Population

The study protocol was approved by the Fujita Health University Human Research Committee and conformed to the ethical guidelines of the 1975 Declaration of Helsinki. The requirement for informed consent was waived because of the retrospective study design, and an opt-out approach was used in accordance with institutional guidelines. All patients with cardiac sarcoidosis were diagnosed according to the JCS 2016 Guideline on Diagnosis and Treatment of Cardiac Sarcoidosis [9]. We included consecutive cases of FDG-PET/CT involving patients with cardiac sarcoidosis who had started immunosuppressive therapy, specifically those who underwent PET/CT at least 12 months after initiating immunosuppressive therapy between June 2013 and October 2023. Immunosuppressive therapy eligibility was determined based on the guidelines of the Japanese Circulation Society and treatment was administered following the established protocol [9]. Patients were excluded if they were considered to be in an unstable clinical phase, defined as those who, within 6 months prior to FDG-PET/CT, had (1) been hospitalized for heart failure or experienced sustained ventricular tachycardia or ventricular fibrillation, (2) received new or upgraded pacemaker implantation, or (3) had any changes in prednisolone dosage or started new immunosuppressive agents. These criteria were applied not to exclude high-risk patients, but to align the clinical phase at the time of imaging, as FDG-PET findings may be transiently influenced by recent acute events or treatment modifications.

The primary endpoint was a composite cardiovascular event, including cardiac death, sustained ventricular tachycardia/fibrillation, appropriate activation of an implantable defibrillator, and hospitalization for heart failure. The clinical data were accessed for research purposes between 01/09/2024 and 31/12/2024. The authors had access to information that could identify individual participants during data collection; however, all data were anonymized prior to analysis.

FDG-PET/CT Protocol

FDG-PET/CT was performed using a dedicated PET-CT scanner (Biograph mCT X-4R; Siemens K.K., Tokyo, Japan). Patients were instructed to follow a low-carbohydrate, fat-rich, and protein-permitted diet the day before the examination, followed by fasting for at least 18 h before the procedure. CT images were acquired in 2-mm slices with a 0.35 helical pitch at 120 kV and an average of 82 mA, with auto exposure control with dose modulation (CARE Dose 4D) and a matrix of 512 × 512 pixels. Whole-body scans were performed 60 min after an intravenous injection of 185 MBq of 18F-FDG, followed by cardiac imaging with electrocardiogram gating to assess left ventricular function. FDG-PET data were collected in a three-dimensional mode for 2 min in each bed position. The FDG-PET data had a matrix of 200 × 200 pixels. CT and FDG-PET images were fused to match a matrix of 512 × 512 pixels [12,13].

FDG-PET/CT analysis

Quantitative assessment

Independent of the visual classification, myocardial 18F-FDG accumulation was quantitatively assessed using RAVAT software (Nihon Medi-Physics Co., Ltd., Tokyo, Japan). Volumes of interest (VOIs) were defined for both the heart and aorta in the PET images at the locations previously described [13–16]. An elliptical VOI was positioned extrinsically around the heart, and columnar VOIs were placed at three points along the inner wall of the descending aorta. These locations included the proximal descending thoracic aorta, upper edge of the liver, and upper portion of the common iliac artery bifurcation. The height of each aortic VOI was fixed at 10.0 mm (equivalent to 5 pixels), while the diameter was adjusted according to the aortic width. The mean standardized uptake values (SUVs) of the three aortic VOIs were measured and used as reference values. In accordance with our previous studies, a threshold of 1.4 times the mean SUV of the descending aorta was used to identify high myocardial 18F-FDG accumulation [13–16]. Physiological myocardial uptake was carefully excluded; diffuse or homogeneous FDG uptake patterns without focal accumulation were interpreted as physiological and were not classified as pathological uptake. Consequently, only focal or focal-on-diffuse uptake exceeding the 1.4-fold aortic SUVmean threshold was included in the calculation of the cardiac metabolic volume (CMV). Cardiac metabolic activity (CMA) was then calculated by multiplying the CMV by the corresponding average SUV [13,17]. In this study, CMA refers to left ventricular values unless otherwise specified.

