Fibroblast activation protein imaging in atrial fibrillation: a proof-of-concept study

To evaluate the feasibility of using radiolabeled fibroblast activation protein inhibitor (FAPI) PET/CT imaging to assess activated fibroblasts in the atria of individuals with AF and to identify factors contributing to enhanced atrial activity. We constructed left atrial appendage (LAA) pacing beagle dog AF models (n = 5) and conducted 18F-FAPI PET/CT imaging at baseline and eight weeks after pacing. Right atrial (RA) specimens were collected from these models. Additionally, 28 AF patients and ten age- and sex-matched healthy volunteers underwent 18F-FAPI PET/CT imaging. RA of AF beagles showed increased 18F-FAPI uptake. Among AF patients, 18 out of 28 (64.3%) exhibited enhanced atrial FAPI activity. No atrial 18F-FAPI uptake was observed in the sham beagle and healthy volunteers. In animal RA specimens, 18F-FAPI activity correlated positively with FAP mRNA (r = .98, P = .002) and protein (r = .82, P = .03) levels, as well as collagen I mRNA expression (r = .85, P = .02). B-type natriuretic peptide levels were associated with atrial 18F-FAPI activity (OR = 3.01, P = .046). This proof-of-concept study suggests that 18F-FAPI PET/CT imaging may be a feasible method for evaluating activated fibroblasts in the atria of AF patients.


INTRODUCTION
Atrial fibrillation (AF) is the most prevalent arrhythmia and is associated with diverse pathophysiological mechanisms. 1 AF is characterized by structural, contractile, and electrical remodeling in the atrium, with atrial fibrosis being a prominent pathological feature of structural remodeling. 2 Studies have demonstrated that atrial fibrosis plays a crucial role in the initiation, maintenance, and recurrence of AF, as well as an increased risk of stroke and heart failure. 3,4Furthermore, there is a strong correlation between the severity of fibrosis and the recurrence rate following radiofrequency ablation, where increased fibrosis grades are associated with higher rates of recurrence. 5Therefore, there is a growing research interest in atrial fibrosis, necessitating the development of reliable techniques for in vivo assessment of fibrotic progression.
Although late gadolinium enhancement (LGE) magnetic resonance imaging (MRI) has been utilized to evaluate atrial fibrosis, 6,7 its application is limited to certain centers and suffers from low reproducibility due to the lack of standardized image protocols.Moreover, LGE primarily detects advanced fibrosis and cannot provide comprehensive information on early fibrotic changes during atrial remodeling.
Myocardial fibrosis is a dynamic process mediated by the activation of myocardial fibroblasts, characterized by the breakdown of extracellular matrix homeostasis and excessive collagen deposition. 8Fibroblast activation is closely associated with early myocardial fibrosis, and fibroblast activation can be identified by the specific expression of membrane-anchored peptidase-fibroblast activation protein (FAP).Recently, radiolabeled fibroblast activation protein inhibitor (FAPI)-targeting tracers have been developed for the detection of activated fibroblasts. 68][9] It is plausible that this technique could also be applied to the atria, enabling the assessment of dynamic fibrotic progression by providing instant information on activated fibroblasts.
In the context of AF, several innovative studies have employed 68 Ga-FAPI PET/CT imaging to assess fibroblast activation in the pulmonary veins after ablation, serving as surrogate markers for tissue damage. 10,11lthough these studies have provided valuable insights, their focus has predominantly been on the effects of ablation therapy.Consequently, there is a knowledge gap regarding fibroblast activation in a broader AF patient population, particularly among those who have not undergone ablation.
In this proof-of-concept study, our objectives were to: (1) explore the feasibility of 18 F-FAPI PET/CT imaging for evaluating activated fibroblasts in the atria of AF, supported by histological evidence from animal models; and (2) preliminarily investigate the factors associated with atrial 18 F-FAPI activity in AF patients.

