Visual Abstract
Abstract
The fibroblast activation protein (FAP) is highly expressed in tumor and stromal cells of mesothelioma and thus is an interesting imaging and therapeutic target. Previous data on PET imaging with radiolabeled FAP inhibitors (FAPIs) suggest high potential for superior tumor detection. Here, we report the data of a large malignant pleural mesothelioma cohort within a 68Ga-FAPI46 PET observational trial (NCT04571086). Methods: Of 43 eligible patients with suspected or proven malignant mesothelioma, 41 could be included in the data analysis of the 68Ga-FAPI46 PET observational trial. All patients underwent 68Ga-FAPI46 PET/CT, contrast-enhanced CT, and 18F-FDG PET/CT. The primary study endpoint was the association of 68Ga-FAPI46 PET uptake intensity and histopathologic FAP expression. Furthermore, secondary endpoints were detection rate and sensitivity, specificity, and positive and negative predictive values as compared with 18F-FDG PET/CT. Datasets were interpreted by 2 masked readers. Results: The primary endpoint was met, and the association between 68Ga-FAPI46 SUVmax or SUVpeak and histopathologic FAP expression was significant (SUVmax: r = 0.49, P = 0.037; SUVpeak: r = 0.51, P = 0.030).68Ga-FAPI46 and 18F-FDG showed similar sensitivity by histopathologic validation on a per-patient (100.0% vs. 97.3%) and per region (98.0% vs. 95.9%) basis. Per-region analysis revealed higher 68Ga-FAPI46 than 18F-FDG specificity (81.1% vs. 36.8%) and positive predictive value (87.5% vs. 66.2%). Conclusion: We confirm an association of 68Ga-FAPI46 uptake and histopathologic FAP expression in mesothelioma patients. Additionally, we report high sensitivity and superior specificity and positive predictive value for 68Ga-FAPI46 versus 18F-FDG.
Malignant mesothelioma is a relatively rare solid malignancy of soft tissue, most frequently affecting the pleura, and is associated with poor survival (1). Because of late detection, mesotheliomas are often diagnosed in unresectable, advanced stages, emphasizing the need for accurate imaging methods to identify suspected lesions (2). In recent years, the fibroblast activation protein (FAP) has become an interesting target for novel molecular probes because of its high expression in tumor cells of mesenchymal origin and in carcinoma-associated fibroblasts in stromal tissue of various solid tumors (3,4). Carcinoma-associated fibroblasts influence tumor cells by producing mediators and can promote tumor angiogenesis, migration, and proliferation (5), and FAP overexpression in solid tumors has been linked with poor outcome and is distinctively overexpressed in tumor tissue compared with normal tissue (4). Especially, mesothelioma has been shown to highly express FAP on tumor tissue in all histopathologic subtypes (3). Therefore, FAP-targeted therapies are of interest for future treatments, and experimental and phase I clinical trials for anti-FAP CAR T-cell therapy have already been evaluated (6,7). In terms of diagnostic approaches, imaging is still focused on radiologic examinations such as contrast-enhanced CT (Ce-CT), MRI, and 18F-FDG PET, but novel FAP-targeted radiotracers have shown promising tumor uptake in mesothelioma patients (8–10).
In 2018, novel FAP-targeted radiotracers were introduced for diagnostic and therapeutic purposes and showed promising diagnostic value for multiple entities, including mesenchymal cancers such as malignant pleural mesothelioma, but have not been investigated in depth (11,12).
We initiated a prospective, single-center, observational FAP inhibitor (FAPI) PET trial to investigate the association of histopathologic FAP expression and 68Ga-FAPI46 PET uptake intensity in various tumor entities (13). We further aimed to analyze the 68Ga-FAPI46 PET/CT sensitivity, specificity, positive and negative predictive values, and detection rates of this new methodology in comparison to established imaging modalities (18F-FDG PET/CT and Ce-CT).
MATERIALS AND METHODS
Study Design and Patients
This was a subgroup analysis of an ongoing 68Ga-FAPI46 PET observational trial (NCT04571086), and the detailed study protocol has been published before (14,15).
Briefly, adult patients scheduled for 68Ga-FAPI46 PET/CT for staging or restaging of proven or suspected mesothelioma as part of clinical workup or routine were enrolled and underwent follow-up prospectively for at least 6 mo after enrollment. In that period, available histopathology specimens, images (CT, MRI, PET/CT, etc.), change-of-management questionnaires, and other clinical data were collected and then analyzed. Correlation of 68Ga-FAPI46 uptake intensity and histopathologic FAP expression was defined as the primary endpoint (Spearman correlation for ordinal data). Secondary endpoints were detection rate and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy confirmed by histopathology per patient and per region or by a composite reference standard (pathology and imaging-based lesion follow-up).
