Visual Abstract
Abstract
Pancreatic intraductal papillary mucinous neoplasms (IPMNs) are grossly visible (typically > 5 mm) intraductal epithelial neoplasms of mucin-producing cells, arising in the main pancreatic duct or its branches. According to the current 2-tiered grading scheme, these lesions are categorized as having either low-grade (LG) dysplasia, which has a benign prognosis, or high-grade (HG) dysplasia, which formally represents a carcinoma in situ and thus can transform to pancreatic ductal adenocarcinoma (PDAC). Because both entities require different treatments according to their risk of becoming malignant, a precise pretherapeutic diagnostic differentiation is inevitable for adequate patient management. Recently, our group has demonstrated that 68Ga-fibroblast activation protein (FAP) inhibitor (FAPI) PET/CT shows great potential for the differentiation of LG IPMNs, HG IPMNs, and PDAC according to marked differences in signal intensity and tracer dynamics. The purpose of this study was to biologically validate FAP as a target for PET imaging by analyzing immunohistochemical FAP expression in LG IPMNs, HG IPMNs, and PDAC and comparing with SUV and time to peak (TTP) measured in our prior study. Methods: To evaluate the correlation of the expression level of FAP and α-smooth muscle actin (αSMA) in neoplasm-associated stroma depending on the degree of dysplasia in IPMNs, 98 patients with a diagnosis of LG IPMN, HG IPMN, PDAC with associated HG IPMN, or PDAC who underwent pancreatic surgery at the University Hospital Heidelberg between 2017 and 2023 were identified using the database of the Institute of Pathology, University Hospital Heidelberg. In a reevaluation of hematoxylin- and eosin-stained tissue sections of formalin-fixed and paraffin-embedded resection material from the archive, which was originally generated for histopathologic routine diagnostics, a regrading of IPMNs was performed by a pathologist according to the current 2-tiered grading scheme, consequently eliminating the former diagnosis of “IPMN with intermediate-grade dysplasia.” For each case, semithin tissue sections of 3 paraffin blocks containing neoplasm were immunohistologically stained with antibodies directed against FAP and αSMA. In a masked approach, a semiquantitative analysis of the immunohistochemically stained slides was finally performed by a pathologist by adapting the immunoreactive score (IRS) and human epidermal growth factor receptor 2 (Her2)/neu score to determine the intensity and percentage of FAP- and αSMA-positive cells. Afterward, the IRS of 14 patients who underwent 68Ga-FAPI-74 PET/CT in our previous study was compared with their SUVmax, SUVmean, and TTP for result validation. Results: From 98 patients, 294 specimens (3 replicates per patient) were immunohistochemically stained for FAP and αSMA. Twenty-three patients had LG IPMNs, 11 had HG IPMNs, 10 had HG IPMNs plus PDAC, and 54 had PDAC. The tumor stroma was in all cases variably positive for FAP. The staining intensity, percentage of FAP-positive stroma, IRS, and Her2/neu score increased with higher malignancy. αSMA expression could be shown in normal pancreatic stroma as well as within peri- and intraneoplastic desmoplastic reaction. No homogeneous increase in intensity, percentage, IRS, and Her2/neu score with higher malignancy was observed for αSMA. The comparison of the mean IRS of FAP with the mean SUVmax, SUVmean, and TTP of 68Ga-GAPI-74 PET/CT showed a matching value increasing with higher malignancy in 68Ga-FAPI-74 PET imaging and immunohistochemical FAP expression. Conclusion: The immunohistochemical staining of IPMNs and PDAC validates FAP as a biology-based stromal target for in vivo imaging. Increasing expression of FAP in lesions with a higher degree of malignancy matches the expectation of a stronger FAP expression in PDAC and HG IPMNs than in LG IPMNs and corroborates our previous findings of higher SUVs and a longer TTP in PDAC and HG IPMNs than in LG IPMNs.
