Carotid plaque inflammation and calcification on somatostatin receptor PET/CT imaging predict stroke and major adverse cardiovascular events

Abstract

Purpose

Carotid atherosclerosis is a risk factor for stroke and major adverse cardiovascular events (MACE) and is characterized by inflammation and calcification at different stages. Simultaneous assessment of both features may improve risk stratification. We therefore evaluated whether combined assessment of carotid plaque inflammation and calcification on [⁶⁸Ga]DOTA-TOC PET/CT predicts stroke and MACE in a real-world oncologic population.

Methods and results

We retrospectively included patients who underwent [⁶⁸Ga]DOTA-TOC PET/CT for suspected or confirmed neuroendocrine tumours. Carotid plaques were assessed visually and semiquantitatively on PET/CT, and patients were categorized into four groups according to the patient-level presence of carotid artery calcification and/or focal [⁶⁸Ga]DOTA-TOC uptake. Median follow-up was 55 months (IQR 42–72). Stroke and MACE, defined as myocardial infarction, hospitalization for myocardial ischaemia, coronary revascularization, and all-cause death, were recorded. A total of 353 patients were evaluated. Carotid artery calcification was present in 123/353 (34.8%) patients, with a median calcification volume of 77.7 cm³ (IQR 27.9–169.8). Forty-six patients (13.0%) showed 67 foci of [⁶⁸Ga]DOTA-TOC uptake, with median SUVmax 1.68 (IQR 1.55–1.93) and median TBR 2.14 (IQR 1.65–2.79). Patients with both calcified and inflamed carotid plaques had the highest rates of stroke (15.2%) and MACE (12.1%) vs. all other groups (P < 0.05). In multivariable Cox analysis, this combined patient-level phenotype remained independently associated with stroke and MACE and remained significant in competing-risk analyses.

Conclusion

[⁶⁸Ga]DOTA-TOC PET/CT enables simultaneous assessment of macrophage-related carotid plaque inflammation and calcification and identifies a subgroup at particularly high risk of stroke and MACE.

Graphical Abstract

Graphical Abstract.

See the editorial comment for this article ‘Is plaque inflammation the missing link for cardiovascular risk prediction?’, by M. Peverelli and J.M. Tarkin, https://doi.org/10.1093/ehjci/jeag137.

Background

Cardiovascular diseases remain the leading cause of death worldwide, accounting for nearly one-third of all global deaths.¹ Among these, 81.8% are caused by ischaemic heart disease and stroke.^1,2 Atherosclerotic plaques involving carotid arteries are not only a well-established risk factor for stroke but also a surrogate marker of systemic atherosclerosis and a predictor of major adverse cardiovascular events (MACE).³

Atherosclerosis is a complex, lipid-driven chronic inflammatory arterial disease whose progression is regulated by multiple factors and involves various pathological processes, such as vascular endothelial dysfunction, lipid deposition, inflammation and oxidative stress, and arterial calcification.⁴ A comprehensive evaluation of carotid plaque morphology, composition, and molecular characteristics can predict plaque stability and disease progression, providing significant value in risk stratification, prognosis prediction, and personalized clinical decision-making.^3,5 Hence, non-invasive imaging of the carotid artery for early identification of high-risk features of atherosclerosis provides vital information for assessing stroke and overall cardiovascular risk.

Among numerous imaging targets, arterial calcification represents a characteristic feature of advanced atherosclerotic lesions and serves as a reliable indicator of overall plaque burden.⁶ In the PROSPECT study, total plaque burden was identified as the strongest predictor of MACE,⁷ although in the early stages of atherosclerosis, plaques may be non-calcified.⁶ On the other hand, active inflammation plays a key role in the vulnerability of vascular plaques and in the occurrence of complications, persisting throughout the entire course of the disease.⁸ As such, integrating information on both calcification and inflammation may enable a more accurate assessment of the severity of atherosclerosis, thereby further enhancing risk stratification.

