18F-FSPG PET/CT Imaging of System xC- Transporter Activity in Patients with Primary and Metastatic Brain Tumors

Abstract

Background The PET tracer (4S)-4-(3-[¹⁸F]fluoropropyl)-l-glutamate (¹⁸F-FSPG) targets the system x_C^- cotransporter, which is overexpressed in various tumors. Purpose To assess the role of ¹⁸F-FSPG PET/CT in intracranial malignancies. Materials and Methods Twenty-six patients (mean age, 54 years ± 12; 17 men; 48 total lesions) with primary brain tumors (n = 17) or brain metastases (n = 9) were enrolled in this prospective, single-center study (ClinicalTrials.gov identifier: NCT02370563) between November 2014 and March 2016. A 30-minute dynamic brain ¹⁸F-FSPG PET/CT scan and a static whole-body (WB) ¹⁸F-FSPG PET/CT scan at 60-75 minutes were acquired. Moreover, all participants underwent MRI, and four participants underwent fluorine 18 (¹⁸F) fluorodeoxyglucose (FDG) PET imaging. PET parameters and their relative changes were obtained for all lesions. Kinetic modeling was used to estimate the ¹⁸F-FSPG tumor rate constants using the dynamic and dynamic plus WB PET data. Imaging parameters were correlated to lesion outcomes, as determined with follow-up MRI and/or pathologic examination. The Mann-Whitney U test or Student t test was used for group mean comparisons. Receiver operating characteristic curve analysis was used for performance comparison of different decision measures. Results ¹⁸F-FSPG PET/CT helped identify all 48 brain lesions. The mean tumor-to-background ratio (TBR) on the whole-brain PET images at the WB time point was 26.6 ± 24.9 (range: 2.6-150.3). When ¹⁸F-FDG PET was performed, ¹⁸F-FSPG permitted visualization of non-¹⁸F-FDG-avid lesions or allowed better lesion differentiation from surrounding tissues. In participants with primary brain tumors, the predictive accuracy of the relative changes in influx rate constant K_i and maximum standardized uptake value to discriminate between poor and good lesion outcomes were 89% and 81%, respectively. There were significant differences in the ¹⁸F-FSPG uptake curves of lesions with good versus poor outcomes in the primary brain tumor group (P < .05) but not in the brain metastases group. Conclusion PET/CT imaging with (4S)-4-(3-[¹⁸F]fluoropropyl)-l-glutamate (¹⁸F-FSPG) helped detect primary brain tumors and brain metastases with a high tumor-to-background ratio. Relative changes in ¹⁸F-FSPG uptake with multi-time-point PET appear to be helpful in predicting lesion outcomes. Clinical trial registration no. NCT02370563 © RSNA, 2022 Online supplemental material is available for this article.

Graphical abstract

Figure 1:

(A) Images in a 63-year-old woman (participant 13) with right temporoparietal glioblastoma multiforme treated with surgery, followed by chemotherapy (temozolomide) and radiation therapy. MRI findings are equivocal for residual disease versus posttreatment change. Static (4S)-4-(3-[¹⁸F]fluoropropyl)-l-glutamate (¹⁸F-FSPG) PET image (left) shows intense uptake in the tumor lesion (arrow), which is visible on the postcontrast three-dimensional spoiled gradient-echo MRI scan (arrow, middle). The tumor time-activity curve (TAC) from the dynamic ¹⁸F-FSPG PET images (right) shows a rapidly increasing uptake over time in the glioblastoma multiforme lesion with an early peak uptake in the first 10 frames. In the background tissue TAC, there is an early peak in the first 10 frames followed by a downward slope and an eventual plateau near zero activity in the final 28 frames. Subsequent pathologic examination confirmed the presence of malignant cells. The outcome score for this lesion was a 5. SUV_max = maximum standardized uptake value. (B) Images in a 44-year-old man (participant 4) with glioblastoma multiforme in the right temporal lobe treated with surgery, followed by chemotherapy and repeat resection showing no tumor. MRI scan (middle) shows an area of contrast enhancement concerning for progression (arrow). Static ¹⁸F-FSPG PET image (left) shows weak uptake (arrow). The lesion TAC from the dynamic ¹⁸F-FSPG PET images (right) shows a plateau line after an early uptake peak, and higher uptake values in the tumor lesion at late times when compared to background, with a slight and slow tendency to increase over time. Subsequent pathologic examination showed necrosis and reactive changes with no definite glial neoplasm. The outcome score for this lesion was a 1. (C) Images in a 77-year-old man (participant 2) with grade III unresectable anaplastic astrocytoma of the temporal lobe (postradiation therapy). Baseline MRI scan (middle) shows a lesion with two separate components, one medial (lesion 1, red arrow) and one lateral (lesion 2, green arrow). Static ¹⁸F-FSPG PET image (left) shows prominent uptake medially (SUV_max = 12), and weak uptake laterally (lesion SUV_max = 3 on whole body [WB] PET). Follow-up MRI scan (middle) obtained 4 months after baseline shows an increase in size of the medial component, while the lateral component remains stable. In the background volume of interest, the TAC on the dynamic ¹⁸F-FSPG PET scan (right) shows an exponential decaying behavior after an early uptake peak. A similar TAC pattern is seen in the lateral lesion, but with higher uptake values at late times when compared to the background volume of interest, with a slight and slow tendency to increase over time. The medial lesion shows rapidly increasing uptake over time throughout the whole duration of the dynamic PET acquisition. The slight tendency of ¹⁸F-FSPG uptake to increase in the lateral (nonenhancing) component of the tumor, and the mild uptake on the WB late scan (SUV_max = 3), suggests a “wait and watch” approach would be best for this lesion. Mild uptake indicates uptake slightly higher than surrounding background (healthy tissue). The outcome score for both lesions in this patient was a 4.

