Cortical abnormalities of synaptic vesicle protein 2A in focal cortical dysplasia type II identified in vivo with ¹⁸F-SynVesT-1 positron emission tomography imaging

Focal cortical dysplasia type II (FCD II) constitutes the most common cause of seizures in patients who undergo surgery before the age of 18 years [1]. Epilepsy in FCD II is commonly pharmacoresistant and thus particularly challenging for antiepileptic treatment [2]. Surgical resection of FCD II lesions may prevent seizures and improve quality of life [3]. It has been well established that the main predictor of favorable surgical outcomes is the complete removal of the dysplastic cortex.

FCD II is predominantly located in extratemporal areas, in particular the eloquent cortex [3]. Magnetic resonance imaging (MRI) features in FCD II have been widely described [4], but so-called negative MRI has been reported in 17–34% of patients and is associated with poor surgical outcomes [5]. Positron emission tomography (PET) imaging with ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) has significantly improved the positive detection rate of lesions. However, maximal hypometabolic areas correspond to both the lesion and seizure onset zone [6]. The accuracy of ¹⁸F-FDG PET in identifying FCD II is limited by hypometabolism frequently extending beyond the lesion. Therefore, for FCD II patients, PET imaging with additional radioligands that can be used to guide more accurate demarcation of the lesion would be of great clinical value.

Observations from a rat model of epilepsy and dysplastic cortical tissue suggested that the loss of synaptic vesicle glycoprotein 2A (SV2A) may lead to alterations in neurotransmission [7]. SV2A loss can cause impairments in γ-aminobutyric acid (GABA)ergic function [8‐10]. ¹¹C-UCB-J, a specific radioligand for SV2A, has been used in the investigation of several neuropsychiatric diseases [10‐13]. Compared to ¹¹C-UCB-J, the newly reported SV2A radioligand ¹⁸F-SynVesT-1 has a longer half-life and superior signal-to-noise ratio [14, 15]. In a preliminary study using static ¹⁸F-SynVesT-1 PET, we demonstrated lower SV2A levels in the epileptogenic zone (EZ) of patients with FCD II [16]. In the present study, we included more FCD II patients with neuropathology data and controls. The FCD II patients were also evaluated with ¹⁸F-FDG PET and high-resolution MRI to allow for direct comparisons.

In the present study, we detected, for the first time, significantly lower SV2A binding in brain lesions of FCD II patients under the age of 18 using in vivo PET imaging with the novel SV2A radioligand ¹⁸F-SynVesT-1. Using visual and semiquantitative analyses of PET images coregistered with MRI images, we assessed the utility of ¹⁸F-SynVesT-1 compared with ¹⁸F-FDG PET for lesion localization in FCD II. Our results indicated that static ¹⁸F-SynVesT-1 PET provided high-quality images and that visual and semiquantitative analyses without arterial blood sampling enabled the identification of lesions in FCD II patients with accuracy comparable to the current standard static ¹⁸F-FDG PET. Furthermore, we demonstrated that the range of low SV2A binding in the lesions was more restricted than that of hypometabolism. The corresponding asymmetry in lesion uptake was highly correlated between the two radiotracers across subjects.

Lower SV2A in epileptic lesions has been repeatedly observed in brain tissue slices from animal models of epilepsy and epileptic patients [7, 8, 27]. In this study, we found an AI based on lesion SUVr of 27.14% ± 10.15% in patients with FCD II, which was much higher than that in the controls. This magnitude of SV2A reduction is in line with the reported 28% decrease for SV2A in FCD lesions measured in vitro with tissues resected from intractable epilepsy patients [10, 20]. Studies have suggested that the loss of SV2A may contribute to instability in neuronal networks [28]. Electrophysiological studies of spontaneous inhibitory neurotransmission revealed that the loss of SV2A led to an imbalance between glutamatergic and GABAergic neurotransmission [8, 29], thus implicating SV2A as a sensitive marker of FCD II, and PET imaging with SV2A-specific tracers as a potentially valuable tool for lesion identification/localization in this disorder. The comparison of lesion AI with ¹⁸F-SynVesT-1 SUVr and ¹⁸F-FDG SUVr indicated a positive correlation. One study in temporal lobe epilepsy also found that hippocampal ¹⁸F-FDG uptake was correlated with ¹¹C-UCB-J binding potential [10]. These findings suggest that the hypometabolism in FCD II lesions may be related to the decrease in SV2A, and SV2A PET provides a complementary measure of the epileptogenic substrate.

