.83 (0.79–4.54)	1.86 ± 0.77 (0.66–3.65)	1.40 ± 0.51 (0.90–3.01)

Values are given as n or mean ± SD (range), where available.

MUS, myocardial uptake score.

MUS calculation and reproducibility

The demographics of different groups and MUS mean values with ranges are shown in Table 1. Mean age and gender distribution were significantly different between cardiac disease positive and cardiac disease negative groups. The MUS measurement reproducibility was tested in a subgroup of randomly selected 18 individuals and was highly reproducible with an ICC of 0.964 (95% CI = 0.903–0.987; P < 0.001). For individual ROI region measurements in the myocardial wall and the LV blood pool, the ICC was lower at 0.706 (95% CI = 0.597–0.789; P < 0.001). Visual score positivity between readers was concordant in 17 out of 18 (94%) patients, and agreement was high with a kappa value of 0.74 (95% CI = 0.57–0.90; P < 0.001). There was strong agreement between visual evaluation (negative, + 1, + 2, + 3) and MUS with rho  =  0.800 and P < 0.0001 (Fig. 3). As binary classification (positive or negative myocardial uptake) for the 145 patients, MUS cutoff of 1.535 correlated with positive visual score with an accuracy of 93.1% (TP85, FP 2, TN50, FN 8) and an area under the receiver operating characteristic curve (AUC – ROC) of 0.977. A close cutoff MUS of 1.52 can differentiate groups of cardiac disease positive (n = 83) from cardiac disease negative (n  =  23) with a sensitivity of 71.2%, specificity of 69.6%, and accuracy of 70.8%.

OPEN IN VIEWER

Fig. 3.

(a) MUS as related to visual score on the x-axis vs. MUS on the y axis. (b) Same box plot with binary classification into negative or positive. MUS, myocardial uptake score.

MUS correlation with cardiac disease

There was significant difference in MUS among the three groups with mean values of 1.40 for no cardiac disease (n  =  23), 1.86 for those with unknown cardiac risk (n  =  39), and 2.03 for cardiac disease (n  =  83) (Kruskal–Wallis, chi-square  =  11.7, df  =  2, and P  =  0.0029) (Fig. 4). MUS was significantly higher for those with cardiac disease than those without (Kruskal–Wallis, chi-square  =  12.3, df  =  1, and P  =  0.00045). The unknown cardiac risk group was significantly different from the patients with no cardiac disease (Kruskal–Wallis, chi-square  =  4.8, df  =  1, and P  =  0.028), but not different from those with cardiac disease (Kruskal–Wallis, chi-square  =  0.76, df  =  1, and P  =  0.38).

OPEN IN VIEWER

Fig. 4.

Myocardial uptake of ⁶⁸Ga-DOTATATE was higher in patients with carcinoid heart disease, as well as other cardiac disease, as evaluated by MUS. Axial fused PET/CT images demonstrate (a) no significant uptake of ⁶⁸Ga-DOTATATE within the myocardium (white arrow) when compared to the blood pool (white circle) in a 62-year-old woman with no known cardiac disease and low Framingham risk score of 7.9% (MUS of 1.0, visual score negative) compared to (b) diffuse myocardial uptake with increased tracer uptake in the myocardium (white arrow) relative to blood pool (white circle) in a 69-year-old man with severe three-vessel CAD, five prior coronary stents, and coronary artery bypass surgery (MUS 3.4, visual score positive). CAD, coronary artery disease; MUS, myocardial uptake score.

Analyzing the population of 145 patients, there was weak correlation between MUS and coronary artery calcifications (Table 2) (Spearman rank correlation coefficient rho  =  0.18; P  =  0.038). After excluding patients who were cardiac disease positive with no coronary calcifications (n  =  23) and patients with unknown cardiac risk (n  =  39), the correlation and the significance became stronger (Spearman rank correlation coefficient rho  =  0.27; P  =  0.011) in the remaining patients (n  =  83), as seen in Table 3. As binary classification of coronary calcification presence, the entire population of 145 patients had rho  =  0.20 with P  =  0.016 and a subgroup of 83 patients excluding cardiac patients without calcifications and unknown risk category had rho  =  0.40 with P  =  0.00017.

OPEN IN VIEWER

Table 2.

MUS as related to severity of coronary artery calcifications.

Calcification score	N (total  =  145)	Mean MUS	Range of MUS
0	85	1.75	0.66–4.12
1	35	2.13	0.81–4.35
2	20	1.89	0.93–4.54
3	5	2.47	1.32–3.37

MUS, myocardial uptake score.

OPEN IN VIEWER

Table 3.

MUS as related to severity of coronary artery calcifications after excluding cardiac patients with no calcifications, and patients with unknown cardiac status.

Calcification score	N (total  =  83)	Mean MUS	Range of MUS
0	23	1.40	0.90–3.01
1	35	2.13	0.81–4.35
2	20	1.89	0.93–4.54
3	5	2.47	1.32–3.37

MUS, myocardial uptake score.

Discussion

In this paper we have developed a reproducible score with high inter-observer concordance for measuring myocardial uptake of ⁶⁸Ga-DOTATATE on PET/CT scan and identified a higher level of uptake in patients with cardiac disease or risk factors compared to those with low risk for cardiac disease, concluding that values >1.52 are suggestive of cardiac disease. Further, we have shown that patients with no known cardiac disease have consistently low MUS, while those with established or risk factors for cardiac disease had significantly elevated MUS. Since the majority of patients with low grade NENs will undergo a ⁶⁸Ga-DOTATATE PET/CT scan as part of their staging and treatment, we feel that MUS has the potential to identify or confirm suspicion of cardiac disease, including CaHD, in this ever-growing population. Our results are in agreement with current literature that suggests that atherosclerosis and myocardial inflammation can demonstrate DOTATATE uptake (6–9).