The study was initiated, planned, conducted, analyzed, and published by the Department of Nuclear Medicine of University Hospital Essen. No financial support was received from commercial entities. All reported investigations were conducted in accordance with the Declaration of Helsinki and with national regulations. This observational trial was registered on clinicaltrails.gov (NCT04571086) and approved by the local Ethics Committee (permits 19-8991-BO and 20-9485-BO). All patients provided their informed consent.
Figure 1 illustrates the patient flow and the inclusion and exclusion of patients and respective cohorts for the respective endpoints (Fig. 1).
Enrollment flowchart. IHC = immunohistochemistry.
Imaging and Image Analysis
PET scans were performed in the craniocaudal direction on a Vereos (Philips), Biograph mCT (Siemens Healthineers), or Biograph mCT VISION (Siemens Healthineers) device. All 68Ga-FAPI46 PET scans were performed as PET/CT (including low-dose CT). The mean injected activity of 68Ga-FAPI46 was 124 ± 32 MBq, and that of 18F-FDG was 260 ± 71 MBq. 68Ga-FAPI46 PET/CT images were acquired approximately 10–60 min (mean, 24 ± 19 min) after injection, and 18F-FDG PET/CT images were acquired approximately 60 min (mean 74 ± 19 min) after injection. Diagnostic Ce-CT (standard protocol: 80–100 mA, 120 kV with iodinated contrast medium), and 18F-FDG PET/CT images were available for all enrolled patients and were included in the image analysis if performed within 2 wk of 68Ga-FAPI46 PET/CT and no relevant therapeutic changes occurred in that time frame (e.g., surgery, chemotherapy, or radiation).
For image analysis, tumor regions were defined as local (pleura), locoregional lymph nodes, and metastatic lesions, as previously published (14). For each scan, the number of lesions per region and per patient and the size and uptake of the lesion with the highest uptake were recorded. Any focal uptake higher than the surrounding background and not physiologic was considered suggestive of malignancy. SUVs (i.e., SUVmax and SUVpeak) were measured with a region-growing algorithm with a threshold of 40% of the maximal uptake (Syngo.via software; Siemens Healthcare) for the lesion with the highest uptake in the respective tumor region. Images were reported independently by 2 masked expert readers (experience with >500 68Ga-FAPI46 PET/CT scans, >1,000 18F-FDG PET/CT scans, and >2,000 Ce-CT scans). Readers were aware of previous surgeries. Divergent findings were discussed and reported in a separate consensus session between readers.
Lesion Detection and Validation
The detection rate was defined as the number and proportion of patients with PET-positive results overall, per region, and per patient, independently of the reference standard or validation (16). On CT imaging, lesions were defined according to RECIST 1.1 and mRECIST 1.1 for pleural mesothelioma (e.g., organ lesions > 10 mm; lymph nodes > 15 mm in the short-axis dimension).
Patients were subjected to histopathologic examination and, when feasible, FAP immunohistochemistry for subsequent evaluation. Lesions and tumor regions were validated either by available histopathology from surgeries or biopsies (e.g., endobronchial ultrasound–guided transbronchial fine-needle aspiration) within a 4-wk period of 68Ga-FAPI46 PET or a composite reference standard that combined histopathology and available imaging follow-up (e.g., time point response, progression, or cross validation). Validation was performed by local investigators, adhering to the study protocol criteria and relying on images and reports. Lesions were validated for 68Ga-FAPI46 PET/CT and 18F-FDG PET/CT separately.
Immunohistochemistry
Available histopathologic specimen were stained with standard hematoxylin and eosin and for FAP immunohistochemistry and evaluated as previously described (14,17). In short, immunohistochemistry samples were visually and semiquantitatively scored as 0 (negative), 1 (1%–10% FAP-positive cells), 2 (11%–50% FAP-positive cells), or 3 (>50% FAP-positive cells).
Statistical Analysis
PET uptake and tissue FAP expression were compared by Pearson correlation for continuous data and Spearman correlation for ordinal data. In addition, uptake and expression data were compared descriptively for each score, uptake, or expression range. Sensitivity, specificity, PPV, NPV, and accuracy on a per-patient and per-region basis for 68Ga-FAPI46 PET and 18F-FDG PET detection of tumor location as confirmed by histopathology, biopsy, or the composite reference standard were calculated and reported along with the corresponding 2-sided 95% CIs. To compare the frequency of detected lesions, contingency testing using the Fisher exact test or χ2 test was used. The CIs were constructed using the Wilson score method. Uptake measurements of tumor regions were tested for statistical differences using the nonparametric Mann–Whitney U test. All statistical analyses were performed using SSPS software (version 28.0; IBM) or Prism (version 9.1.1; GraphPad Software).