Pancreatic intraductal papillary mucinous neoplasms (IPMNs) are a precursor lesion of the highly lethal pancreatic ductal adenocarcinoma (PDAC), which has a 5-y survival rate of less than 10% (1,2). Pathologically, IPMNs can be subdivided into lesions with either low-grade (LG) or high-grade (HG) epithelial dysplasia (3). In clinical routine, IPMNs are usually evaluated via MRI or endoscopic ultrasound using the European guidelines for pancreatic cystic neoplasms or the Fukuoka consensus criteria (2017), which define worrisome features to differentiate between LG and HG IPMNs. Whereas LG IPMNs that lack worrisome features need only be followed up, HG IPMNs have a high risk of transforming into PDAC and thus should undergo surgery (4,5). Although MRI and endoscopic ultrasound have a high sensitivity, they lack high specificity, making them good for detecting IPMNs but not for determining their entity (6). In our prior study, we showed that PET using 68Ga-labeled fibroblast activation protein (FAP) inhibitors (FAPIs) combined with CT has the potential to differentiate between LG IPMNs and HG IPMNs. In particular, PET imaging–derived parameters (such as time to peak [TTP]) showed high ensitivity and specificity for differentiation between IPMN subtypes (7).
PDAC and other epithelial tumors are surrounded by a vast tumor stroma expressing fibroblast activation protein (FAP) and α-smooth muscle actin (αSMA), which act as markers for activated, cancer-associated fibroblasts (CAFs) (8,9). CAFs play an important role in tumor progression by promoting tumor growth, invasion, metastasis and therapy resistance (10). FAP belongs to the serine protease family and is involved in the control of fibroblast growth, tissue repair, and epithelial carcinogenesis (11,12). Thus, a neoplastic transformation process is usually associated with an increase in stromal FAP expression whereas FAP is nearly undetectable in healthy tissue and benign lesions (8). αSMA, being involved in cell motility, is usually expressed in pericytes, myoepithelial cells, smooth muscle cells, and myofibroblasts and is consequently found in normal, reactive, and neoplastic tissue (13). Especially, αSMA is an established marker for CAFs in various cancers, among them PDAC (14).
In previous evaluations of 68Ga-FAPI PET/CT, we observed lower 68Ga-FAPI uptake in precursor lesions (e.g., lung fibrosis and pancreatitis) than in carcinoma (15,16). These results were in line with our IPMN findings showing markedly lower FAPI-74 uptake in LG than HG IPMNs, as well as a shorter TTP in dynamic PET imaging (7). This led to the hypothesis that FAP expression and thus the FAPI avidity of pancreatic lesions increase with higher malignancy. To elucidate differences in FAP expression between LG and HG IPMNs that may underlie their differential appearance on 68Ga-FAPI74 PET/CT and to validate FAP as a biology-based target for PET imaging, we performed immunohistochemistry on a larger series of LG IPMN, HG IPMN, and PDAC specimens.
MATERIALS AND METHODS
All procedures performed in studies involving human participants conformed with the ethical standards of the institutional or national research committee and with the Helsinki declaration (1964) and its later amendments or comparable ethical standards. This retrospective study was approved by the local advisory ethic committee (study S-115/2020).
Histomorphology-Based Classification of Specimens
Hematoxylin- and eosin-stained tissue sections of the surgical material of patients with a diagnosis of LG IPMN, HG IPMN, or PDAC who underwent surgery at the University Hospital Heidelberg between 2017 and 2021 were provided by the EPZ biobank archive (Department of Surgery of the University Hospital Heidelberg) in accordance with the regulations of the tissue bank and the approval of the ethics committee of the University Hospital Heidelberg. All tissue slides were histologically evaluated by an experienced pathologist, who chose representative areas containing IPMN or PDAC tissue and reference tissue. Three representative specimens reflecting the neoplastic growth pattern, especially regarding the ratio of neoplastic epithelial cells to stroma and the grading of neoplasm, were selected for each patient. All specimens were classified as LG IPMN, HG IPMN, HG IPMN plus PDAC, or PDAC. IPMN was defined as a proliferation of intraductal columnar cells producing mucin and was classified as LG or HG dysplasia by the highest degree of cytoarchitectural and nuclear atypia of the epithelium. LG IPMN showed monomorphic columnar cells, which might show mitosis and papillary projections as well as slight to moderate nuclear atypia. In contrast, HG IPMN showed nuclear stratification with loss of polarity, pleomorphism, severe atypia, and papillae with irregular branching and budding (3). PDAC showed ductlike glandular structures arbitrarily infiltrating the pancreatic parenchyma. PDAC was divided into 4 groups from well differentiated to poorly differentiated: grade 1 (G1), grade 2 (G2), grade 3 (G3), and grade 4 (G4). Neoplastic irregular glands surrounded by strong to loosely arranged desmoplastic stroma containing fibroblasts and myofibroblasts, as well as scattered macrophages and lymphocytes, were interpreted as characteristics common to all (17).
