Preoperative evaluation of drug-resistant epilepsy increasingly relies on translocator protein positron emission tomography, a molecular imaging technique that identifies neuroinflammation within seizure foci. Approximately one-third of individuals diagnosed with epilepsy do not achieve seizure control despite medical therapy, leaving surgical resection as the primary curative alternative. For surgical intervention to succeed, clinicians must precisely determine the anatomical boundaries of the epileptogenic focus. Traditional neuroimaging modalities, most notably magnetic resonance imaging and fluorodeoxyglucose positron emission tomography, frequently prove inadequate for detecting subtle structural lesions or yield broad, nonlocalizing abnormalities. By targeting the heightened expression of the eighteen-kilodalton translocator protein on the outer mitochondrial membrane of activated glial cells, this imaging modality achieves superior detection rates for epileptogenic zones compared to conventional scans. Clinical investigations indicate that this targeted approach successfully pinpoints epileptogenic regions in patients whose magnetic resonance imaging scans appear completely normal. Consequently, this imaging modality represents a vital diagnostic advancement for surgical planning.

Under baseline physiological conditions, translocator protein displays a strictly limited distribution throughout the central nervous system, showing minimal presence in most brain regions while remaining restricted to the choroid plexus, ependyma, and olfactory bulb. Under pathological conditions, however, glial activation markedly increases translocator protein concentrations in injured tissues. In temporal lobe epilepsy, surgical tissue analyses reveal pronounced elevations of translocator protein within the hippocampus, specifically localized to sclerotic tissue ipsilateral to the seizure focus. Experimental rodent models of epilepsy, including status epilepticus induced by kainic acid or pilocarpine, confirm these observations, showing elevated glial protein expression across the hippocampus, thalamus, hypothalamus, amygdala, and cortex. In these models, protein upregulation emerges rapidly following seizure onset, peaks within two weeks, and persists for weeks or months. This prolonged elevation is triggered by neuronal injury and the release of damage-associated molecular patterns, which stimulate surrounding microglia and astrocytes, establishing translocator protein as an enduring biomarker of chronic neuroinflammation.

Within epileptic brain regions, resident microglia transition from a resting state into a pro-inflammatory phenotype, releasing cytokines and chemokines that perpetuate tissue damage and exacerbate underlying epilepsy pathologies. Previous mechanistic investigations in rodent models established that seizure-related stimuli activate an intracellular pathway involving protein kinase C epsilon, which sequentially engages Raf-1, MEK1/2, and ERK1/2. This signaling cascade leads to the phosphorylation of transcription factors STAT3 and c-Jun, which form activating protein 1 complexes that bind the translocator protein promoter to stimulate transcription. Importantly, recent translational research demonstrates that this transcriptional upregulation does not operate identically in human microglia. Human microglia and monocytes lack the functional activating protein 1 binding sites essential for this transcriptional cascade. Consequently, inflammatory stimulation in human microglia does not raise translocator protein abundance per cell. Rather, elevated clinical positron emission tomography signals in seizure foci result from the physical accumulation of activated microglia within the epileptogenic focus, alongside secondary contributions from reactive astrocytes and occasionally neurons.

In addition to serving as a diagnostic biomarker, translocator protein regulates several vital cellular functions, including steroidogenesis, heme transport, energy metabolism, apoptosis, and mitophagy. In neurosteroid synthesis, the protein is thought to facilitate mitochondrial cholesterol transport through a conserved recognition consensus motif, although genetic knockout experiments in microglia have debated its absolute necessity. Translocator protein also facilitates mitochondrial heme export via specific conserved amino acid motifs, supporting NADPH oxidase complex assembly and stimulating Nrf2-mediated antioxidant transcription to counteract neuroinflammatory oxidative stress. Deletion or functional impairment of the protein triggers intracellular porphyrin accumulation and heightened mitochondrial reactive oxygen species production, mechanisms intrinsically linked to ferroptosis. Furthermore, translocator protein associates structurally with voltage-dependent anion channel 1 within the mitochondrial permeability transition pore complex. Its overexpression can inhibit mitophagy by impairing sequestosome 1 recruitment and channel ubiquitination, whereas targeted synthetic ligands can restore mitophagy. Although direct evidence in epilepsy remains limited, these mitochondrial regulatory mechanisms likely underpin the close association between protein upregulation and chronic epileptic pathology.

Over the past three decades, the development of radiotracers targeting the translocator protein has evolved across three successive generations to resolve persistent pharmacokinetic and technical obstacles. The initial first-generation compound, carbon-11-labeled PK11195, validated the fundamental feasibility of imaging neuroinflammation in epilepsy, successfully visualizing focus-associated inflammatory changes in temporal structures in animal models and clinical cohorts. Nevertheless, several significant limitations restricted its broader clinical utility. The radiotracer possesses suboptimal blood-brain barrier permeability and excessive lipophilicity, reflected in a logD value of 3.97 that exceeds the optimal target range of approximately 2.2. This high lipophilicity causes pronounced nonspecific binding to peripheral blood cells and plasma proteins, resulting in poor image contrast. Quantitative assessments demonstrate core-to-contralateral uptake ratios of 2.43 plus or minus 0.39 for the first-generation compound. Furthermore, the carbon-11 radioisotope exhibits a short physical half-life of roughly twenty minutes, requiring an on-site cyclotron and restricting its application exclusively to specialized tertiary academic medical centers.

