Heterogeneity of fibroblast activation protein expression in the microenvironment of an intracranial tumor cohort: head-to-head comparison of gallium-68 FAP inhibitor-04 (68Ga-FAPi-04) and fluoride-18 fluoroethyl-L-tyrosine (18F-FET) in positron emission tomography-computed tomography imaging.

Background: Cancer-associated fibroblasts (CAFs) within the tumor microenvironment (TME) can interact with tumor parenchymal cells to promote tumor growth and migration. Fibroblast activation protein (FAP) expressed by CAFs can be targeted with positron emission tomography (PET) tracers, but studies on FAP expression patterns in intracranial tumors remain scarce. We aimed to evaluate FAP expression patterns in intracranial tumors with gallium-68 FAP inhibitor-04 (⁶⁸Ga-FAPi-04) and immunohistochemical staining and to observe the interactions between CAFs and tumor cells with a head-to-head comparison of ⁶⁸Ga-FAPi-04 and fluoride-18 fluoroethyl-L-tyrosine (¹⁸F-FET) for PET quantification analysis.

Methods: We prospectively enrolled 22 adult patients with intracranial mass lesions. ⁶⁸Ga-FAPi-04 and ¹⁸F-FET PET-computed tomography (PET/CT) brain imaging were applied before surgery. Maximal tumor-to-brain ratio (TBRmax), metabolic tumor volume (MTV), and total lesion tracer uptake (TLU) was obtained, and different thresholds were used for ⁶⁸Ga-FAPi-04-positive lesion delineation owing to the lack of relevant guidelines. The MTV and TLU ratios of both tracers were calculated. Linear regression was applied to observe the differential efficacy of semiquantitative PET parameters.

Results: A total of 22 patients with a mean age of 50±13 years (range, 27-69 years) were enrolled. Heterogeneous patterns of ⁶⁸Ga-FAPi-04 uptake [median of maximal standardized uptake value (SUVmax) =3.8; range, 0.1-19.1] were found. More malignant tumors, including brain metastasis, glioblastoma, and medulloblastoma, generally exhibited more significant ⁶⁸Ga-FAPi-04 uptake than did the less malignant tumors, while the SUVmax and TBRmax exhibited nonsignificant differences across three intracranial lesion groups of primary brain tumor, brain metastasis, and noncancerous disease (SUVmax: P=0.092; TBRmax: P=0.189). Immunohistochemistry staining showed different stromal FAP expression status in various intracranial lesions. In 15 patients with positive ⁶⁸Ga-FAPi-04 intracranial tumor uptake, the MTV_FAPi:MTV_FET ratio had differential efficacy in various types of intracranial tumors [95% confidence interval (CI): 0.572-7.712; P=0.027], and further quantification analyses confirmed the differential ability of the MTV_FAPi:MTV_FET ratio (95% CI: -0.045 to 11.013, P=0.052; 95% CI: 0.044-17.903, P=0.049; 95% CI: -1.131 to 30.596, P=0.065) with different isocontour volumetric thresholds.

Conclusions: This head-to-head study demonstrated heterogeneous FAP expression in intracranial tumors. The FAP expression volume percentage in tumor parenchyma may therefore offer benefit with respect to differentiating between intracranial tumor types.