Sensitivity and specificity of quantitative difference SPECT analysis in seizure localization

True ictal SPECT can accurately demonstrate perfusion increases in the epil eptogenic area but often requires dedicated personnel waiting at the bedsid e to accomplish the injection. We investigated the value of perfusion chang es as measured by ictal or immediate postictal SPECT in localizing the epil eptogenic region in refractory partial epilepsy. Methods: Quantitative perf usion difference images were calculated by registering, normalizing and sub tracting ictal (or immediate postictal) from interictal SPECT for 53 patien ts with refractory epilepsy Perfusion difference SPECT results were compare d with visually interpreted SPECT, scalp electroencephalography (EEG), MRI, PET and intracranial EEG. Results: In 43 patients (81%), discrete areas of increased perfusion (with ictal injections) or decreased perfusion (with p ostictal injections) were noted. interictal scalp EEG was localizing in 28 patients (53%), ictal scalp EEG was localizing in 35 patients (66%) and int racranial EEG was localizing in 22 patients (85%) (of 26 patients who under went invasive study). MRI was localizing in 34 patients (64%), PET was loca lizing in 32 of 45 patients (71%), interictal SPECT was localizing in 26 pa tients (49%) and peri-ictal SPECT (visual interpretation) was localizing in 30 patients (57%). By comparison with an intracranial EEG standard of loca lization, SPECT subtraction analysis had 86% sensitivity and 75% specificit y. Conclusion: Our data provide evidence that SPECT perfusion difference an alysis has higher sensitivity and specificity than any other noninvasive lo calizing criterion and can localize epileptogenic regions with accuracy com parable with that of intracranial EEG. To obtain these results, one must ap ply knowledge of the timing of the ictal injection relative to seizure occu rrence.