Clinical prospective study of Gallium 68 (68Ga)–labeled fibroblast-activation protein inhibitor PET/CT in the diagnosis of biliary tract carcinoma

This study is to investigate the [68Ga]Ga-DOTA-FAPI PET/CT diagnosis performance in biliary tract carcinoma (BTC) and analyze the association between [68Ga]Ga-DOTA-FAPI PET/CT and clinical indexes. A prospective study (NCT 05264688) was performed between January 2022 and July 2022. Fifty participants were scanned using [68Ga]Ga-DOTA-FAPI and [18F]FDG PET/CT and acquired pathological tissue. We employed the Wilcoxon signed-rank test to compare the uptake of [68Ga]Ga-DOTA-FAPI and [18F]FDG, and the McNemar test was used to compare the diagnostic efficacy between the two tracers. Spearman or Pearson correlation was used to assess the association between [68 Ga]Ga-DOTA-FAPI PET/CT and clinical indexes. In total, 47 participants (mean age 59.09 ± 10.98 [range 33–80 years]) were evaluated. The [68Ga]Ga-DOTA-FAPI detection rate was greater than [18F]FDG in primary tumors (97.62% vs. 85.71%), nodal metastases (90.05% vs. 87.06%), and distant metastases (100% vs. 83.67%). The uptake of [68Ga]Ga-DOTA-FAPI was higher than [18F]FDG in primary lesions (intrahepatic cholangiocarcinoma, 18.95 ± 7.47 vs. 11.86 ± 0.70, p = 0.001; extrahepatic cholangiocarcinoma, 14.57 ± 6.16 vs. 8.80 ± 4.74, p = 0.004), abdomen and pelvic cavity nodal metastases (6.91 ± 6.56 vs. 3.94 ± 2.83, p < 0.001), and distant metastases (pleural, peritoneum, omentum, and mesentery, 6.37 ± 4.21 vs. 4.50 ± 1.96, p = 0.01; bone, 12.15 ± 6.43 vs. 7.51 ± 4.54, p = 0.008). There was a significant correlation between [68Ga]Ga-DOTA-FAPI uptake and fibroblast-activation protein (FAP) expression (Spearman r = 0.432, p = 0.009), carcinoembryonic antigen (CEA) (Pearson r = 0.364, p = 0.012), and platelet (PLT) (Pearson r = 0.35, p = 0.016). Meanwhile, a significant relationship between [68Ga]Ga-DOTA-FAPI metabolic tumor volume and carbohydrate antigen199 (CA199) (Pearson r = 0.436, p = 0.002) was confirmed. [68Ga]Ga-DOTA-FAPI had a higher uptake and sensitivity than [18F]FDG in the diagnosis of BTC primary and metastatic lesions. The correlation between [68Ga]Ga-DOTA-FAPI PET/CT indexes and FAP expression, CEA, PLT, and CA199 were confirmed. clinicaltrials.gov: NCT 05,264,688.