Composite cardiac events

A major cardiac event was defined as a composite of cardiac death, sustained ventricular tachycardia, ventricular fibrillation, appropriate activation of an implantable cardioverter defibrillator, and hospitalization for heart failure. LVEF was assessed using transthoracic echocardiography performed as part of standard clinical evaluation. Follow-up data were collected by observers, who were blinded to the FDG-PET results, during clinical visits or standardized telephone interviews.

Statistical analysis

The Shapiro–Wilk test was applied to evaluate the normality of continuous variables. Continuous variables are presented as mean ± standard deviation if normally distributed or as median (interquartile range) if not normally distributed. To compare intergroup differences, Student’s t-test was used for parametric variables and the Mann–Whitney U test for nonparametric variables. For categorical variables, the chi-squared test or Fisher’s exact test was employed. The Wilcoxon signed-rank test was used to assess within-subject changes. Cumulative survival curves were generated using Kaplan–Meier analysis, with the time from the initial FDG-PET/CT to the first event. Patients were categorized into the cardiac accumulation (+) and (-). Differences between time-to-event curves were analyzed using log-rank tests. The event rates for the cardiac accumulation (+) and (-) were compared using the Kaplan–Meier estimator. Univariate Cox proportional hazards models were used to assess predictors of cardiac events. Variables with a significance level of P < 0.05 in the univariate analyses, along with age and sex, were included in multivariable analysis. In the short-term analysis, troponin I was included in the multivariable model because it was statistically significant in the univariate analysis. The proportional hazards assumption was assessed using Schoenfeld residuals and log-minus-log survival plots, and no significant violations were observed. Wilcoxon signed-rank tests were used to analyze changes in quantitative cardiac uptake values in patients who underwent multiple FDG-PET scans. The results are presented as hazard ratios (HRs) with 95% confidence intervals (CIs). All statistical analyses were conducted using JMP Pro software (version 17.0, SAS Institute), with statistical significance set at P < 0.05.

Patient Characteristics

The median age of the 79 patients was 67 (range 57–72) years, with 53 (67%) being women. Patients were categorized into the cardiac accumulation (+) and (-). Significant differences were observed in CMA (P < 0.0001), FDG accumulation in the right ventricle (P = 0.0001) and extramyocardial areas (P = 0.017) between the groups. The median prednisolone dose was 5 mg in both groups, with methotrexate used in one patient from each group (Table 1).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Patient characteristics.

https://doi.org/10.1371/journal.pone.0347595.t001

Clinical outcomes

Cardiac events by group are summarized in Table 2. Over a median follow-up of 4 years, 9 patients in the cardiac accumulation (+) experienced cardiac events, with seven of these events occurring within 2 years post FDG-PET. In the cardiac accumulation (-), 4 patients had cardiac events, with only one event occurring within the first 2 years.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Major cardiovascular events after FDG-PET/CT during the phase of stable medical and device therapy.

https://doi.org/10.1371/journal.pone.0347595.t002

The Kaplan–Meier analysis showed no significant difference in composite cardiac events between the cardiac accumulation (+) and (-) over the entire follow-up period (log-rank P = 0.078). However, within the first 2 years, the outcomes were significantly worse in the cardiac accumulation (+) than in the cardiac accumulation (-) (log-rank P = 0.030) (Fig 1). Supplemental analyses using SUVmax, SUVmean, and CMV showed trends consistent with the CMA-based results, and none of these parameters predicted short-term or long-term outcomes in Cox regression models. These findings indicate that the choice of CMA as the primary quantitative index did not alter the overall conclusions of the study. When the cardiac accumulation (+) was further stratified into three subgroups based on CMA tertiles, the incidence of long-term events did not show a stepwise increase with higher CMA values. Specifically, the cumulative cardiac event rates were similar between the middle (CMA 30–199) and highest (CMA ≥ 200) CMA tertiles, both of which were higher than the rate observed in the lowest CMA tertile (CMA < 30). (Supplementary S1 Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Composite of cardiac events after FDG-PET/CT.

