Patient Characteristics
Between October 2019 and June 2022, 28 patients (median age, 52.5 years; interquartile range, 28–80 years) were included in this study. The flow diagram of the study design is presented in Figure 1. Among the 28 patients, 5 patients (18%) underwent PET/CT for definitive diagnosis of suspicious breast lesions on conventional imaging, 9 patients (32%) for initial staging, and 14 patients (50%) who previously underwent surgical resection underwent PET/CT for re-staging. Among the 14 treatment-naïve patients, 1 patient was confirmed as having a benign breast lesion (intraductal papilloma, confirmed via surgical resection), and 13 patients had invasive breast cancer (3 confirmed via surgical resection and 10 via breast biopsy). Regarding the other 14 patients with recurrent/progressive disease, at least one recurrent/metastatic lesion was pathologically confirmed as metastatic breast cancer (via percutaneous biopsy) from each patient. Based on the histological examination, the types of breast cancer included invasive ductal carcinoma (n=21), invasive lobular carcinoma (n=2), invasive micropapillary carcinoma (n=2), and mixed type (n=2). The detailed patient characteristics are summarized in Table 1.
Patient-based tumour detectability of [68Ga]Ga-FAPI and [18F]FDG PET/CT
Regarding the patient-based tumour detectability, although [68Ga]Ga-FAPI PET/CT seemed superior to [18F]FDG PET/CT in visualizing the primary tumour (92% [12/13] vs. 85% [11/13], p=1.000), local recurrence (100% [21/21] vs. 81% [17/21], p=0.125), neck LN metastases (100% [4/4] vs. 50% [2/4], p=0.500), abdomen LN metastases (100% [4/4] vs. 25% [1/4], p=0.250), bone metastases (100% [12/12] vs. 83% [10/12] ,p=0.500) and liver metastases (100% [6/6] vs. 50% [3/6], p=0.250), the difference indicated no statistical significance. The two imaging modalities were equivalent in detecting metastatic lesions in the mediastinal LNs (6/6 vs. 5/6), pelvic LNs (1/1 vs. 1/1), and lung (8/8 vs. 8/8).
Lesion-based tumour detectability of [68Ga]Ga-FAPI and [18F]FDG PET/CT in primary breast lesions
Based on imaging findings from breast US, [18F]FDG, and [68Ga]Ga-FAPI PET/CT, a total number of 22 breast lesions were detected in 13 treatment-naïve patients (among the 14 treatment-naïve patients, 1 was excluded since she was diagnosed with occult breast cancer, and the primary tumour was not radiologically detectable). Among them, 20 lesions from 12 patients were confirmed as malignant lesions, and 2 from another 2 patients were determined to be benign breast diseases (intraductal papilloma and atypical breast hyperplasia). There were no false-positive lesions on the US, [18F]FDG, and [68Ga]Ga-FAPI PET/CT. In addition, regarding the 13 patients with detectable primary tumours (4 patients with multifocal breast cancer), [68Ga]Ga-FAPI PET/CT was able to visualize all breast tumours (20/20), whereas [18F]FDG PET/CT could not detect 6 of them (14/20) (Table 2). The representative case is shown in Figure 2. Compared to [18F]FDG, [68Ga]Ga-FAPI PET/CT showed higher lesion detectability of primary breast tumours (100% vs. 70%, p=0.027) with higher radiotracer uptake (SUVmax, 12.1 [5.0-46.1] vs. 4.1 [0.9-17.3], p=0.001) and TBR (10.7 [4.0-41.9] vs. 3.8 [0.8-15.7], p=0.001) (Table 3).
Since all patients underwent the breast US as SCI, the sensitivity of breast US and [68Ga]Ga-FAPI PET/CT for breast cancer diagnosis was evaluated and compared. As expected, [68Ga]Ga-FAPI PET/CT showed higher sensitivity than breast US (100% [20/20] vs. 80% [16/20]) for the diagnosis of primary breast tumours. However, no statistical difference in diagnostic sensitivity was observed between the two imaging modalities (p=0.114) (Table 2). Not all treatment-naïve patients underwent breast mammography, so the sensitivity between [68Ga]Ga-FAPI PET/CT and mammography was not compared.
Lesion-based tumour detectability of [68Ga]Ga-FAPI and [18F]FDG PET/CT in local recurrence and regional/distant metastases
As for the detection of local recurrence and regional LN metastases,[68Ga]Ga-FAPI PET/CT revealed a greater number of positive lesions than [18F]FDG in detecting chest wall recurrent lesions (18 vs. 8), internal mammary LN metastases (10 vs. 3), axillary LN metastases (62 vs. 52), and clavicula LN metastases (38 vs. 25), with higher radiotracer uptake (SUVmax, 8.2 [0.9-38.0] vs. 3.1 [0.5-16.2], p<0.001) and TBR (7.4 [1.1-23.8] vs. 3.8 [0.9-21.2], p<0.001) than [18F]FDG PET/CT. The representative PET/CT images are shown in Figure 3.
