Expression of FAP in Lung Cancer and Its Correlation With the Tumor Glucose Metabolism and the Histopathology

Purpose To explore the expression of broblast activation protein (FAP) in lung cancer (LC) and its correlation with tumor glucose metabolism and histopathology. Methods From June 2018 to November 2020, a total of 73 patients with newly diagnosed LC were included. Immunohistochemical staining was used to quantify the FAP expression in the tumors. The histopathological type and tumor grade were determined on the histopathological examination. Tumor glucose metabolic parameters and tumor maximal diameter were measured on [ 18 F] F-FDG PET/CT. Univariate and multivariate analysis were used to study the correlation between FAP expression level and glucose metabolic variables and various histopathology. Results Positive FAP expression was observed in 97.3% (71/73) of LC lesions, which was signicantly higher than 87.7% (64/73) of [ 18 F] F-FDG positivity on PET/CT (χ 2 =4.818, P=0.028). In 12 early adenocarcinomas (ADCs), only 3 lesions (25%) were positive on PET/CT, however, 10 lesions (83.3%) were positive with FAP expression. When FAP expression was classied to low level (scores ≤ 3) and high level (scores>4), high FAP level was found in 80.8% tumors and low FAP level in other 19.2% tumors. High FAP level was identied in 100.0% of squamous cell carcinoma (SCC), 85.7% of ADCs, 66.7% (4/6) of large cell neuroendocrine carcinoma (LCNC) and 40.0% (4/10) of small cell lung cancer (SCLC)(P<0.05). In the non-mucinous ADC lesions, on univariate analysis, FAP expression level showed close relationships with tumor metabolism parameters (SUVmax, SUVmean, and TLG), tumor diameter, tumor grade and lesion attenuation (P<0.05). Conclusion The present study demonstrates that FAP expresses widely in LC and has a great variant level in different histopathological types. High positive rate of FAP expression implies FAP targeted imaging may be a sensitive modality for diagnosing LC, especially for early ADCs, and may serve as an alternative of [ 18 F] F-FDG PET/CT. attenuation and tumor pathologic grade in the patients with non- mucinous ADCs (a) FAP expression level had a close relationship with tumor maximum diameter(P 0.028); expression level had a close relationship with a close close


Introduction
Lung Cancer (LC) is the leading cause of cancer-related deaths in the world [1] . Despite recent developments in the diagnosis, classi cation and therapy, the overall survival is still poor [2] .
Unfortunately, most patients present with advanced stage LC at diagnosis, mainly because of lack of disease-speci c symptoms [3] . More and more studies have shown that the occurrence and progression of tumors are the result of dynamic interactions between tumor cells and various components in the tumor microenvironment (TME) [4] . Therefore, TME emerges to be a promising target for diagnosis and therapy for LC, which is hoped to play a role in tumor early detection and improving the current unfavorable treatment effect.
Fibroblast activation protein (FAP), a type membrane-bound glycoprotein belonging to the dipeptidyl peptidase 4 family, is an imperative factor in cancer-associated broblasts (CAFs), which is a fundamental component of the TME [5] . This protein is observed to be usually highly upregulated during tissue remodeling events, including tumorigenesis [6,7] and is relevant to tumor growth, migration, and progression via the secretion of cytokines and chemokines [8,9] . Currently, FAP-targeting radioligands have been developed for PET/CT imaging and targeted radionuclide therapy for a variety of cancer including LC, colorectal cancer (CRC) and so on [10,11] , which may produce a signi cant impact on the diagnosis and treatment of LC in the future.
It is well known that LC contains a group of heterogeneous tumors, with several histopathological types including adenocarcinoma (ADC), squamous cell carcinoma (SCC), small cell lung cancer (SCLC), large cell neuroendocrine carcinoma(LCNC)and so on [1] .This disease shows great variations in biological behavior in different histopathological types and different stages [2] .It is postulated that FAP, as a fundamental factor of the TME, may have some relationship to the histopathology and the disease stages in LC.
[ 18 F] F-FDG PET/CT is a functional imaging modality, which plays an important role in the diagnosis, staging and treatment response evaluation in variety of tumors [12,13] . [ 18 F] F-FDG PET/CT can be used to quantify the metabolism of the glucose in the cancer cells and the TME in the tumor, which is usually positively related to the aggressiveness and prognosis [14,15] . A tumor with highly aggressiveness always presents with high glucose metabolism. It is also postulated that FAP may have some relationship to tumor glucose metabolism.
In the present study, we performed an immunohistochemical analysis to measure the expression of FAP in the primary tumors of LC in 73 patients, who simultaneously received the [ 18 F] F-FDG PET/CT recently. Furthermore, we also investigated the relationships of the expression level of FAP with tumor histopathological types and other clinical indicators. This study may contribute to a better understanding of the biological role of FAP in LC and could be useful to further guide clinical translation of novel FAPtargeting PET/CT imaging and therapy.