(a) Kaplan–Meier curves for the entire follow-up period. (b) Comparisons of events within the first 2 years and after 2 years post FDG-PET/CT. Vertical dashed line indicates the 2-year follow-up point. P values calculated using log-rank test. FDG-PET/CT, ¹⁸F-fluorodeoxyglucose-positron emission tomography/computed tomography. Supplementary S1 Fig Association between composite cardiac events and cardiac metabolic activity (CMA).

https://doi.org/10.1371/journal.pone.0347595.g001

Table 3 presents the Cox proportional hazards analyses for major cardiac events. In the univariate analyses, long-term composite cardiac events were associated with cardiac troponin I levels, sustained VT/VF, and LV dysfunction, but not with the presence of LV accumulation (P = 0.077), LV-CMA as a continuous variable (P = 0.78), or log-transformed LV-CMA on PET (P = 0.46). In the multivariate analysis adjusted for age, sex, and sustained VT/VF, LV dysfunction (EF < 50%) was an independent predictor (HR 5.22, 95% CI 1.08–25.43, P = 0.040) as well as cardiac troponin I level (HR 1.03, 95% CI 1.00–1.08, P = 0.034) (Table 3A). In the short-term analysis, only cardiac troponin I level was predictive in univariate analysis (HR 1.04, 95% CI 1.00–1.07, P = 0.045), whereas the presence of LV accumulation (P = 0.064), LV-CMA as a continuous variable (P = 0.81), and log-transformed LV-CMA on PET (P = 0.16) were not. In the multivariate analysis adjusted for age and sex, cardiac troponin I remained an independent predictor (HR 1.05, 95% CI 1.00–1.08, P = 0.029) (Table 3B).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Cox analysis for major cardiac events in the (A) long-term and at (B) 2-year follow-ups.

https://doi.org/10.1371/journal.pone.0347595.t003

In patients with LVEF < 50% (n = 37), the incidence of cardiac events (n = 11) tended to be higher in the cardiac accumulation (+) group than in the (-) group, although this difference did not reach statistical significance (log-rank P = 0.097). In patients with LVEF≥50% (n = 42), cardiac events occurred in just 2 patients. The rates of composite cardiac events were also similar between the cardiac accumulation (+) and (-) groups (annual cardiac event rates 1.3% vs. 0.8%; log-rank P = 0.91) (Fig 2).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Composite of cardiac events after FDG-PET/CT stratified by baseline left ventricular ejection fraction (LVEF).

(a) LVEF<50% and (b) LVEF≥50%. Vertical dashed line indicates the 2-year follow-up point. P values calculated using log-rank test. FDG-PET/CT, ¹⁸F-fluorodeoxyglucose-positron emission tomography/computed tomography; LVEF, left ventricular ejection fraction.

https://doi.org/10.1371/journal.pone.0347595.g002

Observational periods are presented as median (interquartile range). Values are number of patients; numbers in parentheses indicate events within 2 years after FDG-PET/CT. FDG-PET/CT: 18F-fluorodeoxyglucose-positron emission tomography/computed tomography, ICD: implantable cardioverter defibrillator, LVDd: left ventricular end-diastolic dimension, LVEF: left ventricular ejection fraction, VF: ventricular fibrillation, VT: ventricular tachycardia

