The characteristics of the registered cases are shown in Table 1. Clinical diagnosis was confirmed through histopathology for well-differentiated adenocarcinoma (ADC), medullary thyroid cancer (MTC), papillary thyroid cancer (PTC), cervical, gastric, glioblastoma multiform (GBM), colon, Ewing's sarcoma, and breast cancer who underwent 68Ga-FAPI-46 and 18F-FDG PET/CT scans. The mean tumor size was 2.6 ± 1.7 cm, with a minimum and maximum length of 1.2 cm and 5.4 cm, respectively. Overall, 9 patients underwent surgical resection, 3 patients received a concurrent cycle of 177Lu-DOTATATE and 177Lu-Trastozumab. In addition, 7 patients received chemotherapy cycles at least eight months before dual-tracer imaging. The remaining two patients underwent non-surgical antitumor treatment.
Table 1
Characteristics of patients involved in this study.
No. | Sex | Age | Pathology | Metastases sites | Indication |
1 | M | 46 | Well differentiated ADC | 5 liver mass lesions in both liver lobes | Response |
2 | F | 46 | PTC | Possibility of initial phases of heterotopic ossification lateral to the resected region of right sacroiliac bone, cannot be ruled out | Metastatic |
3 | F | 58 | Cervix cancer | Peritoneum and mass lesion in the left side of the pelvic more compatible with peritoneal seeding and mesenteric involvement. | Recurrence |
4 | M | 58 | Gastric ADC | Hypermetabolic cervical lymph node (LN) in the right side of the neck (level 2) without FAPI uptake. For definite diagnosis tissue sampling is recommended. | Recurrence |
5 | M | 55 | GBM | Right frontotemporal lobes | Metastatic |
6 | M | 53 | Gastric ADC | Cardia region | Recurrence |
7 | F | 41 | Colon ADC | Peritoneal seeding and mesenteric LNs metastasis | Metastatic |
8 | M | 10 | Ewing sarcoma | Left fibula | Recurrence |
9 | F | 53 | Breast cancer | Left breast, left axillary LNs | Metastatic |
10 | M | 33 | MTC | Right upper paratracheal | Metastatic |
11 | F | 31 | MTC | LNs in the superior mediastinum, paratracheal and pre-carinal | Metastatic |
Comparison of visual assessment
In another case, a patient with well-differentiated adenocarcinoma underwent left hemicolectomy two years before imaging and received chemotherapy continuously for 1 year. The patient had a history of radiofrequency ablation (RF) of liver metastasis immediately after the last chemotherapy course. After undergoing 68Ga-FAPI-46 PET/CT for response evaluation, the patient underwent liver metastasectomy followed by chemotherapy sessions. Four days later, 18F-FDG PET scan was performed to detect the metastatic lesions. Five liver mass lesions in both liver lobes were observed in both FDG and FAPI PET scans (Fig. 2).
In two patients with histopathological confirmation of gastric adenocarcinomas and rising tumor markers (CA = 19 − 9), there were 2 hypermetabolic cervical lymph nodes in the neck region without FAPI uptake. For another patient with a history of papillary thyroid cancer who underwent total thyroidectomy and resection of the right iliac bone, the possibility of initial phases of heterotopic ossification lateral to the resected region of the right sacroiliac bone, could not be ruled out. In this case, both tracers indicated the same involved regions. A patient with a history of GBM in the right frontotemporal lobe underwent surgical resection, radiotherapy, and continuous 177Lu-DOTATATE radionuclide therapy as final treatment procedure. In this case, we performed three different scans for a better assessment of treatment options available in our center. 68Ga-FAPI-46 and 68Ga-Pentixafor PET scans exhibited the same uptake in the target region with SUVmax of 1.02 and 4.17, respectively, while 68Ga-DOTATATE PET scan did not indicate any remarkable uptake in the lesion (similar to background). In the known case of Ewing sarcoma (left fibula) who underwent surgical resection and chemotherapy, no abnormal hypermetabolic lesion at the surgical bed (left fibula) was observed. Two hypermetabolic nodules were found in the pulmonary region and paratracheal without FAPI uptake. Discordant uptake of FDG and FAPI in pulmonary nodule (RUL) and paratracheal lymph node required tissue sampling in order to exclude metastatic involvement. Interestingly, two lesions were pathologically diagnosed as malignant and having metastatic involvement. In short, this case showed that if FAPI is negative for recurrence evaluation, imaging using another tracer, such as routine FDG PET should be performed to distinguish malignant from benign lesions or inflammation (Fig. 4).