RESULTS
Patient Characteristics
From April 2020 till August 2023, 43 patients with suspected or proven malignant mesothelioma were enrolled in the 68Ga-FAPI46 PET observational trial, of whom 41 were included in the analysis (2 were excluded because of additional cancer disease or missing clinical data) (Fig. 1). For all patients, 18F-FDG PET/CT and Ce-CT were available within 14 d of 68Ga-FAPI46 PET/CT. In no patient was a therapeutic change made in the interval between 68Ga-FAPI46, 18F-FDG, and Ce-CT. Table 1 details the clinical characteristics of the included patients. No adverse events were reported.
Patient Characteristics (n = 41)
Primary Endpoint: Correlation of 68Ga-FAPI46 PET Uptake with FAP Expression
Histopathologic specimens of 22 patients were available for FAP immunohistopathology, of which 18 were of adequate quality for histopathologic evaluation and scoring. No negative immunohistochemistry was reported in this cohort (score 1, n = 3; score 2, n = 9; and score 3, n = 6); representative immunohistochemistry images are depicted in Figure 2A. 68Ga-FAPI46 SUVmax and SUVpeak showed a moderate, significant linear relationship with the established immunohistochemical scoring system (n = 18) (SUVmax: r = 0.49, P = 0.04; SUVpeak: r = 0.51, P = 0.03) (Fig. 2B). 68Ga-FAPI46 and 18F-FDG showed comparably high uptake for all tumor regions (e.g., local SUVmax: 14.6 ± 8.8 for 68Ga-FAPI46 vs. 12.4 ± 7.4 for 18F-FDG), with a significant difference only in SUVpeak for local tumor (68Ga-FAPI46, 10.9 ± 6.3; 18F-FDG SUVpeak, 8.1 ± 4.9; P = 0.02) (Supplemental Table 1; supplemental materials are available at http://jnm.snmjournals.org).
Primary endpoint. (A) DAB immunohistochemical staining of FAP-α in pleural mesothelioma specimen and respective positive scores. No negative sample was reported in this cohort. (B) Spearman correlation shows moderate association between 68Ga-FAPI46 PET uptake intensity and FAP immunohistochemistry score for available malignant pleural mesothelioma specimen (n = 18) (SUVmax: r = 0.49, P = 0.037; SUVpeak: r = 0.51, P = 0.030).
Secondary Endpoints: Diagnostic Performance
On a per-patient basis, validation by histopathology was possible in all 41 (100%) patients. Lesions were validated in 87 regions by histopathology and in 137 regions by the composite reference standard. The detailed diagnostic performance and CI on a per-patient and per-region basis are shown in Table 2. 68Ga-FAPI46 and 18F-FDG showed similar sensitivity by histopathologic validation on a per-patient (100.0% vs. 97.3%) and per-region (98.0% vs. 95.9%) basis. But per-region analysis revealed a distinct difference in diagnostic performance between 68Ga-FAPI46 and 18F-FDG, with higher specificity (81.1% vs. 36.8%) and PPV (87.5% vs. 66.2%) for 68Ga-FAPI46 by histopathologic validation. Further validation by the composite reference standard showed excellent performance for 68Ga-FAPI46 on a per-region basis in all 4 parameters (sensitivity, 96.6%; specificity, 90.3%; PPV, 90.0%; and NPV, 97.1%) compared with 18F-FDG (sensitivity, 92.2%; specificity, 57.5%; PPV, 65.6%; and NPV, 89.4%). Regardless of the type of validation, 68Ga-FAPI46 showed better diagnostic accuracy than 18F-FDG (Table 2) and especially higher specificity and PPVs. This result is explained by more false-positive regions (68Ga-FAPI46, n = 7, vs. 18F-FDG, n = 31) and false-negative regions (68Ga-FAPI46, n = 2, vs. 18F-FDG, n = 5) for 18F-FDG.
