Increased uptake of 68Ga-DOTA-FAPI-04 in bones and joints: metastases and beyond

To describe the uptake of 68Gallium-labelled fibroblast activation protein inhibitor (68Ga-FAPI) in the bones and joints for better understanding of the role of 68Ga-FAPI PET in benign and malignant bone lesions and joint diseases. All 129 68Ga-FAPI PET/MR or PET/CT scans from June 1, 2020, to February 20, 2021, performed at our PET center were retrospectively reviewed. Foci of elevated 68Ga-FAPI uptake in the bones and joints were identified. All lesions were divided into malignant and benign diseases. Benign lesions included osteofibrous dysplasia, periodontitis, degenerative bone diseases, arthritis, and other inflammatory or trauma-related abnormalities. The number, locations, and SUVmax of all lesions were recorded and analyzed. The detectability of 68Ga-FAPI PET and 18F-FDG PET in patients who had two scans was also compared. Elevated uptake of 68Ga-FAPI in/around the bones and joints was found in 82 cases (63.57%). A total of 295 lesions were identified, including 94 (31.9%) malignant lesions (all were metastases) and 201 (68.1%) benign lesions. The benign lesions consisted of 13 osteofibrous dysplasia, 48 degenerative bone disease, 33 periodontitis, 56 arthritis, and 51 other inflammatory or trauma-related abnormalities. The spine, shoulder joint, alveolar ridge, and pelvis were the most commonly involved locations. Bone metastases were mainly distributed in the spine, pelvis, and ribs. Among benign diseases, periodontitis and arthritis are site-specific. The mean SUVmax of bone metastases was significantly higher than that of benign diseases (7.14 ± 4.33 vs. 3.57 ± 1.60, p < 0.001), but overlap existed. The differences in SUVmax among subgroups of benign diseases were statistically significant (p < 0.001), with much higher uptake in periodontitis (4.45 ± 1.17). 68Ga-FAPI PET identified much more lesions than 18F-FDG PET (104 vs. 48) with higher uptake value. 68Ga-FAPI accumulated in both bone metastases and some benign diseases of the bones and joints. Although the uptake of 68Ga-FAPI was often higher in bone metastases, this finding cannot be used to distinguish between benign and malignant lesions. 68Ga-FAPI PET also has the potential to locate and evaluate the extent of both malignant tumor and benign diseases in bones and joints. NCT04554719, NCT04605939. Registered September 8, 2020 and October 25, 2020—retrospectively registered, http://clinicaltrails.gov/show/NCT04554719; http://clinicaltrails.gov/show/NCT04605939


Background
Fibroblast activation protein (FAP) is overexpressed in cancer-associated fibroblasts in > 90% of epithelial tumors and in activated fibroblasts associated with remodelling of the extracellular matrix [1]. It is a promising target for the diagnosis of tumors and some nonmalignant diseases [2,3]. Various quinolone-based FAP inhibitors (FAPIs) radiolabelled with different radionuclides demonstrated specific binding to FAP [4][5][6]. Their clinical application has been of interest over the past 2 years [7,8]. 68 Gallium-labeled FAPI ( 68 Ga-FAPI) positron emission tomography/computed tomography (PET/ CT) has been used for imaging various types of cancers and outperforms fluorine-18 fludeoxyglucose ( 18 F-FDG) PET in cancers such as gastrointestinal cancers, cholangiocarcinoma, and hepatocellular cancer [9][10][11]. The potential of FAPI as a theranostic tool also warrants special attention [12,13].
Several studies have reported that 68 Ga-FAPI outperformed 18 F-FDG PET/CT in identifying bone metastases [11,[22][23][24]. Two case reports revealed that intense 68 Ga-FAPI uptake was observed in shoulder arthritis [25] and degenerative osteophyte [26]. FAP expression could be found in some benign diseases and normal tissues during remodelling by stromal cells and mesenchymal stem cells, like embryogenesis, wound healing, fibrotic reactions, inflammatory conditions, atherosclerotic plaques, degenerative osteophytes, and rheumatoid myofibroblastlike synoviocytes [27][28][29][30][31], which explains the uptake of 68 Ga-FAPI in these entities. Therefore, we hypothesize that 68 Ga-FAPI has valuable diagnostic implications in both malignant and benign bone and joint diseases. To date, there is no data on the analysis of 68 Ga-FAPI uptake in human bones and joints in a large patient cohort.