Introduction malignant neoplasms worldwide in males and is the seventh highest among females [3]. Mostly, patients with poor prognoses are diagnosed in the middle or late stage, negatively impacting the treatment of BTC and patient recovery. Therefore, timely and accurate diagnosis and staging determination are crucial for improving patient outcomes.
High-quality imaging examination is a prerequisite for determining resectable and accurate preoperative staging. The primary imaging methods for diagnosing BTC are computer tomography (CT) and magnetic resonance imaging (MRI) [4]. However, many studies have found these methods less sensitive in judging lymph node metastases. And gadolinium acid-enhanced MRI could not provide satisfactory hepatobiliary phase images in hilar cholangiocarcinoma and distal cholangiocarcinoma patients with biliary obstruction [5,6]. Several articles have validated the use of positron emission tomography (PET) in diagnosing BTC because it displays important supplementary details missed on CT and MRI images [7,8]. Fluorodeoxyglucose labeled with radionuclide 18F ([ 18 F]FDG) is now the most frequently developed PET molecular imaging probe; nonetheless, [ 18 F]FDG is not tumor-specific. Although [ 18 F]FDG is helpful in diagnosing BTC, it has several limitations. For instance, the liver and gastrointestinal tract physiological uptake increases the rate of [ 18 F]FDG uptake, resulting in improved false positive measurements during BTC imaging [9][10][11][12]. Therefore, developing and designing novel molecular probes with high specificity for tumor imaging is critical.
Cancer-associated fibroblast (CAF), one of the dominating components of the tumor microenvironment, is a special fibroblast when activated. It promotes tumor proliferation, invasion, and metastases by secreting cytokines, growth factors, chemokines, and extracellular matrix degradation enzymes [13,14]. Many studies have stated that the fibroblast activation protein (FAP) is the hallmark of CAF as it is highly specific in active fibroblasts [15,16]. In addition, many studies have reported that FAP is overexpressed in more than 90% of cancers. The overexpression of FAP has been verified as an independent adverse prognostic factor in liver, colon, and lung cancer, making it a highly desirable molecular target for tumor diagnosis and treatment. As a result, FAP inhibitors (FAPI) have recently attracted extensive tumor imaging and treatment research. In most of these studies on FAPI, researchers have applied 68 Ga and other radionuclide-labeled FAPI molecular imaging probes in lung, breast, and stomach cancers [17,18]. The results show that FAPI has better outcomes than FDG [19][20][21]. Therefore, FAPI molecular imaging probes should have broad application prospects in BTC.
A study preliminarily explored the diagnostic performance of FAPI in 18 patients with BTC [22]. More cases and the association between [ 68 Ga]Ga-DOTA-FAPI uptake and clinical indexes need to be further researched for the clinical translation of [ 68 Ga]Ga-DOTA-FAPI, and additional clinical research is necessary to distinguish these scenarios. This study aims to analyze more cases to assess the potential diagnosis of [ 68 Ga]Ga-DOTA-FAPI PET/CT in BTC and investigate the correlation between [ 68 Ga]Ga-DOTA-FAPI PET/CT and clinical indexes.

Study design and participants
The prospective clinical study was conducted at the Zhongnan Hospital of Wuhan University with no financial support from commercial entities. This study was conducted per the Declaration of Helsinki and the national regulations. This clinical trial was registered on clinicaltrails.gov (NCT 05,264,688) and approved by the Clinical Research Ethics Committee of Zhongnan Hospital of Wuhan University. Before participating in this study, the participants must provide written informed consent.
This study recruited participants between January 2022 and July 2022, as shown in Fig. 1. Participants were included in this study if they met the following inclusion criteria: (1) volunteered and signed written informed consent; (2) 18-80 years old male or female; (3) suspected to have BTC based on medical history and CT/MRI examination; and (4) willing and able to participate in all research procedures. On the other hand, participants were excluded from the study if: (1) pregnant or breastfeeding; (2) allergic to alcohol; (3) had a history of other malignant tumors; (4) fasting blood glucose was over 11.0 mmol/L; (5) claustrophobic patients; and (6) failed to maintain a supine position for about 30 min. After recruitment, [ 68 Ga]Ga-DOTA-FAPI and [ 18 F]FDG PET/CT were executed within one week without sequencing rules. The mean time interval between the two studies was 2.6 days (range: 1-7 days). General and clinical information was collected from each participant, and pathological results acted as the gold standard of diagnosis. However, due to clinical and ethical limitations, not all lesions could be obtained for examination in this study, especially for the tumor infiltrated along the bile duct wall, risk area, lymph node, and distant metastases. The diagnosis reference standard was clinical and/or imaging follow-up data (mean, 6 months ± 0.8 [SD]; range, 3-9 months), including physical examination, laboratory tests, and medical imaging. Criteria for determining malignant lesions are as follows: typical malignant features determined by multiple imaging, such as numerous lymph nodes with short-axis diameter > 1 cm, CT demonstrated scan showed equal or slightly hypodense, and the enhancement was nodular or annular. Follow-up imaging showed obvious progression; the extent of the lesion was significantly reduced after treatment.

Preparation of [ 68 Ga]Ga-DOTA-FAPI and [ 18 F]FDG
This study synthesized [ 68 Ga]Ga-DOTA-FAPI using a fully automated synthesis module (Isotope Technologies Munich GmBH, Munich, Germany) in an aseptic ventilation environment. The FAPI precursor (DOTA-FAPI-04) was purchased from CSBio Ltd. (Shanghai, China). The radiolabeling process of the FAPI precursor was provided in the supplementary materials. The study employed several quality control parameters, including physical properties, pH value, bacterial endotoxin, and radiochemical purity, to determine the quality of the synthesized [ 68 Ga]Ga-DOTA-FAPI. [  Ga-DOTA-FAPI. All participants were observed for 30 min after the examination to monitor for any hypersensitivity reactions or adverse events. The PET/CT images were reconstructed with an iterative algorithm, with measured CT for attenuation correction.
Cholangitis and pancreatitis could cause intense uptake of [ 68 Ga]Ga-DOTA-FAPI. To evaluate whether the delayed scan could help distinguish cancerous from inflammatory lesions, the participants whose imaging findings were