Post-FDG-PET Strategy and Cardiac Uptake Changes

Among the 38 patients in the cardiac accumulation (+), the immunosuppressive regimen was changed in two patients, cardiac events occurred before FDG-PET re-examination in four patients, and re-examination was not performed in 9 patients. Of the remaining 23 patients, 15 showed the cardiac accumulation on follow-up, whereas 8 did not. In the cardiac accumulation (-) (41 patients), immunosuppressive regimens were changed in 1 patient, cardiac events occurred before re-examination in 2 patients, and re-examination was not performed in 20 patients. Of the other 18 patients, 2 developed cardiac uptake, while 16 remained without cardiac uptake on FDG-PET re-examination (Fig 3). The follow-up CMA decreased from 4.83 to 0.82 (P = 0.038), and in patients with initial follow-up cardiac uptake, CMA also decreased from 66.2 to 15.7 (P = 0.048) (Fig 4).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Post-FDG-PET/CT clinical management strategy and changes in cardiac uptake between baseline and additional scans.

CMA, cardiac metabolic activity; FDG-PET/CT, ¹⁸F-fluorodeoxyglucose-positron emission tomography/computed tomography.

https://doi.org/10.1371/journal.pone.0347595.g003

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Quantitative cardiac uptake values in consecutive FDG-PET/CT scans.

CMA values are expressed in g·SUV (CMA = cardiac metabolic volume × mean standardized uptake value). FDG-PET/CT, ¹⁸F-fluorodeoxyglucose-positron emission tomography/computed tomography.

https://doi.org/10.1371/journal.pone.0347595.g004

Representative Cases

Case 1 involved a 76-year-old woman who presented with a CMV of 101.5, SUV_max of 6.0, and CMA of 321.3. She developed appropriate shock from implantation of a cardiac resynchronization therapy device using a defibrillator 23 months later. Case 2 involved a 57-year-old woman with a CMV of 31.6, SUV_max of 5.8, and CMA of 101.6 who showed no cardiac uptake on FDG-PET re-examination after 12 months, with reduced values (CMV of 0.4, SUV_max of 2.7, and CMA of 0.9) (Fig 5).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Representative cases of patients with cardiac uptake.

Both cases demonstrated FDG cardiac uptake during the phase of stable medical and device therapy; however, their subsequent clinical courses differed substantially. Case 1: A 76-year-old woman with sustained ventricular tachycardia (VT), reduced left ventricular ejection fraction (LVEF = 22%), and an aneurysm in the left ventricular apex was diagnosed with systemic sarcoidosis based on lung tissue biopsy findings. She underwent implantation of a cardiac resynchronization therapy device with defibrillator (CRT-D). FDG-PET was performed 2 years after the initiation of corticosteroid therapy. Twenty-three months after therapy, CRT-D was appropriately activated in response to sustained VT. Case 2: A 57-year-old woman with non-sustained VT, mildly reduced LVEF (48%), and basal septal thinning was diagnosed with isolated cardiac sarcoidosis through additional evaluations using cardiac MRI and FDG-PET/CT. One year after the initiation of corticosteroid therapy, FDG-PET was performed. At a follow-up examination 12 months later, the FDG cardiac uptake had resolved. CMV, cardiac metabolic volume; CMA, cardiac metabolic activity; FDG-PET/CT, 18F-fluorodeoxyglucose-positron emission tomography/computed tomography; MRI, magnetic resonance imaging.

https://doi.org/10.1371/journal.pone.0347595.g005

Previous Studies on Prognostic Roles of FDG-PET

FDG-PET is valuable for detecting myocardial inflammation in cardiac sarcoidosis, a condition often difficult to diagnose through biopsy [19]. However, its role in assessing treatment response and predicting prognosis remains debated. A few studies have investigated the prognostic role of cardiac PET in sarcoidosis, and they may lack uniform patient backgrounds [11]. A meta-analysis identified six retrospective studies, but these studies often involved patients with suspected cardiac sarcoidosis or those already undergoing immunosuppressive therapy. The results typically showed worse outcomes in patients with cardiac FDG uptake, but this was expected as most patients in the uptake group had cardiac sarcoidosis, whereas many in the non-uptake group did not.