Diagnostic Performance of 68Ga-FAPI46 PET and 18F-FDG PET Compared with Confirmed Reference Validation
Detection Rate
Detection rates for 68Ga-FAPI46 versus 18F-FDG versus Ce-CT (n = 41) are summarized in Table 3. Overall, 252 lesions were detected by any modalities combined. 18F-FDG PET/CT had a significantly higher lesion detection rate than 68Ga-FAPI46 and Ce-CT (18F-FDG [n = 219], 86.8%, vs. 68Ga-FAPI46 [n = 144], 57.1%, vs. Ce-CT [n = 128], 50.8%; P < 0.0001). No significant difference in detection was observed on a per-patient (P = 0.3) or per-region (P = 0.06) basis. But especially in the local nodal region, more PET-positive results were reported for 18F-FDG (18F-FDG [n = 27], 65.9%, vs. 68Ga-FAPI46 [n = 13], 31.7%, vs. Ce-CT [n = 21], 51.2%) but with a large proportion of false-positive results for 18F-FDG (i.e., 18F-FDG, 31/137 false-positive results; 68Ga-FAPI46, 7/137 false-positive results). Figure 3 shows a case example of discordant imaging findings, with false-positive mediastinal lymph nodes on 18F-FDG. Endobronchial ultrasound–guided transbronchial fine-needle aspiration ruled out metastatic lymph nodes and diagnosed nodal anthracosis.
Detection Rates Compared for 68Ga-FAPI46 vs. Ce-CT vs. 18F-FDG
Male patient with suspected pleural mesothelioma before surgical biopsy. Ce-CT shows pleural mass and nodular thickening of left pleura with no morphologic signs of local nodal involvement (lymph nodes < 10 mm). 18F-FDG shows high circular uptake in pleura (SUVmax, 11.4) and additionally positive lymph nodes (SUVmax, 5.5) in levels 4 R/L, 5, 7, and 10 L (red arrows), whereas 68Ga-FAPI46 shows only high uptake in pleura (SUVmax, 8.7). Thoracic lymph node levels are determined by International Associate for the Study of Lung Cancer. R = right; L = left. Endobronchial ultrasound–guided transbronchial fine-needle aspiration ruled out metastatic diseases of respective lymph node levels and revealed nodal anthracosis. MIP = maximum-intensity projection.
DISCUSSION
We established this single-center, prospective, observational trial to investigate the diagnostic value of novel FAP-targeted radioligands and to explore the relationship between 68Ga-FAPI46 uptake and FAP expression levels. Furthermore, we report on the diagnostic performance of 68Ga-FAPI46 in a large group of mesothelioma patients for, what is to our knowledge, the first time.
In line with findings for other tumor entities, we found a moderate association of 68Ga-FAPI46 uptake with immunohistochemical FAP expression, proving our hypothesis that 68Ga-FAPI46 PET is a noninvasive tool for FAP measurement in malignant mesothelioma. Measuring FAP expression is a prerequisite for FAP-targeted treatment, and our results therefore indicate that 68Ga-FAPI46 PET is a suitable tool for noninvasive measurement. Theranostic approaches with 68Ga- and 177Lu-/90Y-FAPI have not, to our knowledge, been investigated in mesothelioma patients but could enable novel cancer treatment strategies targeting both tumor cells and the microenvironment simultaneously in other mesenchymal tumor entities (18,19).
However, the degree of correlation of immunohistochemical FAP expression and 68Ga-FAPI46 PET signal may be constrained by factors such as specimen heterogeneity and the limited sample size in this study. Consequently, our findings must be interpreted with caution. Aside from that, we demonstrated 68Ga-FAPI46 PET/CT to have high diagnostic performance by histopathologic and composite reference validation. Sensitivity was similar for 68Ga-FAPI46 and 18F-FDG PET. However, 68Ga-FAPI46 showed substantially higher specificity and PPV than did 18F-FDG (e.g., sensitivity of 90.3% vs. 57.5% and PPV of 90.0% vs. 65.6%). Possible reasons for this discrepancy are reactive lymph nodes, anthracosis, or inflammation, which results in false-positive findings on 18F-FDG PET/CT (20). These are known general pitfalls in 18F-FDG PET, with, however, high relevance in patients with mesothelioma, who often have other underlying inflammatory disease such as asbestosis or other pneumoconiosis (21,22). Independently of validation and reference standard, 18F-FDG had higher detection rates with an overall mediocre specificity, PPV, and NPV, which reflects these findings. In this setting, 68Ga-FAPI46 PET/CT provides better discrimination between unspecific or reactive processes and actual tumor manifestations. Chronic inflammation accompanying fibrotic or granulomatous changes might demonstrate increased FAPI accumulation (20,23,24). However, in the context of malignant mesothelioma, inflammation-induced microscopic fibrotic changes in lymph nodes might be too subtle to produce relevant FAPI uptake.