We retrospectively reviewed the images of all patients undergoing 68 Ga-FAPI PET/MRI or PET/CT in our PET center, and reassessed 68 Ga-FAPI accumulation in the bones and joints with a focus on the uptake pattern of 68 Ga-FAPI in malignant and benign bone and joint lesions, aiming to avoid misdiagnosis in ever-increasing applications of 68 Ga-FAPI PET in tumor diagnosis.

Patients
We retrospectively reviewed the images of all patients who underwent 68 Ga-FAPI PET/CT or PET/MR in our PET center from June 1, 2020, to February 20, 2021. The patient cohort is a subgroup of two prospective trials focusing on the diagnostic accuracy of 68 Ga-FAPI PET in malignant tumors and fibrous diseases (ClinicalTrials.gov Identifier: NCT04554719 and NCT04605939). The study was approved by the Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology. All patients signed an informed consent before undergoing 68 Ga-FAPI PET imaging. 68 Ga-FAPI PET/MR or PET/CT scanning 68 Ga-FAPI was synthesized according to a previously published method [11]. In brief, 68 Gallium ( 68 Ga) was eluted from a 68 Ge/ 68 Ga generator (ITG, Munich, Germany) and reacted with 25 μg (28.6 nmol) of DOTA-FAPI-04 ligand (CS Bio Company, Menlo Park, CA, USA) in a manual synthesis module (ITG). The pH value was adjusted to 4.0 using 2.5 M sodium acetate buffer. After heating to 100 °C for 20 min, the product was purified and isolated by a solid-phase column (Sep-Pak, Waters Associates, Milford, MA, USA). Quality control of the radiosynthesis was performed by high-performance liquid chromatography, and radiochemical purity exceeding 95% was accepted for injection. A total of 1.85-3.70 MBq (0.05-0.10 mCi)/kg 68 Ga-FAPI was administered intravenously.
PET/MR or PET/CT scanning was started at 20-60 min after 68 Ga-FAPI administration. PET/MR was performed on a hybrid PET/MR scanner (3.0 T, SIGNA TOF-PET/MR, GE Healthcare, Milwaukee, WI, USA). MR sequences included an axial high-resolution T1-weighted fast spoiled gradientecho sequence (LAVA Flex sequence) and T2-weighted (T2W) sequence from the skull base to the upper thighs, and diffusion-weighted sequences were acquired if needed. PET was acquired simultaneously in four to five bed positions with an acquisition time of 3 min per bed position (axial FOV: 25.8 cm, matrix size: 344 × 344). Images of the head were acquired separately. PET/CT (Discovery VCT, GE Healthcare) was performed in patients who were suspected of pulmonary fibrosis or with contraindications to MR. Both CT and PET were acquired from the skull vertex to the upper thighs with PET acquisition for 3 min per bed position. All images were reconstructed using a 3D attenuation-weighted iterative ordered-subset expectation maximization algorithm with 3 iterations and 21 subsets. 1 3 24 years of experience and C. Q., with 12 years of experience in nuclear medicine) and two radiologists (X.L., with 24 years of experience in orthopaedic radiology and F.L. with 20 years of experience in radiology and 5 years of experience in nuclear oncology) reviewed PET/MR and PET/CT images and conventional images (including X-ray, CT, or MR). All reviewers analyzed the images independently. Any differences in opinion were resolved by consensus.
Positive lesions were defined as visually elevated uptake of 68 Ga-FAPI or 18 F-FDG in/around the bones or joints compared to the surrounding background tissue and enrolled for analysis. For each patient, the assessed dataset included: -Number of lesions Number of lesions was recorded, if the number in one patient was > 20, the 20 lesions with the highest uptake were included. -Locations of lesions The locations were identified as being located in the joints or bones including the skull, alveolar ridge, sternum, spine, ribs, and appendicular skeleton. -Classifications of lesions All of the lesions were determined to be malignant (bone metastases) or benign based on a comprehensive analysis of the patient's medical history and previous or follow-up imaging.
The criteria for metastatic lesions were as follows: (1) patient who was diagnosed with malignancy; (2) solitary neo-growth bone lesion or multiple bone lesions; (3) with at least one of the following high-risk features of bone metastases [32,33]: 1) lesions with abnormal signal changes on MRI; 2) bone destruction on CT; 3) continued to increase in size or number; 4) patients with local bone pain which aggravates at night; 5) elevated tumor markers; 6) concomitant metastases in other sites; 7) patients with hypercalcemia.