In a study by Gowani et al., 50 patients with cardiac sarcoidosis who underwent both cardiac MRI and FDG-PET/CT were evaluated. They found that late gadolinium enhancement on MRI was prognostically useful, but residual FDG uptake on PET was not [20]. Similarly, a large study in Japan concluded that the degree of FDG uptake was not linked to prognosis [6]. A few studies have quantitatively assessed FDG-PET findings before and after immunosuppressive therapy, though most involved fewer than 20 cases [21–23]. Rojulpote et al. analyzed FDG-PET findings of 83 patients with cardiac sarcoidosis and found that the treatment response varied, with no significant link between PET response and prognosis, which aligns with our findings. However, their study had limitations, including variability in immunosuppressive therapy doses and timing of FDG-PET scans [24]. Kaneta et al. qualitatively evaluated 80 patients with cardiac sarcoidosis (68 with PET and 12 with gallium scintigraphy) and reported that recurrent radiologic activity was frequently observed during follow-up. They also demonstrated that intensified immunosuppressive therapy suppressed this recurrent activity and was associated with more favorable outcomes in selected patients with active inflammation. However, our clinical approach differs in that we do not routinely intensify immunosuppression solely on the basis of residual FDG uptake during the phase of stable medical and device therapy. In patients without high-risk features—such as reduced LVEF, progressive ventricular dysfunction, or malignant arrhythmias—the potential risks of excessive immunosuppression, including infection and steroid-related adverse effects, must be carefully considered. Our findings suggest that, in the phase of stable medical and device therapy for cardiac sarcoidosis, long-term prognosis is largely determined by irreversible myocardial damage rather than by residual metabolic activity alone. Accordingly, treatment escalation should be individualized rather than uniformly applied to all patients with residual FDG uptake [25]. Imamura et al. showed that higher SUVmax after immunosuppressive therapy and impaired left ventricular ejection fraction predicted adverse outcomes in cardiac sarcoidosis. While their study incorporated serial pretreatment and follow-up FDG-PET data, our study focused on patients who appeared clinically stable on medical and device therapy during outpatient follow-up and did not include pretreatment imaging [26]. Our cohort included patients diagnosed before FDG-PET became widely used in clinical practice, allowing inclusion of a larger single-center cohort and detailed quantitative assessment of myocardial FDG uptake on follow-up FDG-PET. Despite differences in study design, both studies consistently identified impaired left ventricular systolic function as a major prognostic determinant, and prior long-term data indicate that patients clinically stable on medical and device therapy remain at substantial risk.

Physiological Implications of FDG Uptake in the Phase of Stable Medical and Device Therapy after Initiating Immunosuppressive Therapy

A central question in our study is the physiological meaning of metabolic activity observed on follow-up PET. Generally, this is thought to reflect persistent active granulomatous inflammation. However, while FDG is a marker for the enhanced glucose utilization characteristic of inflammatory cells, it is known that hibernating or stunned myocardium can also exhibit increased glucose uptake. Therefore, it is possible that the observed activity reflects a non-specific metabolic shift in scarred or dysfunctional myocardial tissue. Distinguishing between these possibilities could be facilitated by complementary imaging with iodine-123-beta-methyl-iodophenyl-pentadecanoic acid (¹²³I-BMIPP) scintigraphy, which assesses fatty acid metabolism, although this was not performed in our cohort.

Nevertheless, several factors support the interpretation of residual uptake as primarily inflammatory. For instance, it has been reported that a reduction in FDG uptake was associated with an improvement in LVEF, suggesting that FDG activity is linked to a reversible pathological process such as inflammation rather than a fixed scar [27]. This aligns with our observation, that is, a significant number of patients showed spontaneous resolution of FDG uptake without treatment escalation. Collectively, our findings suggest that inflammatory activity in stable CS can be dynamic, with the potential to wane over time even under low-dose maintenance therapy.

