Abstracts

Rationale: Volumetric measurements of mesial temporal structures have gained a significant role in the presurgical evaluations of patients with focal epilepsy of presumed mesial temporal origin. The manual tracing method (MTM) is the gold standard to assess hippocampal atrophy (HA). This method, however, is time-consuming and requires skilled personnel. In recent years, a normative calculator [Potvin et al.NeuroImage 137(2016)9-20] compatible with an automated method (AM) Freesurfer v5.3 software [Fischl et al.Neuroimage 9(1999)195-207; Current version is 6.0] has been validated in healthy adults, in patients with Schizophrenia and dementia. The calculator is an open source spreadsheet with formulas that are easy to use but has been not been applied to adults with Epilepsy. The purpose of this study was to establish whether the AM and MTM yielded comparable results in the identification of HA in patients with temporal lobe epilepsy. Methods: The quantitative volumetric measurements were performed in 3-Tesla high-resolution brain MRI studies of 83 epilepsy patients [male=44(53%); mean age=39yo (+14); TLE=57(69%) of whom MTS=40 (48%);]. The established MTM was considered the gold standard method for identifying HA.HA in MTM was considered to be present when the volume difference of Right-Left hippocampi was >715 mm³ or <-300 mm³, for left and right sides respectively. Bilateral HA was considered when the normalized values were simultaneously < 1664 mm³ on the left and <1926 mm³ on the right side.HA in AM was confirmed by the calculator significant results based on False Discovery Rate – FDR adjusted p-value (q-value). Kappa measure of agreement (Kappa) was used to compare the two methods. The two methods had their results compared based on the 4 possible results: No atrophy; Left atrophy; Right atrophy; Bilateral atrophy. Results: For the entire group of patients, there was a moderate agreement between MTM and AM (Kappa = 0.5, p <.001). Among 24 patients with unilateral HA, based on MTM, AM yielded concordant data in 16 (67%). Among the 16 patients with bilateral HA by MTM, AM yielded concordant findings in only two (12%).  Among the 43 patients without HA, based on MTM, AM suggested the presence of HA in three patients (7%). Conclusions: These data suggest that an AM cannot replace the MTM, as it may miss the presence of bilateral HA and in one third of cases, unilateral HA.  Funding: None