A known case of colon adenocarcinoma underwent right hemicolectomy and received chemotherapy before FDG and FAPI imaging. A hypermetabolic peritoneal mass lesion in the right lower quadrant (RLQ) was observed, which is more compatible with peritoneal seeding. In addition, peritoneal nodularities were observed at the surface of the liver, with FAPI uptake compatible with peritoneal seeding. Moreover, two hypermetabolic right renal artery lymph nodes, linked to patients' history, were more compatible with metastatic involvement. Lung infiltration in the basal left lung observed in FDG and FAPI scans was more compatible with inflammatory reaction. Three hypermetabolic small mass lesions existed in the right liver lobe (segment 8) that were reflected as metastatic involvement on FDG PET, while FAPI PET/CT showed no abnormal uptake throughout the liver (Fig. 5).
Two patients with a history of MTC, who underwent peptide receptor radionuclide therapy (PRRT) using one cycle of 177Lu-DOTATATE, were referred for metastatic work-up due to rising calcitonin score (calcitonin scores for patient number 10 and 11 was 2435 and 11173, respectively). In patient #10, paratracheal lymph node metastasis, multiple bilateral non-FAPI avid pulmonary nodules, and a left peri-bronchial lymph node with mild FAPI uptake were found in the FAPI PET scan. Patient #11 showed multiple FAPI-avid lymphadenopathies in the superior mediastinum, paratracheal, pre-carinal, and sub-nodular lesions in both lungs with no FAPI uptake. 68Ga-FAPI-46 PET/CT scan revealed focal uptake without a corresponding 18F-FDG uptake in a sclerotic lesion at the 9th right rib, which was suspected of bone metastasis according to our criteria. Images of the discordant cases were presented in Fig. 7.
The false positive result observed in patient #8 in the FAPI PET scan could be justified by post-operative inflammation caused by primary tumor excision about one month before PET studies. Comparative results of 18F-FDG and 68Ga-FAPI-46 PET/CT for recurrence detection are shown in Table 2. A difference in recurrence detection between the FDG and FAPI scans was observed in 5 out of 11 patients. The SUVmax and TBR values for 68Ga-FAPI and 18F-FDG PET/CT did not differ significantly for the cervix and gastric metastases detection, while the SUVmax and TBR values of 68Ga-FAPI were superior to those of 18F-FDG-PET/CT in Ewing sarcoma, MTC, breast, and colon cancer. Although there was no significant difference in SUVmax between these two imaging modalities for the detection of liver metastases, 68Ga-FAPI scan exhibited significantly higher TBR values compared to 18F-FDG scan (Table 3). Figures 2 and 3 show examples of maximum intensity projections from 18F-FDG PET/CT and 68Ga-FAPI PET/CT scans of different types of cancer.
Table 2
Comparison of 18F-FDG and 68Ga-FAPI-46 PET/CT scans for recurrence detection.
