Ictal hyperperfusion and hypoperfusion in subtracted SPECT are generally considered significant when the difference of rCBF in each pixel is greater than two SD. The location of significant ictal hyperperfusion is visually decided on MRI by SPECT-MRI coregistration. For the determination of significant rCBF changes, only brain regions showing three and over three voxels (3.56 mm × 3 = 10.68 mm) with significantly different intensity values between ictal and interictal SPECT images are included. Because subtracted SPECT may reflect the perictal or postictal state when the injection of radioisotope was delayed or seizures were very short lasting, subtracted SPECT should be interpreted with considering the interval between radioisotope injection and seizure onset, duration of seizure, and time of seizure end. The propagation pattern of ictal EEG discharges at the time of radiotracer injection is also very important because it may affect the final result of subtracted SPECT. Frontal lobe seizures show a tendency of rapid propagation compared to temporal lobe seizures. The interpretation of SISCOM should include careful simultaneous review of seizure semiology, sequences of ictal EEG patterns, injection time in relation to seizure duration, structural lesion, and other clinical information. SISCOM reflects only cerebral perfusion changes (significantly different from the interictal state) at the time of tracer injection, not always the zone of seizure onset.

The first study using subtracted SPECT was published by Zubal et al. (1995). From this difference image technique, the epileptic focus was localized in five of six temporal lobe focus patients and in four of six for extratemporal seizure patients. This result was superior to visual analysis, and the difference images provided more precise information regarding location of the seizure focus.

Subtraction of interictal from ictal SPECT coregistered to 3-dimensional MRI (SISCOM) is an improved and better validated technique for ictal and interictal SPECT subtraction. Various clinical studies have been performed using this SISCOM technique in the field of epilepsy (O’Brien et al. 1998a, b, 1999; Lee et al. 2000). In the first study on patients with intractable partial epilepsy (O’Brien et al. 1998a), SISCOM was able to localize the epileptic focus in 45 out of 51 (88.2 %) patients, while the traditional side-by-side inspection of ictal and interictal SPECT images localized 20 out of 51 (39.2 %). However, late injection of the radiotracer (>45 s after seizure onset) but not secondary generalization was associated with a falsely localizing or nonlocalizing SISCOM. To overcome the problem of delayed injection, automated equipment for self-injection was developed which improved the sensitivity of ictal SPECT (Van Paesschen et al. 2000). To circumvent the difficulty of early injection during seizures, a postictal SPECT study was performed (O’Brien et al. 1999). The combined interpretation of hyperperfusion (positive values) and hypoperfusion (negative values) of SISCOM showed a better concordance rate with the final diagnosis than the analysis of either the hyperperfusion or hypoperfusion SISCOM image alone. In a similar study (Lee et al. 2000), the sensitivity of the combined hyperperfusion and hypoperfusion SISCOM (91.8 %) was superior to only hyperperfusion SISCOM (85.2 %) and visual inspection of ictal SPECT (68.9 %). Ictal hypoperfusion at the epileptic focus was first reported in that study. Figure 39.2e demonstrated that a patient with an epileptic focus at the left posterior lateral temporal cortex showed ictal hypoperfusion at the epileptic focus, while ictal hyperperfusion was seen at left anterior temporal region where ictal discharges propagated rapidly at the time of radiotracer injection. Rapid propagation of ictal discharges to other brain regions with early postictal state of the epileptic focus despite persistence of clinical seizures is the probable explanation for ictal hypoperfusion. In cases where early radiotracer injection is difficult, postictal SPECT subtraction can also be helpful in the evaluation and localization of the epileptic focus (Zubal et al. 1999).

Other study compared the ability of MSI, FDG-PET, and ictal SPECT to predict seizure-free outcome after epilepsy surgery with intracranial EEG monitoring. The odds ratios of MSI (magnetic source imaging), FDG-PET, and ictal SPECT for prediction of seizure outcome were 4.4, 7.1, and 9.1, respectively, although the number of patients who had MSI, FDG-PET, or ictal SPECT was different (Knowlton at al. 2008).

Another study compared different diagnostic tools in the prognostic factors for surgical outcome in neocortical epilepsy patients. Diagnostic sensitivities of interictal EEG, ictal scalp EEG, FDG-PET, and subtraction ictal single photon emission computed tomography were 37.1 %, 70.8 %, 44.3 %, and 41.1 %, respectively. Localization by FDG-PET and interictal EEG was correlated with a seizure-free outcome. Concordance with two or more presurgical evaluations was significantly related to a seizure-free outcome. But this study showed lower sensitivity of SISCOM compared to other studies (Lee et al. 2005). So there are many factors affecting the result of ictal SPECT and SISCOM including seizure type and duration, radiotracer type, injection time, epilepsy monitoring unit setting, experience of technicians, and the method of SISCOM.

Ahnlide et al. (2007) analyzed SISCOM in 26 patients who had intracranial EEG monitoring and epilepsy surgery. SISCOM findings altered and extended the strategy for electrode placement at invasive recording in 15 patients. In this study, SISCOM was a prerequisite for seizure outcome in all six patients with favorable outcomes, and SISCOM is valuable for the identification of the epileptogenic zone in patients with drug-resistant epilepsy scheduled for invasive video-EEG.

Other two studies also indicated that SISCOM may have an impact on surgical decision (Tan et al. 2008; von Oertzen et al. 2011) (Fig. 39.3).