The criteria for benign lesions were as follows: (1) lesions occurred at specific sites, including the joint, tendon attachment site, alveolar bone, vertebral body edge, or intervertebral space; (2) or no significant changes were observed in the current images compared to the previous or follow-up images, and metastases were excluded; (3) or characteristic imaging features, for example, presence of fracture line or cortical distortion in fracture, ground-glass density shadow with/without cystic lesions in osteofibrous dysplasia. Benign lesions were classified into five subgroups: osteofibrous dysplasia, degenerative bone disease, periodontitis, arthritis, and other inflammatory or trauma-related abnormalities (such as tendinitis, fracture, or healing injuries).
-SUVmax Maximum standardized uptake values (SUVmax) for all lesions were measured.

Statistical analyses
A commercial software package (SPSS 26.0, IBM Inc., Armonk, NY, USA) was employed for statistical analysis. Continuous variables are expressed as mean ± SD. Categorical variables are expressed as number and percentage. For quantitative data, normal distribution and homogeneity of variances were tested first. Quantitative data with non-normal distribution were compared using Mann-Whitney U-test between two groups, and Kruskal-Wallis H-test among the five groups. Independent sample t-test was used for comprising the differences of SUVmax between 68 Ga-FAPI and 18 F-FDG. A p-value < 0.05 was considered statistically significant. 3) and tendon attachment in bilateral ischial tuberosities (a and c, blue arrows; SUVmax 8.1), which were diagnosed as benign inflammatory lesions. A focus of elevated activity with SUVmax 5.5 in the left ilium was diagnosed as osteofibrous dysplasia because of characteristic ground-glass opacity with the low-density shadow which was the same as 2 years ago (d, blue arrows). Multiple aggregation of 68 Ga-FAPI distributed in the ribs (e, red arrows; SUVmax 5.5), the fifth lumbar and first sacral vertebrae (f, red arrows; SUVmax 3.5, 7.6) with new emerging high-density shadows on CT images compared with the CT images 5 months ago, which were considered to be metastases

Results
The image data of 129 cases were retrospectively reviewed. Elevated uptake of 68 Ga-FAPI in/around the bone or joint was found in 82 cases (63.6%), which were enrolled in this study. All patients (47 males and 35 females) had a mean age of 56.65 ± 10.15 years (range, 26-80 years). A total of 73 patients underwent 68 Ga-FAPI PET/MR and 9 patients underwent PET/CT. Table 1 presents the general information of the enrolled patients. In 82 cases, a total of 295 68 Ga-FAPI-avid bone/joint lesions were identified, including 94 malignant bone lesions in 15 cases and 201 benign bone/joint lesions in 76 cases, of which 9 cases had both malignant and benign lesions ( Fig. 1 and Fig. 2). All malignant lesions were diagnosed as bone metastases. The benign diagnoses were as follows: osteofibrous dysplasia (n = 13, 4.4%), degenerative bone disease (n = 48, 16.3%), periodontitis (n = 33, 11.2%), arthritis (n = 56, 19.0%), and other inflammatory or trauma-related lesions (n = 51, 17.3%).
For the subgroup of benign diseases, osteofibrous dysplasia occurred at many sites, including the skull, clavicle, ribs, vertebral bodies, pelvis, and long bones, featuring ground-glass opacities with well-defined borders and intact overlying bone on CT images, and these lesions had not changed significantly for years ( Fig. 1d and Fig. 4). Elevated uptake in the alveolar ridges was diagnosed as periodontitis (Fig. 5). The majority of degenerative bone diseases occurred in the spine, usually at the edge of the vertebral body, including osteodysgenesis, osteophyte and Schmorl's node (n = 39, 81.3%; Fig. 2d, Fig. 6d), and intervertebral discs (n = 4, 8.3%). The most representative arthritis occurred in the shoulder (n = 45, 80.4%; Fig. 1b, Fig. 6b) and hip (n = 11, 19.6%; Fig. 6e). Other inflammatory or traumarelated abnormalities mainly appeared at the acromial or sternal end of clavicle (n = 21, 41.2%), spinal facet joints (n = 9, 17.6%; Fig. 6c), sternal angle/sternocostal joint (n = 6, 11.8%), and tendon attachment in ischial tuberosities (n = 6, 11.8%; Fig. 1c). Figure 6 presents one representative case with 17 foci in the bones and joints, all diagnosed as benign entities. One case showed healed fractures in three consecutive ribs (Fig. 7). In areas with coexisting malignant and benign lesions, the proportion of the diseases was different. In the pelvis and ribs, predominating malignant foci were identified (25/31, 20/27). In the vertebral bodies and posterior elements (malignant n = 40; benign n = 42) and skull (malignant n = 3; benign n = 2), as well as the femur (malignant n = 2; benign n = 3), the ratios were ~ 1, which suggests that more attention should be paid to distinguishing between benign and malignant foci in these sites.