Patient No. | Base tumor site | Recurrence site | 18F-FDG avidity | 68Ga-FAPI-46 avidity | PET/CT result |
1 | Well differentiated ADC | Liver | Y | Y | *5 liver mass lesion in both liver lobes in the segments of 4,5, 6, 7 showed both FDG and FAPI uptake and compatible with metastatic involvement. *Progression of disease is noted. (Interval time is 4 days). |
2 | PTC | Sacroiliac | Y | Y | Possibility of initial phases of heterotopic ossification lateral to the resected region of right sacroiliac bone, cannot be ruled out. |
3 | Cervix cancer | Peritoneal | N | Y | Comparing with FDG PET/CT, Ga-68-FAPI PET/CT showed increased activity in the peritoneum and mass lesion in the left side of the pelvic (which is not FDG avid) more compatible with peritoneal seeding and mesenteric involvement. |
4 | Gastric ADC | Cervical | Y | N | *Hypermetabolic cervical LN in the right side of the neck (level 2) without FAPI uptake. *FAPI uptake in the focal peritoneal thickening in the mid line of the abdomen at the level of L2 (without hypermetabolism) regarding patient's history more compatible with deposited. For definite diagnosis, tissue sampling is recommended. |
5 | GBM | Frontotemporal | - | Y | CXCR4 and FAPI avid lesions in the right frontal and right temporal lobes (three mass lesions) more compatible with tumoral recurrence. |
6 | Gastric ADC | Cardia | Y | N | *Two hepatic metastases are noted in the dome of the liver and in segment VI *No other active lesion is noted in the rest of the body |
7 | Colon ADC | Peritoneal | Y | N | *Hypermetabolic peritoneal mass lesion in the RLQ more compatible with peritoneal seeding. In addition, there are peritoneal nodularities at the surface of the liver with FAPI uptake more compatible with peritoneal seeding. *Two hypermetabolic right renal artery LNs regarding patient's history more compatible with metastatic involvement. *Lung infiltration in the basal left lung with FDG and FAPI uptake more compatible with inflammatory reaction. *Three hypermetabolic small mass lesions in the right liver lobe (segment 8) more compatible with metastatic involvement. In FAPI PET/CT the scan showed no abnormal uptake throughout the liver. (No evidence of liver metastatic lesion) |
8 | Ewing sarcoma | Fibula | Y | N | *Hypermetabolic pulmonary nodule in the RUL without FAPI uptake. *Hypermetabolic right paratracheal LN without FAPI uptake. *Remainder of the study is negative for FDG and FAPI uptake. *Discordant between FDG and FAPI uptake in pulmonary nodule (RUL) and paratracheal LN needs tissue sampling in order to exclude metastatic involvement. |
9 | Breast cancer | Breast | Y | Y | Possibility of left breast involvement. Lymph node in the left axilla, right axilla, mediastinum (paratracheal, prevascular, subcarinaland in the neck of the right side. Left-sided parietal bone metastasis. Multiple lung metastases mainly in the right lung. Pleural involvement in the right lung. |
10 | MTC | Paratracheal | Y | Y | Right upper paratracheal LN metastasis. Multiple bilateral non-FAPI avid pulmonary nodules. Left peri-bronchial LN with mild FAPI uptake. |
11 | MTC | Mediastinum | Y | N | Multiple FAPI-avid lymphadenopathies in the superior mediastinum, paratracheal, pre-carinal, and sub-nodular lesions in both lungs with no FAPI uptake. |
Table 3
Comparison of 18F-FDG, 68Ga-FAPI-46, DOTATATE, and Pentixafor SUVs in the largest involved lesions.
Patient | SUVmax (lesion/background) FDG | SUVmax (lesion/background) FAPI-46 | SUVmax (lesion/background) DOTATATE | SUVmax (lesion/background) Pentixafor | TBR |
1 | 5.29/2.63 | 6.45/1.28 | - | - | 2.011/5.39 |
2 | 12.32/3.99 | 3.01/2.6 | - | - | 3.087/1.15 |
3 | 4.73/3.61 | 4.86/1.94 | - | - | 1.31/2.50 |
4 | 9.13/3.8 | 3.13/1.87 | - | - | 2.40/1.67 |
5 | - | 1.02/0.49 | 2.45/1.75 | 4.17/0.23 | 2.081/1.4/18.13 |
6 | 8.50/3.32 | 24.26/3.32 | - | - | 2.56/7.37 |
7 | 4.59/3.33 | 7.76/1.44 | - | - | 1.37/5.38 |
8 | 31.24/3.5 | 17.35/1.2 | - | - | 8.92/14.45 |
9 | 4.52/2.48 | 14.42/2.9 | - | - | 1.82/4.97 |
10 | 6.46/3.6 | 5.38/1.52 | 4.79/10.91 | - | 1.79/3.53/0.43 |
11 | - | 22.44/2.12 | 5.86/9.42 | - | 10.58/0.62 |
Comparison of the uptake of 68Ga-FAPI, 18F-FDG, 68Ga-Pentixafor and 68Ga-DOTATATE in different types of cancer