We found that lesions in 2 FCD type IIa patients could not be detected by the initial visual assessment of ¹⁸F-SynVesT-1 PET images. In one patient, the lesion could be located after coregistration with MRI (and there was a small AI, 11.00%, in this patient). We assumed that FCD IIa might have resulted in a nonsignificant reduction in SV2A that was not clearly visible in the ¹⁸F-SynVesT-1 PET images but was associated with the pathological findings. Although cytological differences have indicated biological differences between the 2 FCD subtypes, no associated differences in clinical or imaging findings have been consistently identified. A previous study in FCD IIa showed decreased neuropil expression for SV2A, but strong perikaryal SV2A immunoreactivity has been observed around cytomegalic neurons [30]. The cytomegalic neuron expressed strong SV2A, which could obscure the originally decreased neuropil expression of FCD, and PET cannot differentiate these changes at the cellular level.

The detection of FCD II by MRI remains a challenge, despite the use of best-practice MRI sequences, new high-field MRI, and postprocessing software [5]. In the present study, MRI was negative in 56.2% of patients. ¹⁸F-FDG PET was reported to have a high sensitivity to detect FCD II in 60–92% of patients [6], and our ¹⁸F-FDG PET data were concordant with the findings from these contemporary studies. We found the visual assessments of SV2A PET images to have 87.5% accuracy in lesion localization, and there were no statistically significant differences between ¹⁸F-FDG and ¹⁸F-SynVesT-1 PET in the accuracy of lesion location (93.8% vs. 87.5%, respectively) and AI (27.14% ± 10.11% vs. 27.08% ± 9.89%). ¹⁸F-SynVesT-1 PET coregistered with MRI increased detection accuracy and helped delineate FCD II in MRI-negative/doubtful patients, with 93.8% accuracy in FCD detection.

More importantly, we demonstrated that ¹⁸F-SynVesT-1 PET images showed a more restricted pattern of reduced SV2A than the region of hypometabolism in FCD II patients. These findings may have practical importance, as in these cases, seizure freedom may be obtained by smaller resections that are precisely limited to the lesion. Hence, ¹⁸F-SynVesT-1 PET might be helpful to guide the placement of intracranial electrodes and presurgical planning. We assume that SEEG may remain indicated to limit the extent of resection, which is an important consideration when planning surgical resection in the eloquent cortex. However, such an invasive procedure may be progressively discontinued once multimodal imaging becomes capable of delineating the dysplastic cortex. Whether the more restricted pattern of reduced SV2A can become a real advantage for surgical planning of FCD patients requires more clinical research and data analysis in the future.

¹⁸F-SynVesT-1 has been shown to have excellent kinetic and in vivo binding properties in the human brain. SUVr from 60–90 min postinjection provided a good match with the one-tissue compartment binding potential of ¹⁸F-SynVesT-1 and can serve as a surrogate quantitative measurement of specific binding in a short scan time without invasive arterial sampling [15]. As it is more difficult for younger patients with FCD to maintain the same posture for a long time, the longer half-life of ¹⁸F coupled with a quantitative measurement in a short scan time with ¹⁸F-SynVesT-1 will facilitate its broad application in studies of SV2A in FCD and other neuropsychiatric populations.

There are several limitations that warrant mentioning. First, our sample size was modest, although it is well within the range for a PET study evaluating a new radiotracer in a new patient group such as FCD patients [6]. The limited sample size may have obscured possible correlations with clinical measures. Second, we used the CS as a reference region to adjust for nonspecific uptake in the brain. However, it may not be the optimal reference area because its tissue composition is different from gray matter, and the use of the CS as a reference region might lead to underestimation of the SUVr [31]. Last, the age difference between the patient and control groups might have had an effect on the study findings. We analyzed the SUVr of nonpathological brain regions for all participants to observe the trend in synaptic density changes (Supplemental Fig. S3), which was consistent with results from previous molecular biology studies on synaptic pruning and synaptic density changes with age [32, 33]. The changes in myelination with age is another factor that needs to be considered [34]. Postmortem and neuroimaging studies have suggested a quadratic relationship between myelination status and age, which might impact synaptic changes [35, 36]. Due to the age difference between the two groups, objective analyses, such as statistical parametric mapping (SPM), could not be conducted. As a result, we performed only a relatively subjective visual assessment and semiquantitative analysis/comparison between the child patient group and young adult control group. Further analysis will be conducted in future in-depth studies as we continue our investigation in this patient population.

To the best of our knowledge, this is the first in vivo study to investigate SV2A in the lesions of living people with FCD II by PET imaging with the radioligand ¹⁸F-SynVesT-1. ¹⁸F-SynVesT-1 PET demonstrated a higher accuracy than MRI for the localization of FCD II lesions, with a more restricted pattern of SV2A abnormality than that of hypometabolism detected by ¹⁸F-FDG PET. In conclusion, SV2A PET imaging may provide a more specific localization of lesions in FCD II, and in presurgical evaluation and planning, it can serve as a complementary measure of the epileptogenic substrate in addition to the established clinical assessments.