FAP and αSMA Immunohistochemistry
For the lesion areas chosen as described above, semithin tissue sections 4 μm thick were prepared from corresponding paraffin blocks generated from resected tissue after fixation in 4% buffered formalin for 24 h at room temperature. Tissue sections were treated with Ultra Cell Conditioning Solution (Roche) buffer (pH 8.0) for antigen retrieval. Immunohistochemical staining was performed using the following antibodies: anti-FAPa (1:100; Abcam [reference number ab207178]) and anti-αSMA (ready to use; Cell Signaling Technology [reference number 760-2833]). Automated immunostaining was done using the BenchMark Ultra automated staining platform (Roche) with the OptiView DAB immunohistochemistry detection kit (Roche), Autostainer Link 48 (Agilent), and the EnVision Flex kit (Agilent). Stained tissue sections were mounted with Consul-Mount (Thermo Fisher Scientific) and scanned by Aperio AT2 (Leica; magnification, 1:400) for analysis.
Semiquantitative Analysis
FAP and αSMA expression was assessed both in adjacent nonmalignant tissue and in IPMN or PDAC regions marked by premalignant or malignant cell clusters with surrounding stroma. A semiquantitative analysis adapted from Remmele and Stegner (18) was used to interpret the immunohistochemistry. Scores were assigned according to the percentage stain distribution, with a score of 0 indicating 0%; 1 indicating less than 10%, 2 indicating 10%–50%, 3 indicating 51%–80%, and 4 indicating more than 80%. Additionally, stain intensity was scored, with a score of 0 indicating no stain, 1 indicating a low stain intensity, 2 indicating a medium stain intensity, and 3 indicating a high stain intensity. By multiplying the percentage score (0–4 points) with the intensity score (0–3 points), the immunoreactive score (IRS) consisting of 0–12 points was obtained. These points were further classified into 3 groups, with IRS scores 0–1 becoming group 0 (negative), 2–3 becoming group 1 (positive with weak expression), 4–8 becoming group 2 (positive with mild expression), and 9–12 becoming group 3 (positive with strong expression), to match the human epidermal growth factor receptor 2 (Her2)/neu score of the American Society of Clinical Oncology/College of American Pathologists guideline for the evaluation of estrogen and progesterone receptor status in breast cancer. In this Her2/neu score, 0 means that less than 10% of cells are stained, 1+ means that more than 10% cells are minimally stained, 2+ means that more than 10% cells are moderately stained, and 3+ means that more than 10% cells are strongly stained, with a score of 0 or 1+ being classified as negative, 2+ as mildly positive, and 3+ as strongly positive (19).
68Ga-FAPI-74 PET/CT
The synthesis and radioactive labeling of FAPI-74 and the PET imaging procedures are described in a report about our recent PET-imaging project on IPMNs (7) and were previously described by others (20–22); 14 of the 98 patients included in this work had undergone 68Ga-FAPI-74 PET/CT in that project. For PET imaging, a Siemens Biograph mCT Flow scanner was used, according to previously published protocols (16). Static PET scans were acquired 60 min after injection. Dynamic PET scans were additionally performed as previously described (16). PET data were analyzed retrospectively, and static (SUVmax and SUVmean) and dynamic (TTP) PET parameters were extracted as previously described (7).
Statistical Analysis
When there were 3 or more different subgroups, significant differences between staining results or imaging parameters were determined using 1-way ANOVA with Bonferroni multiple-comparison tests, and 4 levels of significance were discriminated (P < 0.05, P < 0.01, P < 0.001, and P < 0.0001). For the subgroup analysis of 2 groups for Supplemental Figure 3 (supplemental materials are available at http://jnm.snmjournals.org), an unpaired t test was used and P values of less than 0.05 were defined as statistically significant. All statistical analyses were performed using GraphPad Prism, version 10.
