Cardiac fibroblast activation detected by Ga-68 FAPI PET imaging as a potential novel biomarker of cardiac injury/remodeling

Background Fibroblast activation protein (FAP) as a specific marker of activated fibroblasts can be visualized by positron emission tomography (PET) using Ga-68-FAP inhibitors (FAPI). Gallium-68-labeled FAPI is increasingly used in the staging of various cancers. In addition, the first cases of theranostic approaches have been reported. In this work, we describe the phenomenon of myocardial FAPI uptake in patients who received a Ga-68 FAPI PET for tumor staging. Method and results Ga-68 FAPI PET examinations for cancer staging were retrospectively analyzed with respect to cardiac tracer uptake. Standardized uptake values (SUV) were correlated to clinical covariates in a univariate regression model. From 09/2018 to 11/2019 N = 32 patients underwent FAPI PET at our institution. Six out of 32 patients (18.8%) demonstrated increased localized myocardial tracer accumulation, with remote FAPI uptake being significantly higher in patients with vs without localized focal myocardial uptake (SUVmax 2.2 ± .6 vs 1.5 ± .4, P < .05 and SUVmean 1.6 ± .4 vs 1.2 ± .3, P < .05, respectively). Univariate regression demonstrated a significant correlation of coronary artery disease (CAD), age and left ventricular ejection fraction (LVEF) with remote SUVmean uptake, the latter with a very strong correlation with remote uptake (R2 = .74, P < .01). Conclusion Our study indicates an association of CAD, age, and LVEF with FAPI uptake. Further studies are warranted to assess if fibroblast activation can be reliably measured and may be used for risk stratification regarding early detection or progression of CAD and left ventricular remodeling. Electronic supplementary material The online version of this article (10.1007/s12350-020-02307-w) contains supplementary material, which is available to authorized users.

Conclusion. Our study indicates an association of CAD, age, and LVEF with FAPI uptake. Further studies are warranted to assess if fibroblast activation can be reliably measured and may be used for risk stratification regarding early detection or progression of CAD and left ventricular remodeling. (J Nucl Cardiol 2021;28:812-21.)
Conclusión. Nuestro estudio indica una asociación de CAD, edad y FEVI con la captación de FAPI. Se necesitan más estudios para evaluar si la activación de fibroblastos se puede medir de manera confiable y se puede usar para la estratificación de riesgo con respecto a la detección temprana o la progresión de la CAD y la remodelación ventricular izquierda. (

INTRODUCTION
Activation of fibroblasts is mandatory for repair and regeneration after myocardial injury including myocardial infarction (MI), progressive heart failure, or chemotherapy-induced injury, with the main mechanism of extracellular matrix remodeling. 1,2 Several pathways and mediators are suggested to play a role in fibrosis development, among others the renin-angiotensin-aldosterone (RAA) system, fibrogenic growth factors like transforming growth factor beta (TGF-β), or inflammatory cytokines and chemokines. 3,4 In addition, fibroblast activation protein (FAP) alpha has been demonstrated to be a specific marker of activated mature fibroblasts. 5 The fibroblast activation protein, also known as dipeptidyl peptidase (DPP) 4, is an enzyme and represents a serine proteinase. During embryonic development, FAP expression can be detected in a variety of tissues. 6 However, in adult humans, intensive expression is only found in the setting of wound healing, in fibrotic remodeling processes like in liver fibrosis and cirrhosis 7 and in the stroma of the majority of malignant cancer types. [8][9][10] The highest FAPI uptake was reported in sarcoma, esophageal, breast and lung cancer, and cholangiocarcinoma pointing at the dominant expression of FAP in solid tumors. 9 A study analyzing brain samples could even demonstrate increased FAP expression in high-grade gliomas. 11 For this reason, it is not only a very promising target in the diagnosis and potentially also in radio-ligand therapy of various tumor types, but it also represents a highly interesting target to diagnose and monitor tissue alterations such as remodeling processes due to different noxious agents including ischemia, chemo-, or radiation therapy. Recently, a Ga-68-labeled tracer has been developed, which has already been increasingly used in the diagnosis and staging of malignancy. 12 The quinolone-based FAP inhibitors can be labeled using the chelator DOTA. Of the Ga-68labeled FAP inhibitors described, FAPI-04 has been See related editorial, pp. 822-824 considered the most promising one due to its characteristics such as low nano-molar affinity, an almost complete internalization of more than 90% and rapid blood clearance. Advantages of FAPI over FDG are, among other things, that it represents a more specific signal with very low uptake in healthy organs and thus a very low background, which also provided the possibility for first theranostic concepts using Y-90 FAPI. 12 Furthermore, FAPI could be used in tumors that often have no or only limited increased FDG uptake resulting in a low sensitivity, such as pancreatic cancer or hepatocellular carcinoma. 13 As there is no approved tracer so far, FAPI can only be used on a compassionate use basis. As many aspects of ventricular fibrosis development are still unclear, 3 this novel imaging technique has the potential to significantly improve the understanding of ventricular fibrosis development under various clinical settings (e.g., cardiac side effects of modern cancer therapy). The latter consists of radio-, conventional chemo-, immuno-, and targeted therapy. 14 Heart failure is among the most common side effects of these multimodal approaches and has been related to the development of fibrosis. 15 Early detection of myocardial remodeling and fibrosis may be critical to prevent development of overt heart failure.
Aim of this study was to descriptively assess patterns of cardiac Ga-68 FAPI uptake in consecutive patients having undergone this imaging technique for staging after cancer treatment. Furthermore, we aimed to link this tracer uptake to clinical characteristics with respect to a cardiac disease history of the study cohort.