Regarding the application of the 68Ga-FAPI tracer in cancer imaging, it is very important to evaluate lesion detection rates and diagnostic efficacy of 68Ga-FAPI compared to 18F-FDG, the dominant radiotracer in oncology. Chen et al. performed a head-to-head comparison of 68Ga-FAPI-04 and 18F-FDG PET/CT in 75 patients (54 patients at initial assessment and 21 patients with recurrence detection), selecting 12 different tumor entities. Their study showed that 68Ga-FAPI-04 compared to 18F-FDG had higher sensitivity in primary tumors (98.2% vs. 82.1%, P = 0.021), lymph node metastases (86.4% vs. 45.5%, P = 0.004), and bone and visceral metastases (83.8% vs. 59.5%, p < 0.005) [16]. Interestingly, Qin et al. reported that 68Ga-FAPI-04 PET/CT detected a lower number of positive lymph nodes compared to 18F-FDG-PET/CT (the number of positive lymph nodes detected was 48 vs. 100) [28]. Their results showed that 68Ga-FAPI PET/CT may be more specific than 18F-FDG to distinguish reactive lymph nodes from tumor-metastatic lymph nodes [29]. In another study of 10 patients with oral squamous cell carcinoma (OSCC), both 68Ga-FAPI-04 PET/CT and 18F-FDG PET/CT had comparable sensitivity and specificity for detecting primary tumors (100% vs. 100%) and cervical lymph node metastases (81.3% vs. 87.5%, p = 0.32; 93.3% vs. 81.3%, p = 0.16) [30]. However, 68Ga-FAPI-04 PET/CT showed a higher TBR compared to 18F-FDG PET/CT for detecting primary tumors (10.90 vs. 4.11) [31]. Komek et al. demonstrated that 68Ga-FAPI-04 PET/CT is capable of detecting a greater number of cancer lesions with a higher SUVmax for primary tumors, including lymph nodes, and distant metastases compared to 18F-FDG PET/CT [32, 33]. In another study, both 18F-FAPI-42 and 18F-FDG had comparable detection rates (100%) for primary tumors in a cohort of 34 patients. However, 18F-FAPI-42 was inferior in detecting brain lesions compared to contrast-enhanced magnetic resonance imaging (CE-MRI) (56 vs. 34, p < 0.005) [34].
Breast cancer
This study showed that the SUVmax of 68Ga-FAPI-46 in breast cancer patients was around 2-fold higher compared to similar studies in the literature. Lymph nodes were observed in the right breast with increased uptake in both FDG and FAPI scans. The largest lymph node was 34 mm in diameter, with SUVmax values of 31.24 and 17.35 for FAPI and FDG, respectively. Moreover, there was a lesion in the left parietal lobe with 27 mm diameter with SUVmax equal to 32.66 and 17.74 in FAPI and FDG scans, respectively. It should be emphasized that due to the very high absorption of glucose in the brain in FDG scans, we were not able to find the involved areas and metastases caused by cancers, while the use of FAPI scan bearing minimal brain background, we were able to detect the involved and suspicious areas (Fig. 6). In contrast to 18F-FDG scans, there was no correlation between tracer accumulation and grade, receptor status, or histological type in 68Ga-FAPI scans. Although the number of patients in our study was insufficient to rule out differences, the findings corroborated immunohistochemical data on consistent FAP expression in breast cancer specimens regardless of tumor type [35]. In addition, an immunohistochemical study found almost exclusive FAP expression in invasive breast cancer but not in ductal carcinoma in situ without microinvasion or common ductal hyperplasia [36]. The presence of additional extra-axillary lymph nodes that are not readily amenable to sentinel lymph node biopsy may influence treatment decisions in breast cancer patients [37]. 68Ga-FAPI scanning supported extra-axillary lymph node involvement in 42% of patients (6 of 14) who would have met inclusion criteria in a study evaluating extra-axillary LN metastases versus the 18F-FDG PET scan with a detection rate of 28% [38]. Although lymph node involvement was not demonstrated by biopsy on a node-by-node basis in any of these studies, the present work and the consistent accumulation in all biopsy-proven axillary lymph node metastases support the good sensitivity of FAPI PET scan. However, a high potential for detecting distant metastases is supported by a recent case series comparing directly 68Ga-FAPI and 18F-FDG PET scans [14, 39, 40]. Since FAPI radiotracer was recently introduced in our center, the number of scans performed for breast cancer was limited, preventing us from quantifying generalizable values of diagnostic accuracy. Previous studies have shown that 68Ga-FAPI PET/CT detects a greater number of cancer lesions with a higher SUVmax for primary tumors, including lymph nodes and distant metastases compared to 18F-FDG PET/CT in breast cancer patients [40]. Komek et al. reached a similar conclusion in a cohort of 20 patients [33], which is consistent with previous studies on FAPI imaging for breast cancer, demonstrating the promise of the 68Ga-FAPI tracer in breast cancer diagnosis.