Of the 82 enrolled patients, 29 underwent 18 F-FDG PET/CT within 3 days before or after 68 Ga-FAPI-04 PET imaging. To compare the detectability of 68 Ga-FAPI PET and 18 F-FDG PET for bone and joint lesions, we recorded numbers of positive lesions and SUVmax. The results are shown in Table 4. Seventeen of 29 patients had 18 F-FDG-avid lesions in the bones and joints. In all categories of diseases, 68 Ga-FAPI PET depicted more lesions than 18 F-FDG PET. All lesions with increased 18 F-FDG uptake had increased 68 Ga-FAPI uptake, and no additional 18 F-FDG-avid foci were found. For each type of lesions, mean SUVmax of 68 Ga-FAPI was higher than that of 18 F-FDG except degenerative bone disease, but significant difference existed only in periodontitis.

Discussion
In this study, we found that 68 Ga-FAPI accumulated in both bone metastatic foci and benign bone and joint diseases. 68 Ga-FAPI PET imaging is mainly used for tumor diagnosis and follow-up [3]. Increased 68 Ga-FAPI uptake in primary tumors and metastases occurs in various cancers, including bone metastases [9,10,34,35]. However, in this study, we found that 68.1% of foci with elevated uptake in the bones and joints were benign disease entities. Since benign bone and joint diseases often occur alongside malignant disease (as shown in Fig. 1 and Fig. 2), more attention should be paid to the positive findings of FAPI-PET. Thus, extensive knowledge of the specific patterns of 68 Ga-FAPI uptake in bone and joint is essential for accurate differential diagnosis between benign disease and tumor metastases.
We found that the characteristics of metastases include higher 68 Ga-FAPI uptake, a higher number of foci, and predominant occurrence in the axial skeleton. Based on 201 benign lesions in this study, the features of 68 Ga-FAPI-positive benign diseases included more solitary lesions, lower uptake, and specific locations, including the shoulder joint, hip joint, alveolar ridge, acromial or sternal end of clavicle, spinal facet joint, sternal angle/sternocostal joint, and tendon attachment in ischial tuberosity. Although bone metastases had higher 68 Ga-FAPI uptake than benign bone and joint lesions, our results revealed overlapping uptake values by malignant and benign lesions (Fig. 3). Therefore, differential diagnosis should be carefully made in cancer patients, especially with solitary bone lesion. Periodontitis, spinal degenerative disease, arthritis, and tendonitis were easily diagnosed due to their specific locations. Lesions involving the skull, ribs, vertebral bodies, pelvis, and long bones could be benign or malignant, necessitating comprehensive analysis with medical history and radiographic imaging for accurate diagnosis.
On the other hand, our results demonstrated that 68 Ga-FAPI PET detected more bone metastases and presented higher SUVmax compared to 18 F-FDG PET, consistent with previous studies [10,11,22]. For benign bone and joint diseases, 68 Ga-FAPI PET also revealed much more Fig. 4 A 58-year-old male underwent 68 Ga-FAPI PET/MR for detecting recurrence and metastases after 5-year comprehensive treatment of gastric cancer. Aggregation of 68 Ga-FAPI on the left fourth rib (a and b, blue arrows) and right ninth rib (a and c, blue arrows) appeared with heterogeneous signals on MR (b and c, blue arrows on fused images), corresponding to well-demarcated lesions with irregular ground-glass opacities and part of low-density shadow on the recent CT, same as CT 2 years ago (b and c, blue arrows on CT images). From the medical history and imaging findings, the two lesions were highly suspected as osteofibrous dysplasia lesions and higher tracer uptake than 18 F-FDG PET. Extensive positive uptake of 68 Ga-FAPI indicates its high sensitivity. Besides evaluating malignant tumors and metastases, it also has the potential role to locate and evaluate the extent of some benign diseases, such as inflammatory joint diseases, and benign bone tumors related to fibrosis (like fibrous dysplasia), cardiac, pulmonary, or liver fibrosis [17,[36][37][38][39].