Concerning distant LN metastases,[68Ga]Ga-FAPI PET/CT showed more positive lesions than [18F]FDG PET/CT for neck LNs (33 vs. 15), mediastinal LNs (42 vs. 26), and abdomen LNs (28 vs. 3), but did not exhibit obvious advantage over [18F]FDG in detecting metastatic lesions in the pelvic LNs (15 vs. 15). For detecting bone and visceral metastases, [68Ga]Ga-FAPI PET/CT revealed a greater number of positive lesions than [18F]FDG PET/CT for bone metastases (146 vs. 59) and liver metastases (28 vs. 11) (Table 3). Except for the lung metastases (70 vs. 70) and some unusual metastases (7 vs. 7), [68Ga]Ga-FAPI PET/CT showed significantly greater SUVmax and TBR than [18F]FDG in the LN metastases (SUVmax, 10.8 [4.2-19.3] vs. 3.1 [1.0-12.5], p<0.001; TBR, 6.3 [2.0-23.7] vs. 1.9 [0.6-9.5], p<0.001), bone metastases (SUVmax, 14.9 [5.5-40.7] vs. 4.8 [1.2-19.0], p<0.001; TBR, 13.7 [3.6-40.7] vs. 4.2 [0.8-8.6], p<0.001), and liver metastases (SUVmax, 6.0 [3.6-11.3] vs. 3.7 [2.5-12.2], p=0.033; TBR, 2.3[1.5-11.3] vs. 1.2[1.0-2.9], p<0.001) (Table 3).
Changes in breast cancer staging, re-staging, and clinical management after [68Ga]Ga-FAPI PET/CT
Regarding the initial assessment of 13 patients, compared with [18F]FDG PET/CT, [68Ga]Ga-FAPI PET/CT upgraded the N and M stage in 4 (4/13, 31%) and 2 patients (2/13, 15%), respectively (Table 4). [68Ga]Ga-FAPI PET/CT demonstrated a greater number of metastatic lesions and a more significant disease extent in 8 patients (8/13, 62%). With the new lesions detected by [68Ga]Ga-FAPI PET/CT, TNM staging was eventually upstaged in four patients (4/13, 31%); two from IIIC to IV, one from IIIA to IIIC, and one from IA to IIA. Consequently, the therapeutic regimen was changed for two patients (2/13, 15%; from surgically resectable to unresectable) due to the newly detected liver, bone, and LN metastases (Table 4). In contrast, [18F]FDG PET/CT detected a greater number of bone, liver, and metastases than [68Ga]Ga-FAPI in only one patient. Compared with SCI, [68Ga]Ga-FAPI PET/CT upgraded the N stage in seven patients (7/13, 54%) and the M stage in three patients (3/13, 23%). Consequently, TNM staging was finally upstaged in five patients (5/13, 38%), and the therapeutic regimen was changed in three patients (3/13, 23%), including two from surgically resectable to unresectable.
Among another 14 patients in whom recurrence or/and metastasis was detected, [68Ga]Ga-FAPI PET/CT detected [18F]FDG–negative local recurrent tumours in 4 patients (4/14, 29%), and [18F]FDG–negative metastatic lesions (cervical LN and bone metastases) in 1 patient (1/14, 7%). Moreover, [68Ga]Ga-FAPI PET/CT detected a greater number of metastatic lesions and a more significant disease extent than [18F]FDG in 12 patients (12/14, 86%). Based on additional findings on [68Ga]Ga-FAPI PET/CT, tumour re-staging was upgraded in three patients (3/14, 21%), and their clinical management was subsequently changed (one from routine follow-up to systemic treatment, the other two from routine follow-up to chest wall radiotherapy) (Table 4). Compared to SCI, [68Ga]Ga-FAPI PET/CT detected SCI–negative local recurrent tumours in eight patients (8/14, 57%) and SCI–negative metastatic lesions in two patients (2/14, 14%). Consequently, tumour re-staging was upstaged in four patients (4/14, 29%), and their therapeutic regimen was subsequently changed (two from routine follow-up to systemic treatment, the two from routine follow-up to chest wall radiotherapy).
Relationship between radiotracer uptake and clinicopathologic characteristics in breast and metastatic lesions
Based on the analysis of primary breast tumours, 20 breast cancers were included from 12 treatment-naïve patients. There were no significant differences regarding the FDG/FAPI SUVmax among the four receptor statuses (HR-positive, HR-positive/HER2-positive, HER2-positive, and TNBC), and no difference regarding the SUVmax among different tumour sizes (<30 mm, 30-45 mm, and >45 mm) was observed. All p values are shown in Figure 6, and further details are presented in Supplementary Table 1.
Twenty-five histologically confirmed metastatic lesions were obtained from 20 re-staging patients to evaluate the relationship between radiotracer uptake and clinicopathologic characteristics in patients with recurrent/progressive disease (Supplementary Table 1). However, no significant difference regarding FAPI/FDG SUVmax among the four receptor statuses (HR-positive, HR-positive/HER2-positive, HER2-positive, and TNBC) and pathological subtypes (ductal, lobular, micropapillary, and the special type) was observed. All p values are shown in Figure 7.
Regarding the tumour size from metastatic lesions, [18F]FDG-SUVmax from large (>20 mm) and medium-sized (11-20 mm) were significantly higher than those from the small tumours (5-10 mm) (SUVmax, 7.1 [3.9-10.3] vs. 1.9 [0.5-3.6], p=0.001, 4.9 [1.2-8.8] vs 1.9 [0.5-3.6], p=0.042, respectively). As for the metastatic lesions on [68Ga]Ga-FAPI PET/CT, a significant difference regarding SUVmax was observed between the medium-sized tumours and small tumours (SUVmax, 17.7 [5.7-43.2] vs. 5.1 [1.6-7.8], p=0.02) (Figure 7).