Papillary Thyroid Carcinoma (PTC)
The incidence of thyroid cancer in the world is gradually increasing. Hence, new imaging techniques are being developed [41]. To the best of our knowledge, there are very few studies comparing FAPI PET/CT and FDG PET/CT imaging on differentiated papillary thyroid carcinoma. In addition to FDG PET/CT, it is known that multiple physiological involvements in different organs or 18F-FDG retention at different levels is observed in infectious/inflammatory events and reactive processes. However, it has been found that non-specific involvements are almost never observed in 68Ga-FAPI PET/CT [14]. The use of 18F-FDG PET/CT for PTC is generally prioritized in patients with postoperative high serum thyroglobulin levels, and negative whole-body radioactive iodine (RAI) scans. The preoperative use of 18F-FDG PET/CT is still controversial [42, 43]. In our study, we focused on the comparison of 68Ga-FAPI-46 and 18F-FDG PET scans in a patient with recurrent thyroid cancer. A 46-year-old woman with a history of PTC underwent total thyroidectomy and resection of the right iliac bone. The possibility of initial phases of heterotopic ossification lateral to the resected region of the right sacroiliac bone cannot be ruled out. There was evidence of right sacroiliac region resection and cement insertion with mild hypermetabolism (SUVmax =5.15), with small zones of possible heterotopic calcification lateral to this region. Based on these results and the features of the tumor, our study showed that 68Ga-FAPI was not inferior to 18F-FDG in a patient with recurrent papillary thyroid cancer (Fig. 3). In addition, 68Ga-FAPI PET can be used as a complementary method with 18F-FDG to detect metastatic foci. According to the literature, 18F-FDG PET scans bear high sensitivity in less-differentiated thyroid carcinomas compared to well-differentiated ones. However, such a limitation is not observed in 68Ga-FAPI PET scans [42]. This issue warrants further investigation using a larger cohort of specific patients. There are studies showing that 68Ga-FAPI PET imaging is more specific in tumoral conditions because of the low levels of expression of the fibroblast-activated protein in the body, but high expression of cancer-associated fibroblasts in the presence of tumor cells [16, 44, 45]. This study's findings supported the idea that combining the results of both PET scans would improve the ability to identify the metastatic focus or foci in challenging cases like recurrent papillary thyroid cancer.
Medullary thyroid cancer (MTC)
Initial studies reported intermediate uptake in medullary thyroid cancers, which obscures the role of FAP imaging [14]. Two patients with a history of MTC who underwent PRRT using one cycle of 177Lu-DOTATATE were referred for a metastatic work-up due to rising calcitonin scores (calcitonin scores for patient #10 and #11 were 2435 and 11173, respectively) (Fig. 3). In patient #10, the right upper paratracheal lymph node metastasis, multiple bilateral non-FAPI avid pulmonary nodules, and a left peri-bronchial lymph node with mild FAPI uptake were found. Patient #11 showed multiple FAPI-avid lymphadenopathies in the superior mediastinum, paratracheal, pre-carinal, and subnodular lesions in both lungs with no FAPI uptake.
