Abstracts

Rationale: Previous pilot data have shown the feasibility of an R2-based MRI method for detecting and quantifying pathology in patients with mesial temporal lobe epilepsy (MTLE). These are the preliminary results of a larger ongoing study to assess the clinical utility of this method. Presurgical assessment of pathology is crucial for these patients and surgical resection can be curative if a clear hippocampal lesion is identified by MRI. Standard diagnostic MRI fails to identify a lesion in about 30% of patients and these patients have a poorer post-surgical outcome when resected. Previously developed MRI approaches have not been sufficiently sensitive. The identification of subtle or early abnormalities, undetected by present standard MRI but potentially quantifiable by our MRI method, could help optimize patient selection and guide surgical decision-making. Methods: This ongoing study has recruited 10 healthy volunteers (negative controls) and 10 MTLE patients, 5 with positive diagnostic MRI (positive controls) and 5 with nonlesional MRIs. Laterality was confirmed by intracranial EEG and post-surgical histology. A modified fast-spin echo-based FLAIR sequence was used with 7 TEs. R2-maps were generated for multiple coronal slices through the head and body of both hippocampi. For comparison, conventional FLAIR images were acquired and T2-maps generated for the same slices. Using per-voxel R2 data from the positive and negative controls, a “percent-pathology” range was calculated and a percent-pathology map (PPM) made for each slice. From the PPMs, pathology-density scores were derived for each hippocampus. Asymmetry indices were generated for these scores as well as for two alternative MRI methods (T2 relaxometry and T2w signal intensity (T2wSI)). Results: The Mann-Whitney rank sum test was used to compare the results from each method for the pathological hippocampi to those from the normal control hippocampi. Both the T2 relaxometry (P=0.03) and R2-PPM (P<0.01) methods differentiated between the normal and MRI-positive abnormal hippocampi. T2wSI (P=ns) could not. Only R2-PPM (P<0.005) could differentiate between the normal and MRI-negative epileptogenic hippocampi. T2wSI (P=ns) and T2 relaxometry (P=ns) could not. The mean asymmetry in the five MRI-positive patients was 7% using T2wSI, 9% using T2 relaxometry, and 86% using our PPM method. The mean asymmetry in the five MRI-negative patients was 5% using T2wSI, 3% using T2 relaxometry, and 66% using our PPM method. Conclusions: The results of this ongoing study are promising. Analysis of this small sample suggests that the R2-PPM method can successfully differentiate between abnormal and normal hippocampi even in MRI-negative patients with MTLE. Moreover, the differentiation using R2-PPM is robust and considerably more sensitive than T2 relaxometry. If a larger study confirms these findings, this method would greatly help patient selection and surgical decision-making in MTLE patients, especially those with a normal MRI. (Source of funding: National Epifellows Foundation).