Patient Enrollment
We retrospectively analyzed patients having undergone FAPI PET imaging for cancer staging at the University Hospital Essen, Germany. According to the local standard, patients receiving anti-cancer treatment undergo cardiac diagnostic work-up before chemotherapy as well as half-annually after initiation of therapy to identify potential complications at an early stage. This standardized diagnostic work-up includes assessment of medical records, assessment of NYHA state and medical history by direct questioning as well as transthoracic echocardiography. Baseline characteristics comprised age at FAPI scan, sex, history of atrial fibrillation, coronary artery disease (CAD) including a history of MI, left ventricular ejection fraction (LVEF), cardiovascular risk factors, chronic kidney disease, and cancer entity including applied specific chemo-and immunotherapy. Anti-cancer treatment included alkylating agents, antibodies, anthracycline and platin derivatives, topoisomerase inhibitors, antimetabolites, and taxanes. Radiation therapy plans were used to calculate the heart dose in patients who received radiation during the course of cancer therapy. All relevant information were collected in a self-designed encrypted database, with all patients having signed informed consent before FAPI scan. All procedures performed were in accordance with the ethical standards of the institutional review board (IRB) of the Medical Faculty of the University Duisburg-Essen and with the principles of the 1964 Declaration of Helsinki and its later amendments.

FAPI Imaging
PET was performed 12 ± 7 minutes after injection of 140 ± 24 MBq Ga-68 FAPI-04 as previously described. 13 Acquisitions were performed on one of the following scanners: Biograph mCT PET/computed tomography (CT) (N = 19), Biograph Vision PET/CT (N = 10), or Biograph mMR PET/MRI (N = 3), all Siemens Healthcare GmbH, Erlangen, Germany. 3D image reconstruction (2 9 2 9 2 mm voxel size) of PET data assessed by PET/MRI was performed using ordinary Poisson ordered subset expectation maximization with 3 iterations and 21 subsets, a Gaussian filter with 4.0 mm full-width at half-maximum, and a 344 9 344 image matrix. Attenuation correction of the acquired PET data was performed using a five-compartment model attenuation map calculated from fat-only and water-only Dixon-based sequences by segmentation into background, lung, fat, and soft tissue with the addition of bone using an atlas. In the case of contrast-enhanced acquired images with MRI, only PET-based image analysis was performed because of the small sample size and because no heart-specific images were available in all cases. In the case of CT imaging, low-dose CTs for the purpose of attenuation correction and morphological correlation using automated tube voltage selection (CareKV, preset 120 kV, slice thickness: 5 mm) and CareDose 4D preset with 40 mAs were performed. PET images acquired on PET/CTs were reconstructed using an ordered subset expectation maximization (OSEM) algorithm, including time-of-flight information, with 4 iterations and 8 subsets. A Gaussian filter kernel with a full width at half maximum of 4 mm was used for postreconstruction filtering.

Image Interpretation
In the case of focal tracer uptake in the left ventricular myocardium, a volume of interest (VOI) was defined comprising the entire focal tracer accumulation. Uptake of remote myocardium and blood pool was assessed by placing a spherical 1 cm 3 VOI in the lateral wall of the left ventricle or the right atrium. From these VOIs, the maximum standardized uptake value (SUV max ) and the mean SUV (average uptake in a spheric VOI of 1 cm 3 surrounding the hottest voxel) were measured. In the case of focal tracer accumulation in the lateral wall, cardiac remote uptake has been measured by placing this spherical VOI in the septum (there was no case with tracer uptake both in the lateral wall or the right atrium). All data were analyzed using dedicated medical imaging software (syngo.via, Siemens Healthcare GmbH, Erlangen).