Animal model study
All experiments were approved by the Animal Ethics Committee of Capital Medical University (2021ke-515) and performed in accordance with the U.S. National Institutes of Health Guide for the Care and Use of Laboratory Animals (publication no.85-23, 1996).
The beagle dog models of AF (n = 5) were constructed via rapid atrial pacing (Supplemental Method 1). 12 After left lateral thoracotomy, the left atrial appendage (LAA) was exposed and an electrode was fixed at its base for stimulation.The pacemaker (output: 1.5 V, pulse duration: 2.0 ms; Ensen-1ES?1ST,Ensen, Shanghai, China) was set at 600 beats/min.The sham group (n = 2) underwent the same pacemaker implantation procedure without atrial pacing. 18F-FAPI PET/CT imaging were performed after 8 weeks of continuous atrial pacing. 18F-FAPI PET/CT image acquisition and analysis 18 F-FAPI was radiolabeled as previously described, 13 and PET/CT images were acquired 60 min after injection (3.7 MBq/kg) using a 16-slice PET/ CT scanner (Discovery STE, GE, USA).After a scout CT acquisition (120 kV, 10 mA) used for heart positioning, CT transmission scanning (140 kV, 120 mA) was performed for attenuation correction and anatomical localization.A PET scan was acquired immediately after the CT scan, with a scanning time of 10 minutes.
On the animal PET/CT images, we selected the right atria (RA) for analysis, to avoid the potential influence of surgery to the left atrial activity.To measure visible uptake (higher than the activity of the blood pool) of the right atrial wall, regions of interest (ROI) were carefully placed on the PET transverse axial images, guided by CT.The maximum standardized uptake value (SUVmax) out of all RA slices was taken as the representative right atrial activity.In cases where visible uptake was not observed, a 5 mm diameter circular ROI was placed on the right lateral wall of the RA, at the level of the aortic root.Concurrently, for determining the background value of 18 F-FAPI uptake, a circular ROI of similar size was placed in the right atrial cavity and the mean SUV (SUVmean) was recorded.To calculate the target-to-background ratio (TBR) of the RA, the SUVmax of three consecutive atrial levels was first averaged.This mean value was then divided by the blood pool SUVmean to arrive at the TBR.In addition, comparison was made between the TBR of the left ventricle (LV) and that of the atria.For the AF models, TBR values of the RA exceeding 1.96 standard deviations (SD) beyond their baseline mean value were classified as abnormal.
Histological analysis After imaging, RA samples were retained for further analysis.Immunofluorescence staining of activated fibroblasts was performed in the atrial samples.All samples were fixed with 4% paraformaldehyde overnight.Paraffin sections of the samples were then incubated with FAP antibody (Abcam, ab53066), Vimentin antibody (Abcam, ab92547), and a-smooth muscle actin antibody (a-SMA, Abcam, ab7817).The following day, secondary Alexa Fluor-conjugated antibodies diluted in the same carrier solution (1:200) were added to the sections for 1 h at room temperature.The stained sections were collected by a fluorescence microscope.Quantitative polymerase chain reaction (PCR) and western blot were used to evaluate the mRNA and protein levels of FAP and collagen I in the RA (Supplemental Methods 2-5).

Human study
Study population This prospective study received approval from the Institutional Ethical Committee of Beijing Chaoyang Hospital (2021-ke-247) and adhered to the Declaration of Helsinki.Between February 2021 and March 2022, 28 patients with AF (20 men, mean age: 61 ± 13 years) who underwent hybrid surgical-catheter ablation at the Cardiac Center of Beijing Chaoyang Hospital were prospectively enrolled.All patients underwent baseline 18 F-FAPI PET/CT imaging and echocardiography before ablation procedures.
A control group of 10 age-and sex-matched healthy volunteers (7 men, mean age: 58 ± 9 years) was recruited to establish the normal range of cardiac FAPI activity, following the same PET/CT imaging procedures as those for AF patients.The inclusion criteria included no history of AF or other arrhythmias, no history of cardiovascular disease, no history of malignancy, and no abnormal findings on PET/CT imaging.Informed consent was obtained from all subjects.
For human 18 F-FAPI PET/CT images, FAPI activities in different atrial structures were analyzed using a quantitative method described previously. 14Briefly, visible atrial wall uptake on each PET trans-axial image was measured by placing ROIs under CT guidance.The SUVmax of all slices was chosen to represent atrial activity.If no visible uptake was observed, 18 F-FAPI activity of the RA was analyzed as in the animal studies.For the LA, a circular 5 mm diameter ROI was placed on the right lateral wall of the LA at the level of the right inferior pulmonary vein.To obtain a background value of FAPI uptake, an ROI was placed on the left and right atrial cavities, and the SUVmean was recorded.Subsequently, a TBR was calculated for bilateral atria.The maximum TBR of all slices in both the left and right atria was chosen to represent atrial activity.For AF patients, values exceeding mean ?1.96 SD of the controls were considered abnormal.
Echocardiography Transthoracic echocardiography images were acquired by echocardiologists, who were blinded to the 18 F-FAPI imaging data, using a Vivid E95 echocardiographic system (General Electric Ultrasound, Milwaukee, WI, USA).LA volume index (LAVi) and RA area were employed to evaluate atrial size.

