Abstracts

Rationale: The IAP is widely used to assess language and memory prior to resective neurosurgery. On occasion, it is repeated due to procedural difficulties (e.g., drowsiness/poor cooperation) or clinical need (e.g., years lapsed between the initial IAP and the time of surgery). We reviewed our experience with repeating this procedure to better understand IAP reliability. Methods: A review of the NYU CEC patient registry (2001-2009) revealed 562 IAPs with 16 individuals (6 men) having repeated IAPs. There were two children, ages 5 and 11; adult ages ranged from 20-60 years at the time of their first IAP. Educational experience ranged from attendance to a special needs school to having completed Master's degrees. General intellectual level paralleled this diversity with individuals performing across the spectrum from "Deficient" to "Above Average". Age of seizure onset ranged from 11 months to 49 years. At the time of the first IAP, there were 14 subjects with left hemisphere seizure onset, one with a right sided focus, and one had bilateral independent foci. Reasons for repeating the IAP included procedural difficulties (6), patients considering a first (3) or second (4) surgery years after their first IAP, and the first IAP results were the wrong way (3). The inter-IAP-interval was quite variable (1 month to 15 years). While most IAPs were performed at NYU, some patients had undergone an IAP elsewhere. Thus, the language and/or memory scoring systems were slightly different across procedures which limited direct comparison. Therefore, for each hemisphere injected, the final criterion for language localization (left, right, or bilateral) and memory functionality ("pass" or "fail") were recorded. One patient had 3 IAPs, 3 patients had either a 3rd injection during one of their procedures or did not have both hemispheres retested during the second IAP, and for one patient, the language but not the memory data were available for one IAP. Thus, there were a total of 14 ipsilateral and 17 contralateral injections reviewed. Results: Language functioning was consistently localized (11 L, 2 R, and 2 bilateral) in all cases. In terms of memory, 86% of the ipsilateral and 71% of the contralateral pass/fail outcomes were consistent across procedures. A Yate's corrected chi square was significant for ipsilateral (p=.05) but not for contralateral (p=.29) injections. When just the 6 patients who underwent repeat IAP due to technical difficulties were reviewed, 100% consistency was seen across procedures. Conclusions: Language lateralization findings were robust and unchanged regardless of the indication for repeat IAP. Interestingly, while procedures repeated for "technical reasons" showed no change in memory ability (despite such things as reduced dose and improved alertness), "technically valid" IAPs that were performed following the natural clinical course of epilepsy or surgical intervention showed the most change. Therefore, rather than assume that test-retest variance is due to unreliability of the procedure, it is possible that the IAP on repeat administration is detecting unique information about changes in brain functionality.