Authors :
Presenting Author: Merran Courtney, MD – Monash University
Ana Antonic-Baker, PhD – Monash University; Benjamin Sinclair, PhD – Monash University; Andrew Neal, MBBS, PhD – Monash University; John-Paul Nicolo, MBBS, PhD – Monash University; Cassandra Marotta, BSc (Hons) – Monash University; Jacob Bunyamin, MSc – Monash University; Zhibin Chen, PhD – Monash University; Meng Law, MBBS, PhD – Monash University; Patrick Kwan, MB BChir, PhD – Monash University; Terence O'Brien, MBBS, MD – Monash University; Lucy Vivash, PhD – Monash University
Rationale:
Although 18F-FDG-PET is commonly used in the pre-surgical evaluation of patients with drug-resistant epilepsy, the prognostic value of the detection of localising hypometabolism is uncertain. Previous meta-analyses have had negative results. Globally, 18F-FDG-PET is used in a variety of ways, with some centres using it routinely for all epilepsy surgery candidates, and others only using it in selected cohorts, such as those with non-lesional (MRI negative) epilepsy. We designed a systematic review and meta-analysis to determine whether localising 18F-FDG-PET hypometabolism is predictive of favourable outcome in epilepsy surgery.
Methods:
A systematic literature search of Embase, Medline, and Web of Science was undertaken. Eligible publications included evaluation with 18F-FDG-PET prior to resective epilepsy surgery, with at least ten participants, and reported surgical outcome at ≥12 months. Random effects meta-analysis was used to calculate the proportion of patients with localising 18F-FDG-PET hypometabolism achieving a favourable (Engel class I, ILAE class 1-2, or seizure-free) outcome, compared to those without localising hypometabolism. Localising hypometabolism was defined as concordant with other presurgical investigations and/or the final decision regarding surgical site alone. Secondary analyses included comparisons in outcome where the 18F-FDG-PET finding was concordant with MRI and/or EEG findings, and between different patterns of hypometabolism, such as focal (confined to a single epileptogenic lobe), regional (2 ipsilateral adjacent lobes), diffuse ( >2 adjacent lobes or bilateral), or normal scan. Sources of heterogeneity were investigated using meta-regression. The protocol was prospectively registered in PROSPERO.
Results:
The database search identified 12917 studies. Of these, 98 studies (4104 patients in total) were included. Localising 18F-FDG-PET hypometabolism was predictive of favourable outcome in epilepsy surgery for drug resistant epilepsy with odds ratio (OR) 2.56 (95%CI 1.99 – 3.29), when compared to those without localising hypometabolism. In the secondary analyses, concordant findings on 18F-FDG-PET with EEG (OR 2.34, 95%CI 1.44 – 3.82), MRI (OR 1.66, 95%CI 1.18 – 2.34), and triple concordance with 18F-FDG-PET, EEG and MRI (OR 2.14, 95%CI 1.33 – 3.46) were also predictive of favourable outcome. Compared to focal hypometabolism, diffuse hypometabolism was associated with a worse epilepsy surgery outcome (OR 0.37, 95%CI 0.23 – 0.60).
Conclusions:
This systematic review and meta-analysis across a large number of studies and patients demonstrated that localising 18F-FDG-PET hypometabolism does predict favourable outcome in epilepsy surgery. We also demonstrated that the pattern of hypometabolism on 18F-FDG-PET scan is important, with focal changes being associated with favourable outcome, and diffuse changes being associated with worse outcomes (Engel II-IV, ILAE 3-6, or not seizure free). These findings support the routine use of 18F-FDG-PET for patients undergoing evaluation for epilepsy surgery and support the inclusion of 18F-FDG-PET in multimodal tools for epilepsy surgery outcome prediction.
Funding: MRC receives an Australian Government RTP scholarship.