Our finding of risk dissociation, where FDG uptake suggests short-term but not long-term events, may reflect the different pathophysiological roles of active inflammation and established scar. We hypothesize that the observed residual metabolic activity represents a non-progressive, smoldering inflammation. This inflammation may be insufficient to drive long-term ventricular remodeling but could transiently increase myocardial instability, predisposing patients to adverse cardiac events such as arrhythmias and heart failure exacerbations in the short term. The spontaneous resolution of FDG uptake observed in a significant portion of our cohort supports this hypothesis of a dynamic self-limiting inflammatory state, distinct from the fixed scar-mediated risk that dictates long-term prognosis.

Definition of Relapse and Treatment Strategies for Cardiac Sarcoidosis

An international survey on the treatment of cardiac sarcoidosis revealed significant variability in corticosteroid management strategies, such as dosing, tapering, duration, and adjunctive immunosuppressive agents [28]. This reflects differences in treatment approaches across institutions. The definition of relapse and how to adjust immunosuppressive therapy in such cases also varies widely [28]. At our institution, metabolic activity on PET during the phase of stable medical and device therapy after initiating immunosuppressive therapy does not lead to dosage increase or addition of other immunosuppressive agents. Instead, therapy intensification is based on factors such as LV function and life-threatening ventricular arrhythmias. Even in patients with resolved FDG uptake and stable clinical courses, corticosteroids are maintained as lifelong therapy. Our findings support this conservative approach in patients clinically stable on medical and device therapy, though further prospective validation is warranted.

This study has certain limitations. First, it was a single-center retrospective analysis, which limits the generalizability of the findings. A multicenter prospective study is needed to validate and expand these results, potentially providing a broader dataset and stronger evidence for clinical decision-making. Second, the relatively small sample size (n = 79) may reduce the robustness of the multivariable analysis and increase the risk of overfitting. Therefore, the results should be interpreted as exploratory and require validation in larger cohorts. Third, pre-treatment FDG-PET was not uniformly available because a subset of patients were diagnosed before FDG-PET/CT became available and were evaluated with gallium scintigraphy instead. Therefore, interval changes in FDG uptake from treatment initiation could not be systematically assessed, limiting interpretation regarding treatment response or relapse. However, the primary aim of this study was not to evaluate treatment response but to assess the prognostic significance of FDG uptake during the phase of stable medical and device therapy (≥12 months after treatment initiation); thus, the absence of baseline FDG-PET does not affect the validity of our main findings. Furthermore, we acknowledge that the use of echocardiography for evaluating longitudinal changes in left ventricular function is subject to inter-observer variability, and cardiac MRI is generally considered a more objective modality. However, restricting our analysis solely to MRI-derived parameters was not feasible in this real-world cohort. A substantial proportion of our patients had advanced conduction abnormalities or severe arrhythmias requiring implanted cardiac devices (e.g., pacemakers and ICDs), which often preclude serial MRI assessments or result in non-diagnostic image quality due to artifacts. Because transthoracic echocardiography remains the standard, guideline-recommended modality for routine longitudinal follow-up, its inclusion in our composite endpoint was essential to accurately reflect real-world clinical disease progression and management. Finally, we acknowledge the potential for immortal time bias inherent to our study design. As all patients were required to undergo FDG-PET during the phase of stable medical and device therapy (≥12 months after initiating immunosuppressive therapy), those who experienced early adverse events or were unable to complete imaging were inherently excluded. This selection process may have led to an underestimation of event rates in higher-risk patients and limits generalizability.

Nonetheless, this design enabled us to specifically examine the clinical significance of FDG uptake during the post-treatment phase of stable medical and device therapy, which remains a relevant but underexplored area. These limitations underscore the need for future research with standardized imaging protocols before and after immunosuppressive therapy, as well as consideration of institutional differences in imaging modalities and therapeutic strategies.