PET/CT image evaluation
The datasets were analyzed visually by interpreting coronal, sagittal, and transverse slices. [ 68 Ga]Ga-DOTA-FAPI and [ 18 F]FDG PET/CT images were independently analyzed and interpreted by two experienced nuclear medicine physicians (with more than ten years of experience) in random order and at different time points. During reading, the two readers were blinded to clinical data (CT, MRI, and pathological results) and [ 68 Ga]Ga-DOTA-FAPI/[ 18 F]FDG PET scan but were aware of the primary lesion site. Any differences in opinion were discussed to reach a consensus. The uptake of [ 68 Ga]Ga-DOTA-FAPI or [ 18 F]FDG in the lesions was considered positive if these tracers were higher than the normal hepatic lobe, which was selected as background. Volumes of interest (VOI) were drawn around the positive lesions using a maximum standard uptake value (SUV max ) threshold of isocontour of 40% to perform semiquantitative analysis. Afterward, the SUV max , mean standard uptake value (SUV mean ), and metabolic tumor volume (MTV) were recorded. The tumor-to-background ratio (TBR) was calculated by dividing the SUV max of the lesion and the background SUV mean . These indicators were used to quantify the tracer uptake in tumor lesions and assess diagnosis performance.

Immunohistochemistry analysis
The sections of formalin-fixed and paraffin-embedded (FFPE) biopsy and surgical specimens were deparaffinized. Afterward, a 0.5-μm-thick FFPE tissue section was cut and used for immunohistochemistry. Semiquantitative analysis of FAP expression was rated following the scoring system in the supplementary.

Statistical analysis
This study used descriptive analyses to analyze clinical data and the participants' general characteristics. Categorical data were reported in percentages, and continuous data were presented as mean ± SD or the median (interquartile range, IQR). The Wilcoxon signed-rank test was used to compare the uptake rate of [ 18 F]FDG and [ 68 Ga]Ga-DOTA-FAPI, and the McNemar test was used to compare the diagnostic efficacy. Pearson testing and Spearman testing were used to compare the correlation between continuous and ordinal data. A p value < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS Statistics Version 22 (IBM, Armonk, NY, USA).  Table 1.

[ 68 Ga]Ga-DOTA-FAPI PET/CT patterns of BTC and biodistribution analysis
The increased uptake of [ 68 Ga]Ga-DOTA-FAPI produced clear images and correctly demonstrated the lesion area in the 42 participants with BTC. This study noted that the liver presented minor uptake of [ 68 Ga]Ga-DOTA-FAPI, which contributed to judging intrahepatic metastasis. Due to tumor obstruction of the bile and pancreatic duct, consequent cholangitis and pancreatitis are typical representations in BTC, especially in extrahepatic cholangiocarcinoma. We noticed significant [ 68 Ga]Ga-DOTA-FAPI uptake around dilated bile and pancreas. Representative [ 68 Ga]Ga-DOTA-FAPI PET/CT images of the three types of BTC are presented in Fig. 2. We analyzed the tracer biodistribution in the abdominal organs that interfere with the BTC diagnosis to assess imaging resolution quality and standardization. In the semiquantitative parameter analysis, the uptake of [ 68 Ga]Ga-DOTA-FAPI in the lesion area was more than [ 18 Fig. 4. The two tracers had no significant difference in diagnosing primary gallbladder carcinoma.
The study also found 33 participants had 244 suspected lymph nodes. Of those, 201 lymph nodes were identified as metastatic. And ultimately, 25 participants had lymph node metastasis. The rest 8 participants had no lymph node metastasis. Data were presented in tabular form (   Fig. 7, [ 68 Ga]Ga-DOTA-FAPI PET/CT detected one lesion, which was missed on CT and [ 18 F]FDG PET/CT. The surgery excised the additional foci than originally planned. The lesion was confirmed as benign foci by pathological tests. Figure 8 shows that [ 68 Ga]Ga-DOTA-FAPI PET/CT changed 17 participants' clinical management compared to CT or MRI. Six participants (6/17, 35.29%) experienced a major change (e.g., change in therapeutic regimen) to the treatment, and three (3/17, 17.64%) had minor changes (e.g., modification of intended therapy). On the other hand, eight participants (8/17, 47.06%) had no treatment regimen changes. The most frequent change was the transition to chemotherapy and immunotherapy in four participants (4/9, 44.44%), supportive care in four (4/9, 44.44%), and surgical protocol change in one (1/9, 11.11%).

