Abstracts

RATIONALE: Standard temporal resection in the left, language-dominant hemisphere carries the risk of postoperative decline in naming ability, a risk associated with later epilepsy onset age and the absence of hippocampal sclerosis. Preoperative language mapping has been performed routinely at some centers to minimize postoperative primary language impairment, but its effect on changes in naming performance has not been explored. This study specifically examined naming outcome in relation to onset age in patients who had temporal resection after language mapping.
METHODS: The sample consisted of 24 patients (15 male) undergoing left temporal tailored resection after extraoperative language mapping. Mean onset age was 12 years and age at surgery 27 years. Mapping involved identification of sites where stimulation resulted in errors of automatic speech, naming, repetition, comprehension and reading. Primary temporal language areas were defined based on error clusters and resection always spared these areas. However, resection did not necessarily spare every language error site. Patients were administered the Boston Naming Test (BNT) pre- and six months postoperatively. All patients were left hemisphere dominant for language by amobarbital testing, had FSIQ[gt]69, and no lesion on MRI apart from changes of hippocampal sclerosis. Reliable Change Index (RCI) (5 for BNT) was used as an indication of meaningful change.
RESULTS: No patient was aphasic postoperatively. There was significant correlation between onset age and change in BNT, later onset being associated with greater BNT decline (r=-0.51, p[lt]0.05). There was a significant difference in BNT decline for patients with onsets [lt]12 years and [gt]12 years (0.82 and [ndash]12.00 respectively, t=3.1, p=0.005). Overall, 11 (46%) patients had BNT decline greater than RCI. Proportions of patients undergoing RCI decline for patients with onset [lt]12 years and [gt]12 years were 18% and 69% respectively ([chi]2=6.25, p[lt]0.05).
CONCLUSIONS: Postoperative naming ability in cases with early onset ([lt]12 years) remains stable with tailored resection based on language mapping. In patients with later onset age a decline in naming is more likely. The relationship between naming decline and the pattern and type of language error sites will be reported. The practical implications of these findings will be discussed.