Statistical Methods
Continuous variables are expressed as mean ± standard deviation unless otherwise annotated. A Kolmogorov-Smirnov test was conducted to test for normality where indicated. Non-parametric and parametric variables were compared using the Mann-Whitney U test and an unpaired t test, respectively, Chi-square test was applied to compare dichotomized variables. SUV mean and SUV max were used for further assessment of variables associated with remote FAPI uptake in a univariate linear regression model. To compare matched continuous variables, the 2-tailed paired Student t test was applied. A P value of less than .05 was considered statistically significant. Statistical analyses were performed using SPSS for Windows (Version 22.0, SPSS Inc., Chicago, IL, USA).
Also, FAPI uptake in remote myocardium was higher in patients with vs without focal myocardial tracer uptake (SUV max 2.2 ± .6 vs 1.5 ± .4, P \ .05, SUV mean 1.6 ± .4 vs 1.2 ± .3, P \ .05). Blood pool was not different between patients with vs without focal myocardial tracer uptake (SUV max 2.2 ± .7 vs 2.1 ± .7, P = .6; SUV mean 1.4 ± .4 vs 1.4 ± .5, P = .8). Table 1 lists the baseline characteristics of the study cohort stratified for a visual presence of focal FAPI uptake. With respect to cancer entity and applied chemo-and immunotherapy, no significant differences regarding focal cardiac FAPI uptake were observed except for a slightly higher percentage of patients with pancreatic cancer and a history of platin/antimetabolite chemotherapy in the non-uptake group. Four out of 32 patients (12.5%) had undergone radiation therapy in the course of cancer treatment. One patient received thoracic radiation in 04/2015 and 07/2018 for local therapy of right-sided non-small cell lung cancer (right lower lobe) with a cumulative dose of only 1.04 Gy. This patient showed no focal FAPI tracer uptake. In the remaining three patients the heart dose was 0 Gy.

Association of Myocardial FAPI Uptake with Patient Characteristics and Clinical Parameters
Our data suggest that patients with focal FAPI uptake were older (70.8 ± 10.1 vs 56.0 ± 14.6 years, P = .03), with a lower LVEF (46.0 ± 8.5 vs 60.1 ± 4.2; P \ .01) and a higher percentage of significant CAD (50.0% vs 0%; P \ .01) compared to non-uptake patients. The exact location of visually detectable tracer uptake in the 6 patients is reported in Table 2. Looking at the patients with a known diagnosis of CAD it appears that the location of the primarily treated or affected myocardium (coronary region treated by percutaneous intervention or location of MI) in these 3 patients correlated well with the area of visual FAPI accumulation. In opposite, no patient in the FAPI-negative group (patients without localized uptake) showed a history of CAD or MI (P \ .01 and P = .03 for CAD and MI, respectively). Figures 1 and 2 depict representative examples of 2 patients with localized myocardial FAPI uptake.