In this regard, [¹⁸F]Fluorodeoxyglucose ([¹⁸F]FDG) positron emission tomography/computed tomography (PET/CT) is a sensitive modality to identify inflamed plaques, but lacks specificity in identifying activated macrophages.⁹ In contrast, somatostatin receptor imaging with [⁶⁸Ga]DOTA-TOC PET/CT identifies expressed somatostatin receptor subtype 2 (SSTR2) on pro-inflammatory macrophages, offering a highly specific signal for plaque inflammation. Our previous research showed that patients with coronary artery calcifications and concomitant [⁶⁸Ga]DOTA-TOC uptake had significantly higher rates of stroke and all-cause mortality compared to those without.⁹ Of note, our previous study was limited by a relatively small patient sample (n = 108) and a relatively short follow-up. It now remains unclear whether the presence of calcified and/or inflamed carotid plaques identified by [⁶⁸Ga]DOTA-TOC PET/CT may yield similar prognostic value in the prediction of stroke and MACE. Therefore, the present study aimed to evaluate this in a real-world oncologic population with extended follow-up.

Materials and methods

Patient population

We retrospectively analysed the complete record of patients with known or suspected neuroendocrine tumours (NETs) who underwent [⁶⁸Ga]DOTA-TOC PET/CT imaging at the University Hospital of Bern, Switzerland, between June 2014 and July 2023. From these, we excluded patients with prior history of stroke, carotid artery surgery or stent implantation and neck radiation therapy. Patients undergoing brain-only [⁶⁸Ga]DOTA-TOC PET/CT imaging, with incomplete clinical data records, poor image quality (significant metal artefacts or enlarged cervical lymph nodes) precluding quantitative analysis of carotid plaque or calcification were also excluded. If a patient had multiple [⁶⁸Ga]DOTA-TOC PET/CT imaging, only the one performed first was considered. Baseline characteristics including cardio/cerebrovascular risk factors as well as oncologic data were recorded. The study was conducted according to the Declaration of Helsinki and was approved by the local Ethics Committee (KEK-Nr. 2022–00486). The outline of the present study is shown in Figure 1.

Figure 1.

Flowchart of the present study.

Imaging protocol

Patients undergoing [⁶⁸Ga]DOTA-TOC PET/CT did not require specific preparation. PET imaging was acquired 60 min after intravenous injection of 151.6 ± 14.0 MBq [⁶⁸Ga]DOTA-TOC. The imaging equipment included standard axial field-of-view (SAFOV, Biograph mCT and Biograph Vision 600, Siemens Healthineers, Knoxville, TN, USA) or long-axial field-of-view (LAFOV, Biograph Vision Quadra Siemens Healthineers, Knoxville, TN, USA) PET/CT scanners. As previously published,¹⁰ images were acquired on the SAFOV scanner in continuous bed motion with a table velocity of 1.1 mm/s (skull-base to mid femur), equivalent to 2 min/bed position, while on LAFOV PET were acquired in list-mode for 10 min in a single bed position (skull-vertex to mid femur). The scanning and reconstruction parameters of CT images were performed as previously published.¹⁰ Non-contrast-enhanced, low-dose CT images were used for attenuation correction, identification, and to assess carotid artery calcification.

Image evaluation

Anatomical coregistration of CT and PET was carefully verified prior to assessing [⁶⁸Ga]DOTA-TOC uptake in the carotid arteries. Images were visually analysed slice-by-slice by two experienced nuclear medicine physicians who were blinded to clinical data. Lesions showing focal [⁶⁸Ga]DOTA-TOC uptake visually above the surrounding background activity were additionally assessed semiquantitatively and considered PET-positive when the target-to-background ratio (TBR) was >1. A region of interest (ROI) was manually drawn on active foci in the carotid arteries to calculate maximum standardized uptake values (SUVmax). To normalize for blood pool activity, three standard ROIs (1 cm diameter) were placed in three consecutive slices in the superior vena cava, and SUVs were calculated as the mean value of the three measurements. TBR was defined as the ratio between SUVmax in the carotid artery and SUVmean of the blood pool.^3,11 Visual analysis of carotid artery CT images was performed to assess calcification, defined as structures within the artery wall demonstrating a density greater than 130 HU.¹² The calcification volume in bilateral carotid arteries was quantified by slice-by-slice delineation of ROI. To assess intra- and inter-reader reproducibility, the same reader and a second reader re-analysed the [⁶⁸Ga]DOTA-TOC PET/CT images in a randomly selected sample of 100 cases >4 weeks after the initial assessment, and intraclass correlation coefficients (ICCs) with their 95% confidence intervals (CIs) were calculated.