Statistical analysis
Statistical analysis was conducted using SPSS Statistics (Version 26.0; IBM).The Kolmogorov-Smirnoff test was utilized to evaluate the data's normality distribution.Continuous variables are presented as the mean ± SD or medians with interquartile ranges, while categorical variables are expressed as absolute numbers and percentages.Comparisons between two groups were executed using Student's t-test, while chi-square tests or Fisher's exact tests were employed for comparing frequencies.Pearson or Spearman correlation analysis was conducted to explore the correlation between the RA 18 F-FAPI activity and the mRNA and protein expression of FAP, as well as the mRNA expression of collagen I. Univariate and multivariate logistic regression models were used to identify relevant factors for increased 18 F-FAPI uptake in atria.For animals, values exceeding 1.96 SD from their baseline mean value were considered abnormal.For humans, values surpassing 1.96 SD from the control mean were deemed abnormal.A P value \ .05indicated statistical significance.

Animals study
All five beagles with AF displayed a visually increased accumulation of 18 F-FAPI in the RA compared to their baseline [1.8 (1.2, 3.4) vs 0.9 (0.9, 1.0), P = .014],while no significant difference was observed in the LV (1.7 ± 0.5 vs 1.3 ± 0.2, P = .162).No atrial 18 F-FAPI uptake was detected in the sham group by 18 F-FAPI imaging and autoradiography (Figure 1A).FAP? fibroblasts were observed in RA samples from beagle models with AF but were rare in the sham groups (Figure 1B).Additionally, the mRNA and protein expression of FAP, Vimentin, a-SMA, and collagen I were significantly increased in RA samples from beagle models with AF (Figure 2A-C) as well. 18 FAPI activities in the RA positively correlated with the corresponding mRNA (r = .98,P = .002)and protein expression of FAP (r = .82,P = .03),as well as with the mRNA expression of collagen I (r = .85,P = .02)(Figure 2D).

Human study
Baseline characteristics of patients are shown in Table 1.A total of 28 patients with AF (20 persistent AF, 8 paroxysmal AF; 20 men, mean age: 61 ± 13 years) and 10 age-and sex-matched healthy volunteers (7 men, mean age: 58 ± 9 years) were prospectively enrolled.

DISCUSSION
Fibroblasts, the primary cells regulating cardiac fibrosis, undergo phenotypic differentiation into the activated type under pathological conditions, initiating the fibrotic process. 15Several pathological triggers of AF have been identified, including high-frequency electrical activity, mechanical stress, ischemia, and inflammation, [16][17][18] all of which contribute to the development of time-dependent atrial fibrotic remodeling.
By targeting the emblem of activated fibroblasts, we present the first evidence that 18 F-FAPI PET/CT can detect increased FAPI activity in the atria of AF and show a strong relationship between atrial activity and tissue-level expression of FAP and collagen I.These findings support the notion that 18 F-FAPI PET/CT is a histologically reliable imaging modality for assessing activated fibroblasts in the atria of AF.Our study bridges this gap by investigating the potential of 18 F-FAPI PET/CT imaging in assessing activated fibroblasts in the atria of AF patients.Our scope extends beyond existing research by incorporating patients who haven't undergone ablation and providing histological validation from animal models, and scrutinizing the factors influencing atrial 18 F-FAPI activity in AF patients.This broader perspective enhances our comprehension of fibroblast activation in AF, providing essential insights into this critical facet of AF pathophysiology.
Fibrillar collagen type I is the primary component of the cardiac extracellular matrix, accounting for approximately 80% of total collagen. 19Since collagen I is synthesized by fibroblasts as pro-collagen and FAP is the hallmark of activated fibroblasts and the target of the FAPI tracer, a strong relationship between 18 F-FAPI activity and collagen I mRNA expression is expected.We discovered that 18 F-FAPI activity was unrelated to the protein level of collagen I, which may be due to dynamic collagen degradation regulated by procollagenases.This suggests that atrial fibrotic alterations  detected by 18 F-FAPI imaging are not at the irreversible end-stage, providing guidance for determining the antifibrotic therapy window and selecting suitable candidates.
Our data revealed a connection between BNP level and 18 F-FAPI uptake.According to previous reports, 20,21 BNP correlates with the mean RA pressure and is upregulated in patients with persistent AF.Atrial wall stretching stimulates fibroblasts, so we assume that 18 F-FAPI uptake reflects increased atrial stretch, suggesting that atrial remodeling may also occur under prolonged AF burden.18 F-FAPI imaging, if fully validated in patients with AF, could offer several potential benefits to patient management: (1) Early detection of atrial fibrosis: Unlike late gadolinium enhancement (LGE) MRI, which primarily detects advanced fibrosis, FAPI PET has the potential to identify early fibrotic changes during atrial remodeling.This could enable clinicians to intervene at an earlier stage of the disease, possibly preventing the progression of atrial fibrosis and its associated complications.(2) Personalized treatment strategies: By accurately assessing the extent and progression of atrial fibrosis, FAPI PET could help tailor treatment strategies to individual patients, guiding decisions on antiarrhythmic medications, anticoagulation, or invasive procedures like catheter ablation.This personalized approach could potentially improve the overall effectiveness of AF management and reduce the risk of recurrent AF episodes.(3) Monitoring treatment response: FAPI PET could serve as a non-invasive tool to monitor the response to treatment in patients with AF.By evaluating changes in fibroblast activation and atrial fibrosis over time, FAPI PET may provide valuable information on the efficacy of various therapeutic interventions, allowing clinicians to modify treatment plans as needed.(4) Risk stratification: FAPI PET could potentially be used as a tool for risk stratification in AF patients, identifying those at higher risk for complications like stroke or heart failure due to the presence of significant atrial fibrosis.This information could guide the selection of appropriate therapeutic and preventive measures, ultimately improving patient outcomes.
This preliminary study has several limitations.First, the sample size was small, with only 28 patients with AF enrolled.Second, histological findings in patients with AF were not analyzed, because RA tissue samples were unavailable in AF patients.Third, the thin atrial walls make SUV measurement more influenced by partial volume effects, while partial volume correction is not applicable because atrial 18 F-FAPI uptake is invisible in many cases.Fourth, CT angiography was not performed, which may have affected more objectively defined of the regions of interest in the very thin atrial walls, especially when there was no 18 F-FAPI uptake visible.Fifth, the PET and PET/CT images with the CT alone are still far from definitive in the accurate placement of the ROIs, although the inclusion of the CT alone images with the PET solves, to some extent, the problem of the difficulty in accurately placing the regions of interest.Finally, follow-up 18 F-FAPI imaging was not performed in either the canine model or clinical patients, so the evolution of atrial 18 F-FAPI activity in dynamic atrial remodeling and the prognostic value of atrial 18 F-FAPI imaging remain unknown.A multicenter cohort study recruiting more patients with AF and conducting repeat 18 F-FAPI imaging over an extended follow-up period will provide more evidence on the clinical value of atrial 18 F-FAPI imaging.