Evaluation of [ 68 Ga]Ga-DOTA-FAPI in benign biliary diseases
Five participants highly suspected BTC were PET/ CT positive for benign, as confirmed by pathology.
The pathological results were chronic suppurative inflammation, fibroblastic proliferation, and gallbladder adenomyomatosis.

Association between [ 68 Ga]Ga-DOTA-FAPI PET/CT and clinical indexes
We used Pearson's correlation analysis to analyze the correlation between clinical indexes and primary lesions SUV max in [ 68 Ga]Ga-DOTA-FAPI PET/CT. The results in Fig. 10 showed that SUV max is weakly correlated with carcinoma embryonic antigen (CEA) (Pearson r = 0.364, p = 0.012) and platelet (PLT) (Pearson r = 0.35, p = 0.016). Of those, the normal CEA level was less than 7.2 ng/ml. If the level of CEA is over 14.4 ng/ml, warn the probability of associated malignancy. SUV max in the CEA ≥ 14.4 ng/ml group was statistically higher than in other groups (22. 14.57 ± 3.42 for 7.2-14.4 ng/ml group, p = 0.034). SUV max did not appear to be a difference among various PLT levels.
This study confirmed the moderate association between [ 68 Ga]Ga-DOTA-FAPI uptake intensity and FAP expression. Influenced by the participants' distribution and most tissues obtained through puncture biopsy, the correlation coefficient may be lower than the theoretical prediction. Raised CA199 could predict metastases in BTC patients and may help predict tumor burden and recurrence. The study found that [ 68 Ga] Ga-DOTA-FAPI PET/CT MTV was correlated with CA199 level but not [ 18 F]FDG. This is one of the advantages of [ 68 Ga] Ga-DOTA-FAPI. In addition, we reported weak correlation analyses between [ 68 Ga]Ga-DOTA-FAPI SUV max and CEA and PLT. High serum CEA was associated with an increased risk of presenting with tumor metastases. And the previous study found that [ 68 Ga]Ga-DOTA-FAPI SUV max correlated to tumor stage [25]. The reason for the weak correlation may be that CEA is not a specific marker in BTC. PLT acted as active players in all steps of tumorigenesis. Experimental data suggest that lowering PLT count may reduce tumor growth and metastasis [27,28]. Platelets produce platelet-derived growth factor (PDGF), which was considered a vital driver of the desmoplastic reaction. A study by Cadamuro et al. confirmed that PDGF is overexpressed in cholangiocarcinoma cells and released into the tumor cell environment to bind and activate the homologous receptor PDGFR-β on fibroblasts. This interaction provides a proliferative stimulus and elicits a strong migratory response, leading to fibroblast recruitment by cholangiocarcinoma [29,30]. However, the expression of homologous receptor PDGFR-β on fibroblasts had heterogeneity. That may explain the weak correlation between [ 68 Ga]Ga-DOTA-FAPI SUV max and PLT.
Our study had strength over previous prospective clinical trials evaluating [ 68 Ga]Ga-DOTA-FAPI PET/CT. We assessed the diagnosis performance of [ 68 Ga]Ga-DOTA-FAPI PET/CT in BTC and confirmed the correlation between [ 68 Ga]Ga-DOTA-FAPI PET/CT and clinical indexes by comparing [ 18 F]FDG PET/CT. On the other hand, this study had some limitations. First, the sample size used in this study was relatively small and heterogeneous. Second, enrolled participants were at a relatively late stage, which may have caused patient selection bias. Not all lesions had a pathological gold standard diagnosis, increasing

Conclusion
Our results exhibited that [ 68 Ga]Ga-DOTA-FAPI had a higher uptake and sensitivity than [ 18 F]FDG in the diagnosis of BTC primary and metastatic lesions and confirmed the association between [ 68 Ga]Ga-DOTA-FAPI PET/CT indexes and FAP expression, CEA, PLT, and CA199. The diagnostic and therapeutic application value of [ 68 Ga]Ga-DOTA-FAPI should be explored in future studies to improve malignancy management status.