Follow up

Clinical records of stroke and MACE (myocardial infarction, hospitalization for myocardial ischaemia, coronary artery revascularization, all-cause death), were collected during the follow-up period. The time to the occurrence of MACE was recorded and defined as the time interval between [⁶⁸Ga]DOTA-TOC and the first event. The median follow-up duration was 55 (interquartile range, 42–72) months.

Statistical analysis

Statistical analysis was performed using SPSS software (version 30.0; IBM, USA). Normally distributed continuous variables are expressed as mean ± standard deviation (SD), and non-normally distributed data are reported as median with interquartile range (IQR; P25, P75). Categorical variables are presented as numbers and percentages (n, %). For group comparisons, the independent samples t-test was applied for normally distributed continuous variables, and the Mann–Whitney U test was used for non-normally distributed variables. Comparisons of categorical data were conducted using the chi-square test or Fisher’s exact test, as appropriate. Correlations between two continuous variables were examined using Pearson’s R. The rate of death and MACE was compared using Cox proportional hazard models to estimate hazard ratios (HRs) with 95% CIs and by Kaplan–Meier curves with Log-Rank test, dividing patients into four groups according to the presence of calcified and/or inflamed carotid plaques. Because of the limited number of events, parsimonious multivariable Cox models were used, and the covariates retained in the final models differed for stroke and MACE, as detailed in Supplementary data online, Tables S1 and S2. Considering that tumour-related deaths may represent a competing risk for stroke and MACE, we further performed competing risk analyses using cumulative incidence functions and Fine-Gray sub-distribution hazard models. P values < 0.05 was considered statistically significant.

Results

Clinical characteristics

Of the 403 patients initially screened, 353 met the inclusion criteria. Of these, 94.3% (333/353) had a confirmed NETs. Baseline clinical characteristics are shown in Table 1.

Table 1.

Baseline characteristics of study population (n = 353)

Characteristics	Value
Age (years)	60.3 ± 14.5
Male, n (%)	190 (53.8)
Body mass index (kg/m2)	25.8 ± 5.3
Risk factors, n (%)
Overweight	113 (19.3)
Obesity	68 (32.0)
Current smokers	58 (16.4)
Family history of CVD	15 (4.2)
Hypercholesterolaemia	72 (20.4)
Hypertension	136 (38.5)
Diabetes mellitus	74 (21.0)
Peripheral artery disease	3 (0.8)
Prior cardiovascular events	16 (4.5)
Tumour characteristics, n (%)
Known NETs	333 (94.3)
Small intestine	130 (36.8)
Pancreas	69 (19.5)
Colorectal	38 (10.8)
Lung	11 (3.1)
Others	85 (24.1)
Suspected NETs	20 (5.7)
PRRT therapy	70 (19.8)

PRRT, peptide receptor radionuclide therapy. Qualitative data: numbers (percentages); continuous data: mean ± SD.

Relationship between [⁶⁸Ga]DOTA-TOC uptake in the carotid arteries and calcification

One hundred twenty-three patients (34.8%) showed carotid artery calcification of any degree in at least one side, with a median calcification volume of 77.7 (27.9, 169.8) cm³. In 46 patients (13.0%), a total of 67 foci of increased [⁶⁸Ga]DOTA-TOC uptake were detected in the carotid arteries (37 left, 30 right). The median SUVmax was 1.68 (1.55, 1.93), while the median TBR was 2.14 (1.65, 2.79). The intra-reader ICC for calcification volume and TBR measurements was 0.97 (95% CI: 0.95, 0.98) and 0.96 (95% CI: 0.93, 0.97), and the inter-reader ICC was 0.95 (95% CI: 0.94, 0.96) and 0.91 (95% CI: 0.86, 0.94).