RESULTS
Immunohistochemical staining against FAP and αSMA was performed on 294 specimens (3 specimens per patient) of 98 treatment-naïve patients (47 female and 51 male; average age at diagnosis, 66.47 y [range, 43–83 y] for women and 65.53 y [range, 47–84 y] for men). Twenty-three patients were diagnosed with LG IPMN, 11 with HG IPMN, and 10 with HG IPMN plus PDAC, of whom 8 had G2 PDAC and 2 had G3 PDAC. Fifty-four patients were diagnosed with PDAC, of whom 2 had G1 PDAC, 33 had G2 PDAC, and 19 had G3 PDAC. The most frequent localization of the lesions was the pancreas head (61 patients), followed by the cauda (15 patients) and corpus (13 patients). In 4 patients, the lesion was at the junction of the head and corpus, and in 2 patients, the lesion was at the junction of the corpus and cauda. In 2 patients, the lesion was in the whole pancreas, and in 1 patient, the lesion extended from the processus uncinatus to the cauda (Supplemental Table 1).
Results of Immunohistochemical Staining
FAP
Although a weak immunohistochemical reactivity against FAP could be observed in the tumor-free pancreatic stroma, significantly higher expression was detected in neoplasia-associated peri- and intratumoral desmoplastic stroma, with the tendency increasing with malignancy grade. However, a strong, unspecific immunopositivity for FAP could also be shown for epithelial tumor cells in several specimens.
LG IPMNs showed a less FAP-positive stromal reaction and a lower FAP intensity than HG IPMNs. The mean IRS was 2.09 (±2.13; median, 2.0) for LG IPMNs, 3.0 (±3.77; median, 2.0) for HG IPMNs, 7.4 (±4.6; median, 7.5) for HG IPMNs plus PDAC, and 9.67 (±3.21; median, 12) for PDAC (Fig. 1A). Mean Her2/neu scores were 0.61 (±0.72; median, 1.0) for LG IPMNs, 0.91 (±1.04; median, 1.0) for HG IPMNs, 2.10 (±1.1; median, 2.5) for HG IPMNs plus PDAC, and 2.67 (±0.7; median, 3.0) for PDAC (Fig. 1B). Similarly, both criteria—the intensity of the immunohistochemical reaction (reflected by intensity score) and the percentage of FAP-positive stroma (reflected by percentage score)—increased with a higher risk of malignant transformation (Figs. 1C and 1D).
Mean values (±SEM) of FAP IRS (A), Her2/neu score (B), intensity score (C), and percentage score (D) of LG IPMNs, HG IPMNs, HG IPMNs with PDAC, and PDAC. P values are for 1-way ANOVA with Bonferroni multiple-comparison tests. *P < 0.05. **P < 0.01. ***P < 0.001. ****P < 0.0001.
Regarding grading, the mean IRS, Her2/neu score, intensity score, and percentage score were lower for HG IPMNs plus G2 PDAC than for HG IPMNs plus G3 PDAC, with a mean IRS of 6.75 (±3.96; median, 6.0) versus 12.0 (±0.0; median, 12) and a mean Her2/neu score of 2.13 (±0.99; median, 2.0) versus 3.0 (±0.0; median, 3.0). In G1 PDAC, the mean IRS and the mean Her2/neu score were lower than for HG plus G2 PDAC and G3 PDAC, with a mean IRS of 3.5 (±3.54; median, 3.5) for G1 versus 9.82 (±3.23; median, 12.0) for G2 versus 10.05 (±2.63; median, 12.0) for G3, as well as a mean Her2/neu score of 1.0 (±1.41; median, 1.0) for G1, 2.76 (±0.56; median, 3.0) for G2, and 3.0 (±0.0; median, 3.0) for G3 (Supplemental Figs. 1A and 1B). Only slight differences were observed between the intensity score and the percentage of moderately and poorly differentiated PDAC (Supplemental Figs. 1C and 1D). Of note, in all specimens, tumor-free islets of FAP-positive Langerhans were observed.
αSMA
In this study, αSMA expression could be shown in normal pancreatic stroma and within peri- and intraneoplastic desmoplastic reaction.
The mean IRS of LG IPMNs, at 6.04 (±3.25; median, 6.0), was higher than that of HG IPMNs, at 3.91(±3.70; median, 2.0), whereas the mean IRS of HG IPMNs plus PDAC, at 8.1 (±3; median, 7.0), was slightly lower than that of PDAC, at 8.73 (±2.86; median, 8.0) (Fig. 2A). Similarly, the Her2/neu score showed means of 2.26 (±0.62; median, 2.0) for LG IPMNs, 1.64 (±0.92; median, 2.0) for HG IPMNs, 2.4 (±0.52; median, 2.0) for HG IPMNs plus PDAC, and 2.36 (±0.7; median, 2.0) for PDAC (Fig. 2B). LG IPMNs had a higher mean intensity and percentage score than HG IPMNs. Both LG and HG IPMNs had lower mean intensity and percentage scores than HG IPMNs plus PDAC and PDAC, whereas the differences between HG IPMNs plus PDAC and PDAC were minimal (Figs. 2C and 2D).