Figure 2:

Images in a 65-year-old woman (participant 3) with non–small cell lung cancer brain metastases. (A) Contrast-enhanced MRI scan shows a small right frontal lesion (lesion 1, green arrow) and also an enhancing lesion within the right central sulcus (lesion 2, red arrow), which may represent evolving treatment effects or progression. (B) Fluorine 18 fluorodeoxyglucose PET image shows no uptake in either lesion, whereas (C) (4S)-4-(3-[¹⁸F]fluoropropyl)-l-glutamate (¹⁸F-FSPG) PET image shows mild uptake in the smaller anterior lesion and intense uptake in the bigger posterior lesion. (D) Tissue time-activity curves from the dynamic ¹⁸F-FSPG PET images show a plateau line after an early uptake peak in the background volume of interest, and a similar pattern in the small right frontal lesion (lesion 1), whereas the larger lesion in the right central sulcus (lesion 2) shows rapidly increasing uptake over time throughout the whole duration of the dynamic acquisition. The outcome score was 1 for lesion 1 and 2 for lesion 2. SUV = standardized uptake value, SUV_max = maximum SUV.

Figure 3:

Compartmental model–fitted curves for the grade 4 primary brain tumor lesions (11 participants, 16 lesions), the non–grade 4 primary brain tumor lesions (six participants, 11 lesions), and the brain metastases (nine participants, 21 lesions). Error bars denote the standard error of the mean. C₁, C₂, and C₃ are the tracer concentrations in each tissue compartment. C_p is the tracer concentration in the plasma or blood compartment. K₁ = rate coefficient for transfer of tracer into the extracellular space from arterial plasma or blood compartment, k₂ = rate constant for transfer of tracer out of the extracellular space and into arterial plasma or blood compartment, k₃ = rate constant for transfer of tracer into the cell via system x_C^– from the extracellular space, k₄ = rate constant for transfer of tracer out of the cell and into the extracellular space, k₅ = sequestration rate constant of tracer into amino acid pools within the cell, K_i = net influx rate constant of tracer uptake into tissue, SUV = standardized uptake value, TAC = time-activity curve.

Figure 4:

Compartmental model–fitted curves for grade 4 primary brain tumor lesions as stratified by lesion outcome (determined with follow-up MRI and/or pathologic examination). There were 11 grade 4 lesions in the poor lesion outcome group and five grade 4 lesions in the good lesion outcome group. Error bars denote the standard error of the mean. K₁ = rate coefficient for transfer of tracer into the extracellular space from arterial plasma or blood compartment, k₂ = rate constant for transfer of tracer out of the extracellular space and into arterial plasma or blood compartment, k₃ = rate constant for transfer of tracer into the cell via system x_C^– from the extracellular space, k₄ = rate constant for transfer of tracer out of the cell and into the extracellular space, k₅ = sequestration rate constant of tracer into amino acid pools within the cell, K_i = net influx rate constant of tracer uptake into tissue, SUV = standardized uptake value, TAC = tumor-activity curve.

Figure 5:

Box-and-whisker plots show the percentage change in (4S)-4-(3-[¹⁸F]fluoropropyl)-l-glutamate (¹⁸F-FSPG) influx rate constant (K_i) and maximum standardized uptake (SUV_max) for grade 4–only lesions (A and C) and grade 4 and non–grade 4 lesions pooled together (B and D). A plus sign (+) denotes the mean value in the box plots. The Student t test was used for group mean comparisons.

Figure 6:

Receiver operating characteristic (ROC) curve analysis is comparing the decision measures (A, C) ΔKi and (B, D) ΔSUV_max in the grade 4–only primary brain tumors and the entire primary brain tumor group (grade 4 and non–grade 4 brain lesions pooled together). AUC = area under the ROC curve, ΔK_i = percentage change between the K_i value estimated using the 30-minute dynamic PET data and the K_i value estimated using the dynamic plus whole-body PET data. ΔSUV_max = percentage change in maximum standardized uptake value of the brain lesion between the dynamic PET scan (last frame) and whole-body PET scan.