PMID- 12076102 OWN - NLM STAT- MEDLINE DCOM- 20021104 LR - 20191210 IS - 1059-1311 (Print) IS - 1059-1311 (Linking) VI - 11 IP - 5 DP - 2002 Jul TI - Assessment and cost comparison of sleep-deprived EEG, MRI and PET in the prediction of surgical treatment for epilepsy. PG - 303-9 AB - Our aim was to determine if less expensive interictal indices can predict which epilepsy patients may benefit from the more expensive comprehensive pre-surgical evaluation. Surgical treatment was determined based on the results of a comprehensive inpatient continuous video-EEG monitoring. This evaluation included three interictal tests, which were reviewed retrospectively-2 hour-sleep-deprived electroencephalogram (SDEEG), magnetic resonance imaging (MRI), and positron emission tomography (PET). Sixty-nine patients were evaluated with 35 patients having focal resection (33 temporal, two frontal). When two or more interictal tests were positive, 77% (27 /35) went to surgery, but when one test was positive 23% (8 /34) had surgery. When all tests were negative, only a single patient (1 /13 or 7.7%) had surgery, a frontal resection. The positive predictive value for any single interictal test was 68%, while it was higher for any combination of two positive tests (77-83%). PET was the most sensitive (0.86) single interictal test, compared to SDEEG (0.66) and MRI (0.66). The odds ratio for predicting surgical treatment for a positive PET, SDEEG, or MRI was 8.57, 4.01, and 4.01, respectively. MRI was three and PET was six times the cost of a SDEEG. The combination of SDEEG and MRI had the best cost/PPV ratio. Seventy-nine percent (11 /14) of the patients with three positive tests were seizure free following focal resection compared to 43% (9 /21) when less than three tests were positive ( P