Mean values (±SEM) of αSMA IRS (A), Her2/neu score (B), intensity score (C), and percentage score (D) of LG IPMNs, HG IPMNs, HG IPMNs with PDAC, and PDAC. P values are for 1-way ANOVA with Bonferroni multiple-comparison tests. *P < 0.05. ***P < 0.001. ****P < 0.0001.
Regarding grading, HG IPMNs plus G2 PDAC, HG IPMNs plus G3 PDAC, and PDAC (G1–G3) showed the mean IRS, Her2/neu score, intensity score, and percentage score to be at a similar level, without marked differences (Supplemental Fig. 2).
Immunohistochemical results concerning FAP and αSMA expression in IPMNs plus PDAC and PDAC subdivided by grading are pictured in Supplemental Figures 3 and 4.
Comparison of FAP Expression and 68Ga-FAPI-74b Uptake
For 14 patients who underwent 68Ga-FAPI-74 PET/CT, the mean IRS, SUVmax, SUVmean, and TTP were compared. Of these, 4 had LG IPMNs, 4 had HG IPMNs, 4 had HG IPMNs plus G2 PDAC, and 2 had HG IPMNs plus G3 PDAC.
LG IPMNs had a mean IRS of 1.8 (±1.7; median, 1.5), a mean SUVmax of 3.7 (±2.1; median, 3.5), a mean SUVmean of 2.1 (±1.3; median, 1.8), and a mean TTP of 45.0 s (±0.0 s; median, 45 s). HG IPMNs had a mean IRS of 3.8 (±5.6; median, 1.5), a mean SUVmax of 5.1 (±1.6; median, 4.9), a mean SUVmean of 2.5 (±0.6; median, 2.3), and a mean TTP of 247.5 s (±173.4 s; median, 232.5 s). HG IPMNs plus PDAC had a mean IRS of 6.8 (±5.7; median, 7.0), a mean SUVmax of 9.1 (±6.0; median, 7.7), a mean SUVmean of 4.8 (±3.4; median, 3.7), and a mean TTP of 408.0 s (±147.5 s; median, 450 s) (Fig. 3). The concordant increase in FAP expression and 68Ga-FAPI-74 uptake on PET/CT is visualized in Figure 4.
Comparison of mean values (with SD) of IRS with SUVmax (A), SUVmean (B), and TTP (C) of LG IPMNs, HG IPMNs, and HG IPMNs plus PDAC as parameters of static and dynamic 68Ga-FAPI-74 PET imaging observed in our previous study.
Representative hematoxylin and eosin staining (×20), immunohistochemical staining against FAP (×20), and corresponding axial PET/CT with 68Ga-FAPI-74 images of 3 patients with LG IPMNs in pancreatic tail, HG IPMNs in pancreatic head, and HG IPMNs plus PDAC in pancreatic head (scale bars = 50 μm; arrows indicate pancreatic lesions). HE = hematoxylin and eosin; IHC = immunohistochemistry.
Regarding grading, HG IPMNs plus G2 PDAC showed a mean IRS of 4.3 (±5.2; median, 2.0), a mean SUVmax of 6.1 (±2.3; median, 5.9), a mean SUVmean of 3.0 (±1.0; median, 2.8), and a mean TTP of 360.0 s (±184.3 s; median, 285.0 s). The highest values were observed for HG IPMNs plus G3 PDAC, with a mean IRS of 12.0 (±0; median, 12.0), a mean SUVmax of 15.0 (±7.5; median, 15.0), a mean SUVmean of 8.3 (±4.4; median, 8.3), and a mean TTP of 480.0 s (±42.4 s; median, 480.0 s) (Supplemental Fig. 5). In summary, the IRS reflecting FAP expression increases with higher malignancy in accordance with the increasing mean SUVmax, SUVmean, and TTP, with higher malignancy observed in our previous study.