Considering calcification and/or SSTR2 expression, 38.8% (26/67) of focal [⁶⁸Ga]DOTA-TOC uptakes correlated with a detectable calcification on low-dose CT imaging. SUVmax [1.76 (1.47, 2.10) vs. 1.66 (1.57, 1.85), P = 0.837] and TBR [2.20 (1.53, 2.81) vs. 2.13 (1.71, 2.82), P = 0.842] were not different between calcified and non-calcified lesions.

As illustrated in Figure 2, patients with focal [⁶⁸Ga]DOTA-TOC uptake demonstrated significantly greater carotid calcification volumes than those without uptake. And those with carotid calcification showed significantly higher [⁶⁸Ga]DOTA-TOC uptake values (both SUVmax and TBR) compared to patients without calcification. Supporting this bidirectional relationship, calcification volume correlated moderately with both SUVmax (r = 0.322) and TBR (r = 0.310), with P < 0.001 for both.

Figure 2.

Correlation between carotid artery calcification and focal [⁶⁸Ga]DOTA-TOC uptake. (A) Carotid artery calcification volume was significantly larger in patients with focal [⁶⁸Ga]DOTA-TOC uptake than in those without, 47.2 (0, 149.1) cm³ vs. 0 (0, 13.4) cm³, P < 0.001; (B,C) [⁶⁸Ga]DOTA-TOC uptake (SUVmax and TBR) was significantly higher in patients with carotid artery calcification than in those without, (B) SUVmax: 0 (0, 1.44) vs. 0 (0, 0), P < 0.001, (C) TBR: 0 (0, 1.38) vs. 0 (0, 0), P < 0.001.

Relationship among carotid artery calcification, focal [⁶⁸Ga]DOTA-TOC uptake, and cardiovascular risk factors

As shown in Figure 3, smoking, hypercholesterolaemia, hypertension, diabetes, and prior cardiovascular events were correlated with higher rate of carotid artery calcification [odds ratio (OR): 1.95–8.94, P < 0.05, Figure 3A]. Further, a focal [⁶⁸Ga]DOTA-TOC uptake in the carotid artery was more often seen in patients with smoking habits, family history of cardiovascular disease (CVD), hypercholesterolaemia, and hypertension (OR: 2.39–4.97, P < 0.05, Figure 3B).

Figure 3.

Relationship between cardiovascular risk factors and carotid artery calcification (A), focal [⁶⁸Ga]DOTA-TOC uptake (B). * indicates a statistically significant.

Based on the patient-level presence of carotid artery calcification and focal [⁶⁸Ga]DOTA-TOC uptake, patients were divided into four groups (Group 1: Calcification(-) and [⁶⁸Ga]DOTA-TOC uptake(-); Group 2: Calcification(-) and [⁶⁸Ga]DOTA-TOC uptake(+); Group 3: Calcification(+) and [⁶⁸Ga]DOTA-TOC uptake(-); Group 4: Calcification(+) and [⁶⁸Ga]DOTA-TOC uptake(+)). This classification reflects a patient-level imaging phenotype and does not imply that calcification and tracer uptake were colocalized within the same plaque. As shown in Table 2, patients with smoking, family history of CVD, hypercholesterolaemia, and hypertension showed a higher rate of both calcifications and focal [⁶⁸Ga]DOTA-TOC uptake in the carotid arteries, consistent with inflamed and calcified plaques. If more risk factors are present (prevalence of ≥ 4 cardiovascular risk factors), this trend becomes even clearer (Group 1–4: 5.5%, 15.4%, 17.8%, and 24.2%, P = 0.002).

Table 2.