Patient Cohort
The Institutional Review Board of Nanfang Hospital at Southern Medical University approved the present retrospective study and waived the requirement for written informed consent because of the retrospective nature of the study.
From June 2018 to November 2020, a total of 73 patients with newly diagnosed LC who underwent [ 18 F] F-FDG PET/CT at Nanfang Hospital for staging before the surgery or other anti-tumor therapy were included. The inclusion criteria were as follows: (1)  150µCi/kg) [ 18 F] F-FDG, PET/CT was performed with a 5-minute single bed position three-dimensional acquisition for uEXPLORER scanner and 5-6 bed positions three-dimensional acquisition with 2 min/bed position for Biograph mCTx scanner according to established protocols in our center [16,17] .
The acquired PET and CT images were registered and analyzed using the Syngo MI workplace (Siemens Healthcare, Germany) or Image Workstation (MedEx (Beijing) Technology Limited Corporation, China). All the PET/CT images were independently read by two certi ed nuclear medicine physicians with over ve years of experience. Both physicians were blinded to the clinical and histopathological information and the results of FAP expression. According to the attenuation of lesions, the ADC lesions were classi ed into pure ground glass nodules (pGGN), mixed ground glass nodules (mGGN) and solid nodules (SN). For the semiquantitative analysis, a region of interest (ROI) was drawn along the margin of the lesion on the axial PET image and automatically adapted to a three-dimensional volume-of-interest at a 60% isocontour, which was used to measure the SUVmax, SUVmean, MTV and TLG. TLG was calculated using the formula: TLG = SUVmean×MTV.

Histologic Diagnosis
Resected or biopsied pathologic lung specimens (hematoxylin and eosin-stained slides) were independently reviewed by two certi ed pathologists to determine the histopathological type (ADC, SCC, SCLC, LCNC and so on) and ADC subtype. The malignant grade of non-mucinous ADC subtype was partitioned based on the IASLC/ATS/ERS criteria [18] . According to the literature [19] , patients with nonmucinous ADC were classi ed into three subgroups: low grade group including adenocarcinoma in situ Immunohistochemical Analysis LC tissue samples were acquired via biopsy or surgical resection, including 11 biopsy samples and 60 surgical resection samples. Para n-embedded tissue sections were depara nized with xylene and rehydrated in a graded series of ethanol solutions. Antigen retrieval was performed by heating the slides twice in 0.01 mol/L sodium citrate buffer, pH 6.0, in a microwave oven (13 min, 850 W). Endogenous peroxidase was then blocked with 0.3% H 2 O 2 in methanol for 15 min at room temperature. The staining of FAP in tumor tissue was calculated based on the percent positivity of stained cells and the staining intensity. The percent positivity was scored as 0 (absent or <1%), 1 (1%-10%), 2 (11%-50%), or 3 (>50%) [20] . Five to ten regions of interest were manually selected and each selected region was individually calculated for percent of positive cells. The average percentage of positive cells for all regions was calculated as mean for statistical comparisons. The staining intensity was scored as 0 (no staining), 1 (weak staining), 2 (moderate staining), or 3 (strong staining). The percent positivity and staining intensity were determined in a double-blinded manner. When the score of a case did not coincide, both pathologists came to an agreement after a discussion while viewing specimens under a doubleheaded microscope. The FAP expression score based on immunostaining was calculated as percent positivity score plus staining intensity score and ranged from 0 to 6. The FAP expression level was de ned as -, 0; 1 + , 1-3; 2 + , 4-5; or 3 + , 6.