DISCUSSION
This is the first in-human study describing the pattern of cardiac FAPI uptake in a cohort of patients who underwent Ga-68 FAPI PET imaging for cancer staging. The main findings are that patients with visual detectable (focal) tracer uptake (6/32 patients, 18.8%) were older, with a higher prevalence of CAD/MI and lower LVEF. In patients with CAD, the visual tracer uptake seemed to correlate with the area of primarily diseased myocardium. In patients with increased uptake in certain myocardial areas, an increased uptake was also found in the remote myocardium, while no difference in blood pool activity was detected. Remote myocardial uptake was correlated to CAD, age, and LVEF, the latter showing a highly significant correlation. Fibroblast activation protein is not expressed in normal cardiac fibroblasts 16 ; nevertheless, cardiac FAP activation has been reported under various pathologic conditions in animal models. 5,17 The possibility of detecting fibroblast activity by FAPI PET/CT after MI has already been shown in a murine model. 10 In our present in-human experience on cardiac FAPI tracer accumulation we could demonstrate that in an unselected cohort of individuals with malignancy about 20% of subjects showed distinct cardiac tracer accumulation. This raises the question if this localized tracer accumulation could be attributed to prior myocardial injury due to the chemotherapy applied in the course of cancer treatment as preliminary data suggested a potential role of chemo-therapy-induced cardiotoxicity on myocardial FAPI enrichment. 18 Despite a medical history of several chemotherapeutic drugs applied in the study cohort we Table 2. were not able to link the chemotherapy regimens to distinct FAPI uptake. Detailed analysis of the 20% of patients of the present study with localized tracer accumulation revealed that those patients were older, with a lower LVEF and higher prevalence of CAD. This goes in line with clinical and preclinical data suggesting that LV function, 19,20 age 21 , and especially CAD/MI 10,17 are associated with myocardial fibrosis. With respect to localized tracer uptake we hypothesize that lower EF and higher age show a significant interaction with CAD, as patients with CAD were older and had a significantly lower EF (73.0 vs 57.2 years and 46.0 vs 60.1%), respectively. In the three patients with CAD, the area of mainly diseased myocardium correlated well with the localized FAPI uptake in those patients. This is of particular relevance considering that following myocardial injury due to myocardial ischemia, FAP-α has been shown to be enriched in adjacent myocardium (close to infarcted myocardium, with reduced function and perfusion) compared to remote myocardium (preserved function and perfusion) in a pig model of myocardial infarction. 5 This leads to the hypothesis that visualization of FAP expression could help to assess the extent of myocardial damage following MI in progressive CAD and to further assess the extent of cardiac re-modeling following such ischemic events. Large controlled studies are warranted to investigate this hypothesis.

Potential Significance of FAPI Uptake in Remote Areas
We demonstrate that fibroblast activation in the remote myocardium is also significantly increased in patients with focally increased uptake of certain myocardial areas. This points at left ventricular remodeling processes in those patients besides the localized uptake areas. At the moment we cannot draw a final conclusion on the mechanisms behind higher fibroblast activation in remote areas as the patient number in the present study was low. We speculate that potentially a generalized myocardial injury, as it is the case in microvascular dysfunction present in CAD, could be causative for this phenomenon. This cannot be finally proven at the moment as there are not conclusive histopathological studies validating cardiac FAPI uptake with fibroblast activation as it is the case for FAPI uptake in malignancy. 9,10 Fibroblast activation may also play a role in generalized fibrotic remodeling, as previously shown in heart failure, where the deposition of extracellular matrix proteins increasing myocardial stiffness 1 is primarily driven by cardiac fibroblasts. 22 As fibrosis is involved in the disruption of myocardial excitation-  Bold P values indicate statistical significance R, correlation coefficient; SUV, standardized uptake value; other abbreviations as in Table 1 contraction coupling this is suggested to play a key role in the development of both systolic and diastolic heart failure. 22 Especially looking at the complex entity of heart failure with preserved EF, this imaging modality might bear the potential of increasing diagnostic yield, further guiding specific therapy in the future.

Limitations
This is a retrospective study of FAPI PET in a heterogeneous population of patients. Therefore, the data must be treated with caution and do not allow to draw general conclusions. Some limitations have to be mentioned. First, despite preclinical work showing highly specific binding to FAP, 10 we cannot conclusively prove that myocardial FAPI accumulation is specific for FAP expression or fibroblast activation as we are not able to provide histopathological validation or blocking study data. We acknowledge the limitation of not having profound data on diastolic function in our cohort in order to further investigate the impact of this confounder on our results. As no routine invasive or non-invasive screening for CAD has been performed we are not able to entirely rule out unidentified myocardial ischemia as potential confounder for localized FAPI uptake in patients without a known history of CAD. Last, as this is a pilot study on cardiac FAPI uptake, scans were assessed on a consensus decision of the nuclear medicine specialists. As the analysis was clearly defined and highly standardized requiring only minimal user interaction, a high interobserver reliability can be assumed.

CONCLUSION
This is the first report to assess cardiac FAPI uptake in cancer patients. Higher age, a history of CAD/MI as well as impaired LVEF seem to be associated with increased localized uptake. These results point towards the potential role of FAPI imaging to assess myocardial localized as well as generalized injury. Further research efforts are needed to clearly link myocardial FAPI uptake to myocardial fibroblast activation. Finally, studies are warranted to assess if FAPI uptake might be used to risk-stratify patients with respect to a progression of CAD or left ventricular remodeling.

New Knowledge Gained
This pilot study demonstrates that cardiac fibroblast activity can be assessed by Ga-68 FAPI PET imaging. Fibroblast activity correlated with CAD, MI, age, and especially LVEF, pointing at the potential of this technique to significantly improve the understanding of structural/fibrotic remodeling processes in cardiac pathophysiology.