Comparison of the proportion of patients with cardiovascular risk factors across groups (n = 353)

Risk factors	Group 1 (n = 217)	Group 2 (n = 13)	Group 3 (n = 90)	Group 4 (n = 33)	P
Obesity % (n)	17.1 (37)	46.2 (6)	22.2 (20)	15.2 (5)	0.053
Current smokers % (n)	12.4 (27)	23.1 (3)	18.9 (17)	33.3 (11)	0.018a
Family history of CVD % (n)	3.7 (8)	7.7 (1)	1.1 (1)	15.2 (5)	0.010a
Hypercholesterolaemia % (n)	14.7 (32)	23.1 (3)	26.7 (24)	39.4 (13)	0.004a
Hypertension % (n)	25.8 (56)	46.2 (6)	56.7 (51)	69.7 (23)	<0.001a
Diabetes mellitus % (n)	16.6 (36)	23.1 (3)	27.8 (25)	30.3 (10)	0.073
Peripheral artery disease % (n)	0 (0)	0 (0)	2.2 (2)	3.0 (1)	0.087
Prior cardiovascular events % (n)	1.4 (3)	0 (0)	11.1 (10)	9.1 (3)	<0.001a
High risk (≥ 4 risk factors) % (n)	5.5 (12)	15.4 (2)	17.8 (16)	24.2 (8)	<0.001a

Proportions among the four groups were compared using the chi-square or Fisher's test, with the P-value indicating the overall difference across the four groups.

^aindicates a statistically significant P-value (P < 0.05).

Risk of cardio/cerebrovascular events

During follow-up, non-fatal cardio/cerebrovascular events occurred in 23 patients (stroke: 12 patients, 3.4% and MACE: 11 patients, 3.1%), while 83 all-cause deaths (23.5%) were documented. The incidence of stroke and MACE in each group was as follows: Group 1: 0.5%, 1.4%; Group 2: 0%, 0%; Group 3: 6.7%, 4.4%; Group 4: 15.2%, 12.1% (χ² = 19.090, 9.241; P < 0.001, 0.018, Figure 4). Figure 5 shows KM-survival analysis demonstrating higher rate of MACE and stroke for patients in Group 4.

Figure 4.

The incidence of stroke and MACE in each group. The incidence of stroke in Groups 1–4 was 0.5%, 0%, 6.7%, and 15.2%, respectively; χ² = 19.090, P < 0.001. The incidence of MACE in Groups 1–4 was 1.4%, 0%, 4.4%, and 12.1%; χ² = 9.241, P = 0.018.

Figure 5.

Kaplan–Meier estimates of event probability for stroke (A) and MACE (B) across the four groups. The presence of both carotid artery calcification and [⁶⁸Ga]DOTA-TOC uptake was associated with a significantly higher risk of stroke and MACE compared to other presentations. Group 1: carotid artery calcification (−) [⁶⁸Ga]DOTA-TOC uptake (−), Group 2: calcification (−) uptake (+), Group 3: calcification (+) uptake (−), Group 4: calcification (+) uptake (+).

Univariate Cox regression analysis identified age, family history of CVD, hypertension, prior cardiovascular events, peripheral artery disease, the coexistence of carotid artery calcification, and [⁶⁸Ga]DOTA-TOC uptake as risk factors for stroke (HR = 1.08–58.31, P all <0.05), smoking and the coexistence of carotid artery calcification and [⁶⁸Ga]DOTA-TOC uptake were risk factors for MACE (P = 0.009 and <0.001). Multivariable COX analysis confirmed that the coexistence of carotid artery calcification and [⁶⁸Ga]DOTA-TOC uptake was the independent risk factor for stroke (HR: 29.98, 95% CI: 3.22, 279.18, P = 0.017) and MACE (HR: 10.47, 95% CI: 2.24, 48.90, P = 0.003) in NETs patients (see Supplementary data online, Tables S1 and S2). Through the competing risk analyses, the coexistence of carotid artery calcification and [⁶⁸Ga]DOTA-TOC uptake remained significantly associated with both stroke (sub-distribution HR: 93.03, 95% CI: 9.47, 913.61, P < 0.01) and MACE (sub-distribution HR: 11.91, 95% CI: 2.42, 58.71, P < 0.01). In subgroup analysis, Group 4 had a significantly higher rate of both stroke and MACE compared to Group 3 (χ² = 4.058, 4.397; P = 0.044, 0.036).