In this study, 13 bone foci were diagnosed as osteofibrous dysplasia, which constitutes 5% of all benign bone lesions [40] and generated by replacement of portions of normal bone by focal fibrous proliferation, which results in disorganized trabecular formation [41]. On CT, it usually has typical features of ground-glass density and cystic low-density areas [42]. Elevated 68 Ga-FAPI uptake was observed in the alveolar ridges, which may be related to the fibroblasts centrally involved in the reformation of periodontal tissue structures in the healing response [43]. Our results revealed SUVmax of periodontitis is much higher than that in other benign conditions, which likely indicates more active fibroblasts. High uptake at both ends of clavicle and tendon attachment in ischial tuberosities could be explained as inflammation or tendonitis because FAP is highly expressed during tissue repairing. Similarly, both 68 Ga-FAPI-02 and 68 Ga-FAPI-04 have been reported to have uptake in locations of wound healing after surgical intervention [1]. In our study, 48 degenerative bone lesions revealed high uptake. A recent report demonstrated 68 Ga-FAPI accumulation in degenerative osteophytes [44]. The 68 Ga-FAPI uptake by degenerative osteophytes may be a result of the presence of fibroblasts in early osteophytes [45,46]. FAPα is reported to be expressed in human and rat degenerative intervertebral disc disease [28]. The corresponding CT or  68 Ga-FAPI presented on the right maxillary first and second molars (a, blue arrows) in a 48-year-old man with a history of dermatomyositis and interstitial pneumonia. A 54-year-old woman with a history of gastric cancer also revealed elevated uptake of 68 Ga-FAPI in the mandibular central incisor and first premolar with partial denture (b, blue arrows). The lesions were diagnosed as periodontitis based on clinical symptoms of toothache and imaging findings Fig. 6 A 61-year-old male underwent 68 Ga-FAPI PET/MR for detecting the primary lesion of unexplained ascites. Diffuse intense uptake was observed in the extensive thickened gastric wall, omentum, and peritoneum (a), which suggested gastric cancer with celiac extensive metastases. Unexpectedly, elevated 68 Ga-FAPI uptake appeared on 17 foci of bones and joints which were all diagnosed as benign, just around the left shoulder joint (b, blue arrows), facet joints (c, blue arrows), edge of vertebral body (d, blue arrows), right femoral head (e, blue arrows), and tendon attachment in bilateral ischial tuberosities (f, blue arrows) with SUVmax of 3.3-4.2. MRI revealed no significant signal change. Lesions at the edge of vertebral body were diagnosed as retrogression, while the rest were considered to be inflammations such as arthritis and tendinitis MR images revealed the specific location of the disease, which made it easier to distinguish degenerative disease from bone metastases.
There are several limitations in this study. First, the types of bone diseases included were limited. No primary bone malignancy was included. Only a few types of benign bone diseases were identified and analyzed. Second, no pathological evidence was obtained. The final diagnosis was based on comprehensive analysis, including clinical data, imaging features, and follow-up results. Third, this was an observational study with a limited number of patients, and the original purpose of the clinical researches was not to explore the diagnostic value of 68 Ga-FAPI PET in bone and joint lesions. In spite of these limitations, we found that both benign and malignant bone lesions could accumulate 68 Ga-FAPI, which suggests that we should pay attention to the differential diagnosis for malignant bone lesions, especially for the single lesion, since the diagnosis of bone metastasis may upstage the disease and change treatment. 68 Ga-FAPI can be taken up by both bone metastases and benign osteoarticular lesions. The uptake in bone metastases was higher than that of benign diseases in the bones and joints, but there were overlapping uptake values. Abnormal osseous 68 Ga-FAPI uptake should be carefully assessed in patients with malignant tumors to avoid misdiagnosis, especially in patients with solitary bone lesion. 68 Ga-FAPI PET also has the potential role to locate and evaluate the extent of several benign diseases.