Statistical Analysis
All statistical analyses were performed using SPSS version 25.0 (IBM Corporation, Armonk, NY, USA). The classi cation performance of clinicopathological factors and metabolic parameters between high FAP expression group and low FAP expression group in the tumor were analyzed. Quantitative data that were normally distributed are expressed as mean ± standard deviation, and the independent sample t-test was used for comparison between the two groups. Quantitative data were not normally distributed are expressed as median (interquartile range), and the Mann-Whitney test was used for comparison between the two groups. Qualitative data are expressed as number and percentage (n [%]), and the chi-squared test or Fisher's exact probability method were used for comparison between two groups. Multiple logistic regression analyses were used to analyze the association between FAP expression and [ 18 F] F-FDG uptake, histopathological diagnosis, and other clinical parameters. A p-value < 0.05 indicated statistical signi cance.

Association of FAP Expression Level and Tumor Glucose and Clinicopathological Variables
When all the patients were included to analyze the association between FAP expression level and tumor glucose metabolic parameters or clinicopathological variables, univariate analysis showed only the histopathology of the tumor had a close relationship with the FAP expression level (P<0.05). Neither the semiquantitative parameters of glucose metabolism, including SUVmax, SUVmean, MTV and TLG, nor the tumor maximal diameter, stages, age, gender and smoking history, had close relationship with the FAP expression level (P>0.05) ( Table 2).
However, when only non-mucinous ADCs were included to analyze the association between FAP expression level and various variables, univariate analysis showed that SUVmax, SUVmean and TLG were signi cantly related to FAP level (P<0.01), but not MTV(P>0.05) ( Figure 5). The tumors with high FAP level had signi cantly higher levels of SUVmax, SUVmean and TLG than those of the tumor with low FAP level (P<0.01) ( Table 3). Meanwhile, the tumor maximal diameter, tumor grade and tumor attenuation were also found to be closely associated with FAP level in the tumor (P<0.05) ( Figure 5). The tumors with high FAP level tended to have signi cantly larger tumor size, more solid attenuation on CT and higher grade on histopathology than those of the tumors with low FAP level (P<0.05) ( Table 3).Nevertheless, FAP level showed no signi cant relationship with the age, sex, smoking history and TNM stage(P>0.05) ( Table  3).Multivariate analysis was then used to analyze the association between FAP level and the above six related factors on the univariate analysis (three glucose metabolic parameters, tumor maximal diameter, tumor grade and lesion attenuation). However, it revealed that none of above 6factors had signi cant association with the expression level of FAP in the non-mucinous ADC tumors.