Discussion

In our study, we demonstrated that the calcification burden within carotid plaques correlates with the presence of inflammation. Most importantly, the combined patient-level phenotype characterized by both carotid calcification and focal [⁶⁸Ga]DOTA-TOC uptake was independently associated with stroke and MACE after adjustment for clinical risk factors, identifying a distinct subgroup at substantially elevated risk.

Atherosclerotic plaque formation and progression are driven by a multifactorial pathological process that evolves intermittently over time.¹³ Among the key cellular mediators of this process are pro-inflammatory macrophages, which play a central role in plaque progression and destabilization.¹⁴ When activated macrophages accumulate within the plaque, they may weaken the structural integrity of the fibrous cap, thereby increasing the risk of plaque rupture.¹⁵ In view of its high sensitivity in detecting inflammation, [¹⁸F]FDG PET/CT was firstly used to assess the prognostic impact of patients with inflamed carotid plaques. In a study by Moon et al.¹⁶ 1089 asymptomatic adults were followed up for 4.2 years, and 19 individuals (1.74%) developed cardiocerebrovascular events. On multivariable Cox regression, both carotid [¹⁸F]FDG uptake and intima-media thickness emerged as independent predictors of events (HR 2.98 and 2.82, respectively). These findings are directionally consistent with our results, although our study evaluated SSTR2-targeted inflammation imaging and additionally integrated plaque calcification assessment.

Additional carotid [¹⁸F]FDG studies provide important biological and clinical context for the interpretation of our findings. In a prospective cohort of 49 patients undergoing carotid endarterectomy, Cocker et al.¹⁷ showed that carotid [¹⁸F]FDG uptake correlated with plaque CD68 and CD45 expression and was higher in symptomatic than asymptomatic plaques, particularly in patients imaged within 90 days of a cerebrovascular event. Likewise, Giannotti et al.¹⁸ reported in 25 patients with recently symptomatic carotid stenosis that [¹⁸F]FDG uptake was positively associated with MRI markers of plaque vulnerability, including lipid-rich necrotic core, and inversely associated with fibrous cap thickness and calcium volume. Although these studies did not assess future stroke or MACE directly, they strengthen the biological plausibility of our observations by showing that increased carotid PET signal identifies plaques with inflammatory and morphological features of instability. Compared with these FDG-based studies, our data extend the literature by suggesting that SSTR2-targeted PET/CT, especially when combined with CT-based calcification assessment, may help identify a subgroup at particularly high risk of stroke and MACE. Prior studies also investigated the usefulness of other imaging modalities such as CT angiography¹⁹ or dual-tracer imaging using [¹⁸F]FDG and [¹⁸F]NaF²⁰ in the assessment of inflammation and calcification in atherosclerotic plaques. However, it should be noted that perivascular fat attenuation indices derived from CT angiography offer only an indirect measure of inflammatory activity, and dual-tracer PET imaging typically requires a 2-day protocol and may cause higher radiation exposure.

Conversely, SSTR2 PET/CT imaging represents a more effective methodological approach for elucidating the interplay between inflammation and calcification, as already demonstrated by Rominger et al.,¹¹ Malmberg et al.,²¹ and Mingels et al.⁹ In fact, SSTR2 PET/CT imaging can assess the presence of inflammation in target lesions, capitalizing on the expression of SSTR2 in activated macrophages. Tarkin et al.²² demonstrated that SSTR2 gene expression is closely associated with M1-like macrophages. In their work, strong correlations were seen between SSTR2 mRNA/CD68 mRNA levels and [⁶⁸Ga]DOTA-TATE PET TBRmax in carotid plaques (r = 0.89 and 0.84, respectively), hence indicating that focal carotid uptake on SSTR2-targeted PET imaging directly reflects inflammatory activity and may represent a risk factor for cardio-cerebrovascular events.