Myocardial uptake in high-risk patients is likely reflecting subclinical inflammation in the cardiac muscle and/or proliferating endothelium. The radiotracer is known to highlight inflammation in general, such as in cases of inflammatory arthritis, radiation-related damage, and abscesses or other infections (11) because most white cells express SSTRs. Our results align well with previous research demonstrating SSTR tracer uptake in patients with active myocarditis or sub-acute myocardial infarction (12), probably due to SSTR overexpression in activated macrophages or damaged proliferating endothelium. The connection between various degrees of inflammation and CAD is an ongoing topic of interest. Patients with inflammatory disorders have an increased incidence of CAD and there is a subset of patients with no evidence of arterial occlusion who still develop ischemic heart disease. The CANTOS trial (13) used a monoclonal antibody targeting interleukin−1 beta therapy in patients with an elevated high-sensitivity C-reactive protein (hsCRP) and myocardial infarction (MI) and was effective at preventing adverse cardiac events over a median of 3.7 years. It is not clear if prior cardiac disease and inflammation can adequately explain the elevated MUS in our patients with CaHD given the small number of such patients. Another consideration for the scan positivity would be intracardiac metastasis from neuroendocrine tumors, but this is a rather rare phenomenon, occurring in <5% of patients with CaHD (14), with a very different imaging appearance demonstrating intense focal uptake (Fig. 5) rather than diffuse myocardial uptake. In our study, we did not encounter such lesions, but the example is shown in order to demonstrate the vast difference in imaging appearance. Another explanation for myocardial uptake in the majority of patients is a physiologic low level of tracer giving lower blood pool SUV, which would require a higher cutoff for the MUS. This is felt to be less likely given both low tracer uptake in the group with no cardiac risk factors and expected high prevalence of heart disease in our study population with a mean age of 56.9 years and other co-morbidities.

OPEN IN VIEWER

Fig. 5.

Intracardiac metastasis of neuroendocrine neoplasm. A 67-year-old man with metastatic low-grade neuroendocrine tumor of the terminal ileum status after ileocecal resection in 2004, presented with biopsy-proven metastatic liver lesions in 2019. (a) Axial fused ⁶⁸Ga-DOTATATE PET/CT images demonstrate an intense focus of tracer uptake along the inner aspect of the left ventricular lateral wall with activity greater than liver background consistent with a metastatic deposit (arrowheads) as also seen on (b) cardiac magnetic resonance imaging.

The present study has some limitations. They include its retrospective nature and referral bias – a confirmatory study in a separate validation cohort is being planned. Patients with neuroendocrine tumors have potential impact on the heart from excess serotonin and other vasoactive substances (15). Even though the sample size was significant, there was little representation of patients with no known heart disease or risk factors for cardiac disease. Nevertheless, even with this small number of patients negative for cardiac disease, the differences in MUS were significant, suggesting an underlying phenomenon. We did not account for variability in scanners and reconstruction algorithms that might affect MUS accuracy, although the majority of imaging, 134 of 145 scans (92%), was done in our institution and the way to acquire data is standardized between locations in our tertiary center. Another technical limitation is the proximity of the heart to areas of high tracer uptake in the upper abdomen (liver and spleen), which might have interfered with quantification, especially for the inferior wall; this is why inferior wall measurements were not included in MUS. In order to allow for a ubiquitous score that can be obtained quickly without excessive postprocessing, measurements were done on the axial acquired images and not along the cardiac axis, although measurement along the cardiac axis might have been more robust to prevent influence of the cardiac angle in the chest. Another limitation is the difference in age and gender distribution between the two groups of patients with cardiac disease, which is expected given that age and male gender are risk factors for CAD. We did not account for variations in the uptake time of the tracer or timing of the study relative to the last administered dose of long-acting somatostatin analogues (SSAs), which might have affected the DOTATATE max uptake levels (16). That being said, Ga-DOTATATE scans are no longer timed with SSA administrations for most institutions. We elected to use a ROI rather than a volume of interest in order to ensure that the ROI is indeed overlying the myocardium and excluding any adjacent organs or lesions. SUV mean was preferred to SUV max because we wanted to minimize the effect of noise on the overall measurement. SUV peak was not elected due to the relatively small size of the ROI, although potentially it can provide more accurate estimates. An additional limitation is that coronary artery calcifications were evaluated subjectively due to the technical limitations imposed by available attenuation correction CT images and not by a rigorous method such as a formal calcium score, although subjective evaluation of coronary calcifications have shown significant correlation with death from cardiovascular disease (17). Another limitation is that detailed cardiac health information was not known for all of our patients and we do not have prospective, multi-year cardiac data; therefore, we are not sure if some of our current patients with high MUS will develop CAD or other cardiac disease in the future. Moreover, given the relatively new approval of this imaging modality, we have not had time to longitudinally follow-up the patients for subsequent development of cardiovascular events. Nevertheless, our results still did show a clear and statistically significant difference in uptake between patients known to be without and those with heart disease.

There are numerous areas for future investigation along the lines of this study, including standardizing MUS across scanners, correlating myocardial uptake with timing of SSA administration, prospective and long-term follow-up of patients with elevated MUS with serum markers, such as pro-BNP and troponin levels, and a detailed cardiac evaluation when needed.

In conclusion, we have introduced a simple score for evaluating myocardial uptake of ⁶⁸Ga DOTATATE, proved its concordance with subjective visual evaluation, and demonstrated that myocardial uptake of this tracer might indicate underlying cardiac disease or high-risk factors for cardiac disease at a cutoff level of 1.52. Our results need further consolidation by additional larger studies, and if proven to be reproducible, can be easily translated into clinical care.