Abstracts

Rationale: The Australian Epilepsy Project (AEP) is a national-scale project to provide advanced imaging, cognition and genetic testing for people with epilepsy. Data will be used to drive development of outcome prediction and clinical decision support tools. This pilot study (February 2020 - May 2022) developed an imaging protocol for dual research-clinical use, targeting detection of subtle epileptogenic lesions, and assessing feasibility within this patient group.

Methods: Participants were referred from First Seizure and Epilepsy clinics in Melbourne, Australia. Adults aged 18 to 65 years with either first unprovoked seizure only (FUS), a new diagnosis of epilepsy (NDE) or drug-refractory focal epilepsy (DRE) were included. MRI was acquired on a 3T Siemens Prisma with 32-channel head coil, as a single 90-minute session with optional 10-minute break. The acquired sequences included a subset for clinical review (Table 1) which were reported by a neuroradiologist, noting epileptogenic lesions, any head-motion artefact and need for clinical follow-up.

Results: Recruitment: A total of 177 referred participants were eligible and consented. MRI was successfully acquired in 79% (40 FUS, 38 NDE, 61 DRE).  Six could not complete the scan due to body habitus, psychological discomfort, or seizures during the session. Six did not meet MRI safety requirements and 26 chose to withdraw before the scan (e.g., due to travel or COVID considerations). 
_x000D_ Image quality: Head-motion artefacts were noted to mildly affect or degrade some sequences, in 12% and 17% of cases respectively.
_x000D_ Imaging findings: Thirty-one percent of the first-seizure cohort (11 FUS, 13 NDE) had an epileptogenic lesion detected (Table 2). Most common were temporal pole encephaloceles (12%; a rate comparable to refractory focal epilepsy, Campbell et al. 2018), cavernous hemangioma (8%), and focal cortical dysplasia (4%). No hippocampal sclerosis was found. 
_x000D_ Follow up: Additional clinical imaging was recommended in 14%, for lesion characterisation with contrast (6%), dedicated pituitary imaging (5%) or assessment of interval change.

Conclusions: The AEP MRI research protocol is feasible for the majority of people with epilepsy. It is sensitive for epileptogenic lesions, with detection rates in a first-seizure cohort (31%) that are comparable or exceed previous studies (cf. 17% in King et al. 1998; 23% in Hakami et al. 2013). Temporal pole encephaloceles are emerging as an important potential cause of new-onset focal epilepsy. In contrast, hippocampal sclerosis was not found in this cohort. Routine use of tailored sequences that are sensitive to these subtle epileptic pathologies may assist with early imaging-based diagnosis.

Funding: This project was supported by a Australian Government MRFF Frontiers grant (MRFF75908), NHMRC project and programme grants (1091593 and APP1157145) and Victorian Government Operational Infrastructure grants. We acknowledge the facilities and assistance of the National Imaging Facility at The Florey node, an NCRIS capability.