In this regard, prior studies have explored the role of SSTR2 PET imaging in atherosclerosis, linking increased carotid or coronary plaque tracer uptake to conventional cardiovascular risk factors,^9,11,21 or in the identification of high-risk vulnerable plaques,^22,23 and in the assessment of the severity of cardiovascular disease.²⁴ However, evidence on a direct association between SSTR2 PET findings and hard clinical endpoints such as stroke and MACE is still limited.²⁵

Building upon our group's previous finding of increased stroke incidence in subjects with [⁶⁸Ga]DOTA-TOC uptake in calcified coronary artery plaques,⁹ we now expanded on carotid plaques, also increasing sample size and follow-up duration.

While confirming that focal carotid [⁶⁸Ga]DOTA-TOC uptake is associated with multiple cardiovascular risk factors (smoking, hypertension, hyperlipidaemia, and family history, etc.), our results suggest that the added prognostic value of PET lies in identifying, among patients with calcified carotid plaques, a subgroup at particularly high risk of stroke and MACE.

Calcification is a hallmark of advanced atherosclerotic disease,²⁶ and the relationship between carotid calcification, plaque vulnerability, and clinical events remains debated. Carotid calcium scoring has been identified as an independent marker of carotid stenosis and ischaemic symptoms¹² and predicts MACE and cardiovascular mortality following carotid endarterectomy.^26,27 Yang et al.²⁸ reported strong associations between carotid calcification (presence, multiplicity, superficial type) and intraplaque haemorrhage (OR = 3.0–3.9), a recognized feature of high-risk plaques and a biomarker for ischaemic cerebrovascular events.²⁹ In a meta-analysis encompassing 89 studies and 22 683 patients, Bytyçi et al.³⁰ demonstrated that carotid calcification and a higher calcified plaque burden are more prevalent among patients with significant coronary artery disease. However, the mere presence of calcification does not give precise information on plaque stability; rather, specific calcification characteristics, such as size, number, location, morphology, and composition, are more likely to provide useful markers to predict a worse outcome.^4,31,32 A meta-analysis by Homssi³³ found no significant link between carotid calcium burden/morphology and ischaemic stroke, while the ANTIQUE study³⁴ reported a strong association of large carotid calcifications with coronary artery disease and atrial fibrillation. These inconsistent findings underscore the substantial heterogeneity of arterial calcified plaques, and a clear relationship between carotid calcification and stroke or MACE remains uncertain. In our study, incorporating the assessment of inflammation yielded to identify a subgroup of patients with even higher risk for adverse outcomes compared to patients with calcified, non-active plaques (15.2% vs. 6.7%). These findings seem to be in contrast with what has been reported by studies based on the analysis of coronary plaques, wherein microcalcifications and spotty calcifications (but not macrocalcifications) are more closely tied to vulnerable plaques and cardiovascular events.^19,35,36 Differences in calcification patterns may arise from vascular-bed-specific variations in structure, cellular composition, molecular milieu, and haemodynamic forces.³⁷ The fact that a strong association between macrocalcified carotid plaques and hard events could be found suggest that combined low-dose CT and [⁶⁸Ga]DOTA-TOC PET imaging be used for assessment of cardio-cerebrovascular risk.

Intriguingly, our data suggest that isolated focal [⁶⁸Ga]DOTA-TOC uptake without concomitant calcification does not predict a higher risk of stroke or MACE. From a pathophysiological perspective, inflammation is seen earlier in atherogenesis¹³ and inconsistently persists throughout its progression, whereas calcification may represent an endpoint resulting from sustained and excessive inflammation.^31,38 The phenotypes of carotid plaque inflammation, calcification, and combined inflammation-calcification likely reflect distinct stages within the atherosclerotic continuum, being the ‘focal uptake only’ pattern suggestive of earlier disease. This is consistent with a study featuring CT coronary angiography and optical coherence tomography in 578 CAD patients, wherein Fujimoto et al.¹⁹ observed that vascular inflammation in early-to-intermediate atherosclerosis often occurs without concomitant calcification. During calcification accrual, lipid-rich plaques, macrophages, and microvessels initially increase before subsequently decreasing, indicating that the inflammation-calcification relationship is not merely linear. These insights may help explain the differential cardiovascular outcomes among phenotypic subgroups in our study and reinforce the concept of atherosclerosis as an intermittently progressing process characterized by cycles of ‘inflammation—healing—calcification—recurrent inflammation’.¹³ However, given the very small size of this subgroup, the lack of a statistically significant association may also reflect insufficient statistical power rather than a true negative finding. Accordingly, the prognostic significance of the PET(+)/CT(-) phenotype should be regarded as inconclusive rather than negative. Therefore, larger prospective studies with adequate sample sizes are needed to clarify the biological significance of isolated focal [⁶⁸Ga]DOTA-TOC uptake without calcification, as well as its relationship with stroke and MACE.