Sarcomas
Primary lesion detection rates obtained from 68Ga-FAPI-46 and 18F-FDG PET/CT scans were completely different owing to hypermetabolic pulmonary nodules in the pulmonary nodule (RUL) without any FAPI uptake and hypermetabolic right paratracheal lymph nodes without FAPI uptake. A known case of Ewing sarcoma in the left fibula who underwent surgical resection and chemotherapy was referred to our department for pulmonary nodules and a metastatic workup. The discordant uptake of FDG and FAPI uptake in (RUL) and paratracheal lymph nodes required tissue sampling in order to exclude metastatic involvement. There was a paratracheal lymph node with SUVmax =8.42 (short-axis diameter (SAD) = 9 mm) and a hypermetabolic pulmonary nodule in the RUL with a diameter of 11 mm and SUVmax =6.57 without FAPI uptake. In addition, there is a 5 mm pulmonary nodule in the RML without FDG and FAPI uptake (Fig. 4).
Gastric cancer
In our two gastric cancer cases, we found that 68Ga-FAPI PET was superior to 18F-FDG-PET in detecting primary lesions and metastases in patients with gastric cancer at initial diagnosis and in detecting recurrence. 68Ga-FAPI PET led to the detection of more and/or larger lesions presenting with higher tracer uptake. The low background uptake of 68Ga-FAPI led to the identification of small metastatic lesions of gastric cancer in the peritoneum, abdominal lymph nodes, liver, and bones, which are more difficult to detect in an 18F-FDG PET scan (Fig. 7). In this regard, the 68Ga-FAPI PET scan demonstrated its value in visualizing primary and metastatic gastric cancer, while previous studies have shown that preoperative 18F-FDG PET has a low detection rate for primary gastric cancer [46, 47].
Glioblastoma Multiform (GBM)
While glucose-based PET is the standard for many entities, amino acid-based PET, such as Fluoroethyl-L-tyrosine-based (FET) PET, has shown advantages in glioma studies [48]. A joint practice guideline from the European Association of Nuclear Medicine (EANM), the Society of Nuclear Medicine and Molecular Imaging (SNMMI), the European Association of Neuro-oncology (EANO), and the Working Group for Response Assessment in Neuro-oncology with PET (PETRANO) highlighted the role of PET with radiolabeled amino acids in the diagnosis of gliomas and the distinction between true progression and treatment-related changes such as pseudoprogression or radionecrosis [49]. A head-to-head comparison of FET PET-based and FAP-specific PET-based target volume delineation could further assess FAP specificity in the context of radiation therapy planning for GBM. However, subjecting patients to both FET and FAP-specific PET, ideally with limited time between both procedures, would be challenging given the ethical implications, as no clear benefits are expected for the participants. Given the limited experience with the clinical use of FAP-specific PET in the treatment of cancer in general and GBM in particular, determining a threshold for an individual patient should be done by an experienced nuclear medicine physician [50]. Although amino acids and FDG tracers have been evaluated for glioma imaging, they showed limitations like increased gray matter tracer uptake or decreased uptake in treated areas of the brain [49]. A previous clinical pilot study with 68Ga-pentixafor in glioblastoma patients showed promising results with significant tumor uptake in 11 out of 13 glioblastoma patients [51]. However, the small sample size of this clinical study prompted us to plan measuring the level of expression of CXCR4 in glioblastoma tumor cells and healthy brain tissue in a larger population in a future study to assess the potential of 68Ga-Pentixafor and 177Lu-Pentixather as diagnostic and therapeutic tracers, respectively. In our study, a 55-year-old male with a history of GBM in the right frontotemporal lobes underwent surgical resection, radiotherapy, and 177Lu-DOTATATE for recurrence evaluation. The initial assessment showed that there were three lesions in the cerebral mass of the right frontal and right temporoparietal lobes. These include a lesion in the right frontal lobe with a size of 11×15 mm and SUVmax of 4.15 on 68Ga-Pentixafor and SUVmax of 4.28 on 68Ga-FAPI PET images, a lesion in the right frontal lobe (periventricular) with a size of 27×15 mm and SUVmax of 1.97 on 68Ga-Pentixafor and SUVmax of 2.68 on 68Ga-FAPI PET images (also spotted on MRI), a lesion in the right temporal lobe with a size of 12×14 mm and SUVmax of 2.23 on 68Ga-Pentixafor and SUVmax of 1.82 on 68Ga-FAPI PET images. Therefore, CXCR4 and FAPI avid lesions in the right frontal and right temporal lobes (three mass lesions) were more compatible with tumoral recurrence (Figs. 8 and 9).