To overcome the modest signal-to-noise ratios of first-generation tracers, second-generation radioligands were introduced, including carbon-11-labeled PBR28, DAC, and DPA-713, alongside fluorine-18-labeled agents such as PBR11, PBR111, and DPA-714. Utilizing fluorine-18 extended radiotracer half-life to approximately 110 minutes, facilitating centralized manufacturing and distribution. These second-generation tracers improved imaging performance substantially, providing brain uptake ratios four to six times higher than carbon-11 PK11195 and core-to-contralateral ratios of 3.41 plus or minus 1.09. However, the routine deployment of second-generation tracers encountered a severe genetic obstacle caused by the rs6971 single nucleotide polymorphism in the translocator protein gene. This genetic variation induces an alanine-to-threonine substitution at position 147, which dramatically decreases tracer binding affinity. Individuals with the low-affinity binder genotype display markedly reduced tracer uptake, with binding ratios between high-affinity and low-affinity binders varying from 4 to 55.3. Because this genotype occurs in two to thirty percent of populations globally, second-generation agents necessitate prior genetic screening, significantly hindering routine clinical workflows.

Third-generation radiotracers were developed to overcome the genetic sensitivity of earlier compounds while preserving elevated signal-to-noise ratios and high sensitivity. Novel fluorine-18 radiotracers such as GE-38, LW223, and BS224 demonstrate uniform binding affinity across genetic profiles, maintaining low-affinity to high-affinity binder ratios between 1.0 and 1.8, comparable to first-generation PK11195. By eliminating differential binding caused by the rs6971 polymorphism, these modern tracers eliminate the necessity for pre-scan genetic screening. Demonstrating practical utility, longitudinal studies utilizing the third-generation tracer fluorine-18 GE-180 in a rat model of temporal lobe epilepsy revealed that early uptake increases in the hippocampus and amygdala accurately predicted subsequent spontaneous seizure development. Several third-generation compounds are currently undergoing preclinical evaluation and human clinical trials, establishing them as strong candidates to replace earlier generations. Their genotype-independent binding profile and favorable imaging kinetics are expected to accelerate the routine clinical implementation of neuroinflammatory positron emission tomography for drug-resistant epilepsy.

In clinical settings, translocator protein imaging has demonstrated superior detection rates for epileptogenic foci compared to structural magnetic resonance imaging, particularly in challenging diagnostic scenarios such as extratemporal epilepsy and magnetic resonance imaging-negative presentations. In an evaluation conducted by Cheval and colleagues examining fifteen patients with drug-resistant epilepsy and completely normal structural imaging, translocator protein imaging successfully detected signal abnormalities in one hundred percent of cases, with surgical outcomes subsequently confirming focus localization. Similarly, Qiao and colleagues investigated a clinical cohort with histopathologically verified focal cortical dysplasia and established that translocator protein positron emission tomography correctly localized seizure foci in all thirteen patients who exhibited negative magnetic resonance imaging results. The imaging findings correlated completely with invasive stereoelectroencephalography recordings and postsurgical histopathology. Combining translocator protein imaging with magnetic resonance imaging further enhances diagnostic confidence and focus detection rates, offering a complementary modality that refines presurgical hypotheses when initial clinical evaluations remain inconclusive.

Translocator protein imaging also surpasses metabolic fluorodeoxyglucose positron emission tomography in detection rate and anatomical precision when localizing epileptogenic foci. In patients with extratemporal lobe epilepsy and normal magnetic resonance imaging, the positive detection rate of fluorodeoxyglucose positron emission tomography is limited to approximately thirty percent. Furthermore, fluorodeoxyglucose imaging frequently reveals widespread hypometabolic zones extending beyond the true seizure focus, introducing substantial risks of surgical mislocalization. In contrast, translocator protein signals remain tightly confined to the epileptogenic focus. In a clinical investigation by Dickstein and colleagues involving eleven patients with drug-resistant epilepsy, seven exhibited normal fluorodeoxyglucose findings; among these seven metabolic non-responders, translocator protein imaging yielded positive localizations in five patients, with three undergoing resection confirmed as focal cortical dysplasia. In a related study by Cheval and colleagues, translocator protein imaging provided novel localization details that refined surgical margins in 65.2 percent of patients, compared to only 17.4 percent for metabolic imaging.

Achieving reliable imaging outcomes requires careful management of scan timing, disease severity, and biological limitations. Research tracking post-seizure dynamics indicates that tracer uptake begins rising within five to ten days after a seizure, reaches peak intensity at approximately one week, and gradually declines over four to twelve weeks, indicating that the optimal window for scanning occurs roughly one week postictally. Disease severity also influences imaging intensity, as patients with higher Global Assessment of Severity of Epilepsy scores, increased seizure frequency, or ipsilateral hippocampal sclerosis demonstrate greater radiotracer uptake. Despite these advantages, notable limitations persist. Elevated translocator protein expression reflects general neuroinflammation rather than epilepsy-specific pathology, meaning that non-epileptic conditions such as trauma, reactive gliosis, or signal spillover from the adjacent choroid plexus can yield false-positive signals outside the epileptogenic focus. Addressing these limitations through multimodal imaging combining magnetic resonance imaging and single-photon emission computed tomography, alongside standardized acquisition protocols, will be crucial to integrating translocator protein imaging into standard surgical epilepsy management.

Subject of Research: Preoperative localization of seizure foci using TSPO-PET in drug-resistant epilepsy

Article Title: Refining seizure foci localization: the potential of TSPO-PET

Article References: Wang, Y., Chen, Y., Xu, S., & Wu, X. (2025). Refining seizure foci localization: the potential of TSPO-PET. Acta Epileptologica, 7(1), Article 41. https://doi.org/10.1186/s42494-025-00234-2

Image Credits: AI Generated

DOI: 10.1186/s42494-025-00234-2

Keywords: TSPO-PET, drug-resistant epilepsy, seizure foci, neuroinflammation, positron emission tomography, preoperative evaluation, microglia, radiotracers, Refining, seizure, foci, localization