Discussion
It has long been recognized that cancer-associated broblasts (CAFs) play an important role in tumor growth-promoting, rendering cells to invade surrounding tissues through cell-to-cell interactions and secretion of different invasive molecules, as well as producing angiogenic factors for the formation of new blood vessels, all of which contribute to disease progression [21] . FAP, as an important surface marker of CAFs, has attracted more and more attention, especially after the radiopharmaceuticals targeted to FAP were recently successfully developed for tumors imaging and showed a promising prospect for treating malignant tumors [10,22,23] . LC may be one of the optimum tumors for FAP-targeted imaging and therapy because it was found to be one of 6 cancers presented with the highest uptake level (SUVmax >12) of [ 68 Ga] Ga-FAPI-04, a novel FAP targeting PET tracer [10,11] . However, so far, FAPI PET/CT imaging has mainly been performed in patients with advanced LC [11,24,25] . It will be of great signi cance to uncover whether this novel PET tracer is sensitive in diagnosing LC and which histopathological type of LC can be more bene cial from the FAPI targeted therapy. For a targeted imaging or therapy, the diagnostic and therapeutic effects largely depend on the expression of the target. Therefore, a full understanding of the expression of FAP in LC helps to guide the clinical translation of FAP-targeted imaging and therapy for this disease. Nevertheless, few papers have described the whole picture of the expression of FAP in LC.
In the present study, clinical samples from the patients with LC were analyzed for the expression of FAP using immunohistochemistry staining. It was observed in the present study that nearly all (97.3%) of the LCs had positive expression of FAP, which was signi cantly higher than 87.7% positive rate of [ 18 F] F-FDG PET/CT for LC, indicating that FAP-targeted PET/CT may be a more sensitive modality for diagnosing LC.
[ 18 F] F-FDG PET/CT is a standard imaging modality recommended for diagnosing and staging LC by NCCN guidance [26,27] . However, it is not sensitive for the detection of the low invasive tumors, such as those presented as pGGO or mGGO on CT, which usually are non-FDG avid [28] . In the present study, only 3 lesions of 12 early ADCs, which presented as pGGO and mGGO lesions on CT, were observed to be [ Ga-FAPI-04 was highly taken by an early LC, presented as mGGO on CT, while no uptake of [ 18 F] F-FDG was observed in the lesion [29] .
It was also found for the rst time that there was a great variation of FAP expression level in LCs with different histopathological types. SCC was found to have the most lesions with high FAP level, followed by ADC, LCNC and SCLC. This nding may also be useful for guiding the radionuclide ( 177 Lu or 90 Y) labeled FAP-targeted therapy. It is well known, despite the introduction of multiple new therapy for LC, the survival rate remains relatively low [2] , so it is necessary to nd more effective treatment for LC. Our data indicated that radionuclide labeled FAP-targeted therapy may be an expectable treatment for SCCs and ADCs, which showed high FAP expression level in 100% and 85.7% of the tumors. On the contrary, low FAP expression was found in most of the lesions of SCLC, implying radionuclide labeled FAP-targeted therapy may be not suitable for SCLC. Our result was consistent with that reported by Wei Yuchun et al, which showed lower uptake of [ 68 Ga] Ga-FAPI-04 in SCLCs than other types of LC [30] .
In the ADC tumors, on univariate analysis, our study revealed the FAP level had close relationships with the tumor grade, lesion diameter, lesion attenuation and glucose metabolism (P<0.05). High FAP level in the tumors were often accompanied by larger tumor lesions, more solid attenuation and higher glucose metabolism in ADCs, which were reported to be positively related with high tumor aggressiveness [14,19] . Therefore, it implies that FAPI imaging may be helpful to predict tumor invasiveness. Similar result was also reported by P. Moreno-Ruiz et al [31] . However, this relationship was not identi ed on the multivariate analysis, so further research with larger sample size is needed to determine whether this correlation really exists.
There are some limitations to this study. First, this study was retrospective study, there may have been a certain degree of bias. Second, the included sample sizes of patients were small, especially the number of LCNC and SCLC, which may fail to re ect the whole picture of the expression of FAP in LCNC and SCLC.
Lastly, no benign disease was included, which make the diagnostic speci city of FAP expression for lung cancer could not be determined.

Conclusion
The

Data availability
All data were transparent. The data used in the current study are available from the corresponding author on reasonable request.
Compliance with Ethical Standards

Con ict of interest
The authors declare that they have no con ict of interest.

Ethics approval
Ethical approval was waived by the local Ethics Committee of Southern Medical University in view of the retrospective nature of the study.

Consent to participate
Due to the retrospective nature of the study, informed consent was not obtained in the study.