DISCUSSION
Summary of Results
In this study we performed immunohistochemistry with antibodies directed against FAP and α-SMA on FFPE material of LG IPMNs, HG IPMNs, HG IPMNs plus PDAC, and PDAC. The semiquantitative analysis of FAP staining showed a rising IRS with increasing malignancy, confirming our previous finding and showing concordance between the mean IRS and the mean SUVmax, SUVmean, and TTP. A higher IRS for α-SMA was observed in PDAC than in LG IPMNs, but a clear ascending order of IRS with increasing malignancy was missing.
Validation and Comparison of FAPI PET/CT-Acquired Data and FAP Expression
We observed a higher percentage of FAP-positive stroma and a stronger staining intensity in HG IPMNs than in LG IPMNs. The IRS underlines this outcome in that the mean HG IPMNs were twice as high as the IRS of LG IPMNs, thus demonstrating stronger FAP expression in HG IPMNs than in LG IPMNs. These results are similar to those of other studies, validating their FAPI PET/CT–acquired data by immunohistochemistry (15,23). Consequently, these findings are in line with our previous results showing a higher SUVmax and SUVmean and a longer TTP in HG IPMNs than in LG IPMNs (7). Besides, our results are in accord with the outcome of a further prospective study revealing a strong correlation between SUVs and the IRS of FAP in patients with various solid tumors (24). By adding specimens of HG IPMNs plus PDAC and PDAC to our study, we could demonstrate a constant increase in the IRS of FAP (and thus its stronger expression) with higher malignancy. Because FAP is a marker of CAFs, which are associated with tumor progression and poor prognosis (25), our findings suggest that 68Ga-FAPI PET/CT–acquired data project the histopathologic diagnosis by SUV and TTP and may harbor the potential to predict malignant progression.
FAP Expression in Tumor-Free Pancreatic Stroma and in Epithelial Tumor Cells
FAP, forming a homodimeric integral membrane gelatinase as a member of the serine protease family, promotes extracellular matrix degradation. Consequently, it participates not only in tumor growth but also in nonmalignant processes such as inflammation with a fibrotic component and tissue remodeling. Therefore, a weak immunohistochemical positivity against this protein was observed in tumor-free pancreatic stroma, an observation that is usually characterized by a minimal inflammatory reaction to the production of aggressive, autodigestive enzymes by exocrine ductal cells and of a periductal microleakage of this enzyme-containing secretion, resulting in necrosis of single exocrine cells. Additionally, FAP is involved in pericellular proteolysis of the extracellular matrix and hence promotes cell adhesion and migration. Thus, FAP is nonspecifically expressed in α-cells of Langerhans insulae (26). However, FAP expression has also been shown for β-cells, endothelial cells, macrophages, and ductal cells in normal pancreatic parenchyma (as indicated by a search of the Human Protein Atlas in September 2023). This observation could also explain the strong unspecific immunoreactivity of several epithelial tumor cells in pancreatic carcinomas. Especially in the context of possible protein function in the control of epithelial–mesenchymal interactions during epithelial carcinogenesis (as indicated by a search of the RefSeq database in September 2023), further investigations are required to better understand the role of FAP expression in epithelial tumor cells.
Tumor Microenvironment and PDAC Progression
A possible mechanistic explanation for the differential FAP expression of LG IPMNs, HG IPMNs, and PDAC could be the role of the tumor microenvironment and CAFs during the carcinogenesis of PDAC. At the earlier stages of PDAC development, a proinflammatory tumor microenvironment may support tumor initiation (27,28). During cancer progression, the fibrotic and immune suppressive effects of the tumor microenvironment promote invasion and induce chemoresistance (29). CAFs act as major mediators of these differential, protumorigenic tumor microenvironment functions (30). CAFs are a heterogeneous population because they can develop from different cell types, such as pancreatic stellate cells, resident fibroblasts, epithelial cells, or even fat cells (14,31). Even more importantly, CAFs are a dynamic population that changes its biologic activity and antigen signature during cancer development (27). Our findings suggest that the portion of FAP-positive CAFs or the expression level of FAP in CAFs increases during PDAC cancer genesis, possibly due to changes in the tumor microenvironment composition and its increasing profibrotic function over time. Although we observed an ascending IRS with increasing malignancy for FAP, the IRS of αSMA was higher in HG IPMNs with PDAC and PDAC than in LG and HG IPMNs but showed no clear ascending score with higher malignancy. Although αSMA is a common marker for CAFs besides FAP, it is not expressed consistently in each tissue containing CAFs. Öhlund et al. observed 2 subtypes of CAFs with distinct functions in PDAC and with different expression of αSMA. Although the inflammatory CAFs are characterized by low αSMA expression, the myofibroblastic CAFs show an elevated expression of αSMA (14). In a comparison of PDAC versus IPMNs, Bernard et al. identified inflammatory CAFs only in PDAC whereas myofibroblastic CAFs were present in PDAC as well as in both LG and HG IPMNs, with higher representation in the latter samples (32). On the one hand, our findings of a stronger αSMA positivity in PDAC than in IPMNs suit this observation of different αSMA expression in both lesions depending on the subpopulation of CAFs present. On the other hand, the inverse αSMA positivity in HG versus LG IPMNs and the equal αSMA expression of PDAC (G1–G3) lead us to the hypothesis that αSMA is not as suitable as FAP as a marker for malignant progression of IPMNs and aggressiveness of PDAC.