CONCLUSIONS
In this preliminary investigation, we demonstrated that 18 F-FAPI imaging is a histologically accurate method for evaluating activated fibroblasts in the atria of AF, with potential for assessing dynamic fibrotic progression.Furthermore, our findings suggest a correlation between elevated BNP levels and increased FAPI uptake in the atrium.

NEW KNOWLEDGE GAINED
18 F-FAPI imaging may be a histologically accurate method for assessing fibroblastic activation in the atria and provide new insights into the remodeling of the AF.

Figure 1 . 18 F
Figure 1. 18F-FAPI PET/CT images and histological findings in beagle dog models of AF and sham.(A) Enhanced FAPI uptake in the right atrium of beagles with AF was demonstrated through 18 F-FAPI PET/CT imaging and autoradiography, in contrast to the absence of atrial FAPI uptake in the sham group (arrows indicate atrial 18 F-FAPI uptake with TBR values).(B) Masson and immunofluorescence staining for FAP, Vimentin, and a-SMA proteins in RA sections revealed extensive atrial fibrosis and activated fibroblasts in atrial samples from the beagle model with AF, which were scarce in the sham groups.AF, atrial fibrillation; FAPI, fibroblast activation protein inhibitor; TBR, target-to-background ratio.

Figure 2 .
Figure 2. Elevated mRNA and protein expression of FAP, Vimentin, a-SMA, and collagen I in the right atrium (RA) samples from beagle models with AF. (A-C) The mRNA and protein levels of FAP, Vimentin, a-SMA, and collagen I were significantly increased in the RA samples from beagle models with AF. (D) Correlation analysis revealed that the mRNA and protein expression of FAP, as well as the mRNA expression of collagen I, were strongly associated with the corresponding FAPI activities in the RA.AF, atrial fibrillation.

Figure 3 .
Figure 3. Increased 18 F-FAPI uptake was observed in the right atrium (RA) of patients with AF, but not in the healthy volunteers (arrows indicate atrial FAPI uptake with TBR values).FAPI, fibroblast activation protein inhibitor; AF, atrial fibrillation; TBR, target-to-background ratio.

Table 2 .
Comparisons between AF patients with-and without FAPI uptake

Table 3 .
Univariate regression analysis of risk factors for FAPI uptake