Limitations

Some limitations in our study should be acknowledged. First, this study has a retrospective design, and the studied population consisted of oncological patients with confirmed or suspected NETs, the mortality rate related to NETs is relatively high. Although we have conducted the competing risk analyses, the same results may not be completely transferred to a general population of disease-free patients.

Second, invasive or CT-based carotid angiography was not routinely performed to assess the degree of carotid stenosis. This represents a potential confounding factor that may be associated with a different rate of stroke and MACE. Moreover, as non-calcified, soft plaques were not systematically characterized with contrast enhanced CT, some focal [⁶⁸Ga]DOTA-TOC uptakes without a corresponding CT lesion may be non-specific. Prior studies have reported an association between SSTR2-targeted tracer uptake and non-calcified coronary plaque features,^11,19 thus we cannot exclude that part of the ‘uptake-only’ pattern in our cohort reflects non-calcified plaque substrate that was not detectable on low-dose CT, or alternatively non-specific uptake. This may have contributed to the lack of prognostic significance observed in patients with isolated [⁶⁸Ga]DOTA-TOC uptake.

Third, the total number of stroke and MACE recorded was relatively low, reflecting the incidence rate of a population without known cerebrovascular disease. This limited the assessment of reliable thresholds for [⁶⁸Ga]DOTA-TOC uptake parameters (such as SUVmax or TBR) or calcification volume to predict these events. Consequently, it was not feasible to construct a robust prognostic prediction model based on these quantitative plaque characteristics. Given the limited number of stroke and MACE events, the multivariable Cox models are at risk of overfitting and should be interpreted as exploratory. Furthermore, only 26 focal [⁶⁸Ga]DOTA-TOC uptake sites in the carotid artery colocalized with carotid artery calcification, and due to the relatively small number of events, we were unable to explore the relationship between inflammatory calcified plaques and cardiovascular events at the lesion level.

Finally, the imaging examinations for the enrolled subjects were conducted over a considerable time span. During this period, the imaging instrumentation was upgraded. Differences in detection sensitivity and spatial resolution between the different scanner generations may introduce variability in the PET quantitative parameters.^9,10

Conclusions

Calcified carotid plaques represent a risk factor for the onset of cardio-cerebrovascular hard events. The combined patient-level presence of carotid calcification and M1-like macrophage-related inflammation as detected by [⁶⁸Ga]DOTA-TOC PET/CT identifies a subgroup at particularly high risk of cardio-cerebrovascular hard events. Hence, [⁶⁸Ga]DOTA-TOC PET/CT may be effectively used in order to plan an early start of an effective preventive therapy.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Author contributions

Xiaoliang Shao (Conceptualization, Writing—original draft [equal], Formal analysis [supporting]), Carola Maria Bregenzer (Data curation [equal], Writing—original draft [supporting]), Jiahui Wang [Formal analysis, Methodology (equal)], Yizhou Chen [Formal analysis, Methodology (equal)], Yuetao Wang (Data curation, Methodology [equal], Writing—original draft [supporting]), Luisa Maria Knappe (Data curation [equal], Writing—original draft [supporting]), Axel Rominger (Conceptualization [supporting], Writing—original draft [equal]), Kuangyu Shi (Resources [lead], Writing—original draft [equal]), and Federico Caobelli (Conceptualization, Methodology, Writing—original draft [equal], Supervision [lead])

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.