Clinical Implications
The increasing expression of FAP with higher malignancy was measurable concordantly as well by immunohistochemistry as by 68Ga-FAPI PET. This suggests that FAP expression has the potential to become a malignancy marker for pancreatic lesions. In a primary setting, 68Ga-FAPI PET–based or biopsy-based immunohistochemical evaluation of FAP positivity could be helpful to estimate the risk of malignancy of pancreatic lesions and support decision making regarding a possible resection or watch-and-wait strategy. For PDAC patients, immunohistochemical FAP expression of resected tumors or FAP avidity of tumor manifestations measured by 68Ga-FAPI PET may have prognostic value and support treatment decisions—in particular with respect to the role of FAP-positive CAF in chemo- and radioresistance (33–35). Finally, 68Ga-FAPI PET holds potential as a monitoring tool for systemic PDAC therapies, especially if CAF-targeted therapies are applied in the future (36). The evaluation of FAP immunohistochemistry and 68Ga-FAPI PET in these clinical settings is a promising subject for future prospective clinical trials.
Limitations
Although the results of our histopathologic target validation study are promising, several limitations must be considered. The major limitation is that the method of semiquantitative analysis does not allow exclusion of a certain interobserver deviation. A computer-automated evaluation of the staining results might be a more precise method for further studies. The second limitation is the relatively small number of specimens per diagnosis. Despite staining 3 slides per diagnosis to increase the validity of our findings, the SDs are somewhat high and could be reduced by larger patient cohorts. The unproportioned specimens’ distribution must be considered when interpreting the results. With regard to our comparison of immunohistochemistry results and PET imaging, only a small group (14/98) of patients underwent 68Ga-FAPI PET imaging. Summarizing, our results must be interpreted with caution and should be confirmed by prospective studies with larger patient cohorts.
CONCLUSION
This analysis of FAP immunohistochemistry in synopsis with the static and dynamic parameters of 68Ga-FAPI-74 PET/CT shows an ascending order of FAP positivity from LG IPMNs to PDAC, confirming increasing FAPI avidity with higher malignancy in pancreatic lesions. To allow possible predictions on the progression of the disease through 68Ga-FAPI PET/CT, further analysis of immunohistochemistry and PET parameters in larger prospective studies are needed.
DISCLOSURE
This work was funded by grant 13N 13341 from the Federal Ministry of Education and Research. Uwe Haberkorn has filed a patent application for quinoline-based FAP-targeting agents for imaging and therapy in nuclear medicine and has shares of a consultancy group for iTheranostics. No other potential conflict of interest relevant to this article was reported.
KEY POINTS
QUESTION: Do FAP expression and thus FAPI avidity in pancreatic lesions increase with higher malignancy?
PERTINENT FINDINGS: FAP positivity in LG IPMNs, HG IPMNs, PDAC deriving from HG IPMNs, and PDAC increases with rising malignancy and is in concordance with 68Ga-FAPI PET/CT parameters.
IMPLICATIONS FOR PATIENT CARE: 68Ga-FAPI PET–acquired data reflect the FAP expression of pathologic subclasses of IPMNs and might predict further malign transformation into PDAC. Further analyses of IRS and SUV or TTP in prospective studies with larger cohorts are needed.
ACKNOWLEDGMENTS
We thank the EPZ biobank archive in Heidelberg for providing the slides and Tim Rau (NCT tissue bank in Heidelberg) for excellent technical assistance with staining and immunohistochemistry.
Footnotes
↵* Contributed equally to this work.
- © 2024 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- Received for publication July 18, 2023.
- Revision received October 17, 2023.