Because of the variety of common histopathologic subtypes of mesothelioma, different tracers might prove beneficial for selected subtypes. Studies suggest lower 18F-FDG uptake in epitheloid and nonpleomorphic mesothelioma, but differences in diagnostic performance have not been established so far. A recent study suggested a significantly higher SUVmax for 68Ga-FAPI than for 18F-FDG—a finding that could only partially be verified in our cohort (25). Still, there are no reliable data comparing imaging tracers by subtype, a comparison that appears challenging because of the rarity of some subtypes.
This trial was limited by its observational, single-center character and by a recruited cohort that reflects a real-life patient population rather than a controlled and randomized study population. Accordingly, some patients were pretreated, and specimens for histopathologic evaluation could not be acquired in all cases. Furthermore, a minority of patients underwent follow-up imaging or treatment elsewhere. Nonetheless, the 68Ga-FAPI46 PET data show robust high diagnostic performance as compared with histopathologic and composite reference validation. Although Ce-CT remains the proposed standard imaging method, the superior diagnostic efficacy of 18F-FDG for mesothelioma led us to select it as the primary comparator for 68Ga-FAPI46 (26). In this cohort, we investigated the detection rates of only Ce-CT and 68Ga-FAPI46, but future investigations should also assess the advantages over conventional imaging methods.
CONCLUSION
This prospective observational study demonstrated an association between target FAP expression levels and 68Ga-FAPI46 uptake. 68Ga-FAPI46 PET for mesothelioma imaging revealed specificity and PPV higher than—as well as sensitivity similar to—that of 18F-FDG PET. 68Ga-FAPI46 PET is a valuable diagnostic tool in patients with mesothelioma. The therapeutic potential of radiolabeled FAPIs should be investigated in the future.
DISCLOSURE
Hubertus Hautzel receives personal fees from Roche and Urenco and other fees from Pari, Roche, and Urenco outside the submitted work. Wolfgang Fendler reports fees from SOFIE Biosciences (research funding), Janssen (consultant, speaker), Calyx (consultant, image review), Bayer (consultant, speaker, research funding), Novartis (speaker, consultant), Telix (speaker), GE Healthcare (speaker), Eczacıbaşı Monrol (speaker), Abx (speaker), Amgen (speaker), and Urotrials (speaker) outside the submitted work. Kim Pabst received fees from Bayer (research funding) and Novartis (speakers’ bureau) and travel fees from IPSEN and has received a Clinician Scientist Stipend from the University Medicine Essen Clinician Scientist Academy (UMEA) sponsored by the faculty of medicine and Deutsche Forschungsgemeinschaft (DFG) outside the submitted work. Benedikt Schaarschmidt has received a research grant from the Deutsche Forschungsgemeinschaft for an ongoing investigator-initiated study not related to this work, as well as research grants from PharmaCept and the Else Kröner-Fresenius Foundation for a past investigator-initiated study not related to this work. Benedikt Schaarschmidt has received speaker fees from AstraZeneca and a travel grant from Bayer AG. Ken Herrmann reports personal fees from Bayer, personal fees and other fees from SOFIE Biosciences, personal fees from SIRTEX, nonfinancial support from ABX, personal fees from Adacap, personal fees from Curium, personal fees from Endocyte, grants and personal fees from BTG, personal fees from IPSEN, personal fees from Siemens Healthineers, personal fees from GE Healthcare, personal fees from Amgen, personal fees from Novartis, personal fees from ymabs, personal fees from Aktis Oncology, personal fees from Theragnostics, personal fees from Pharma15, personal fees from Debiopharm, personal fees from AstraZeneca, and personal fees from Janssen. Jens Siveke receives honoraria as a consultant or for continuing medical education presentations from AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Immunocore, MSD, Novartis, Roche/Genentech, and Servier. His institution receives research funding from Abalos Therapeutics, Boehringer Ingelheim, Bristol-Myers Squibb, Celgene, Eisbach Bio, and Roche/Genentech; he holds ownership and serves on the board of directors of Pharma15, all outside the submitted work. No other potential conflict of interest relevant to this article was reported.
KEY POINTS
QUESTION: Is there an association between FAPI uptake intensity and FAP expression in malignant mesothelioma, and what is the diagnostic performance of FAPI PET in mesothelioma patients?
PERTINENT FINDINGS: The SUVmax and SUVpeak of 68Ga-FAPI46 were significantly associated with immunohistochemical FAP expression, and the overall diagnostic performance of 68Ga-FAPI46 PET compared well with that of 18F-FDG.
IMPLICATIONS FOR PATIENT CARE: We established the diagnostic utility of 68Ga-FAPI46 PET in patients with malignant mesothelioma, with implications for FAP-targeted therapies.
Footnotes
Published online Jul. 3, 2024.
- © 2024 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- Received for publication January 21, 2024.
- Accepted for publication May 22, 2024.
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.