Curved reconstructions versus three-dimensional surface rendering in the demonstration of cortical lesions in patients with extratemporal epilepsy

Surgical planning, histopathology findings and postoperative outcome in MR-negative extra-temporal epilepsy using intracranial EEG, functional imaging, magnetoencephalography, neuronavigation and intraoperative MRI

OBJECTIVE

MRI-negative drug-resistant epilepsy presents a challenge when it comes to surgical planning, and surgical outcome is worse than in cases with an identified lesion. Although increasing implementation of more powerful MRI scanners and artificial intelligence has led to the detection of previously unrecognizable lesions, in some cases even postoperative pathological evaluation of electrographically epileptogenic zones shows no structural alterations. While in temporal lobe epilepsy a standardized resection approach can usually be performed, the surgical management of extra-temporal lesions is always individual. Here we present a strategy for treating patients with extra-temporal MRI-negative epilepsy focus and report our histological findings and patient outcome.

METHODS

Patients undergoing epilepsy surgery in the Department of Neurosurgery at the University Hospital Erlangen between 2012 and 2020 were included in the study. Inclusion criteria were: (1) failure to identify a structural lesion on preoperative high-resolution 3 Tesla MRI with a standardized epilepsy protocol and (2) preoperative intracranial EEG (iEEG) diagnostics.

RESULTS

We identified 8 patients corresponding to the inclusion criteria. Second look MRI analysis by an experienced neuroradiologist including the most recent analysis algorithm utilized in our clinic revealed a possible lesion in two patients. One of the patients with a clear focal cortical dysplasia (FCD) finding on a second look was excluded from further analysis. Of the other 7 patients, in one patient iEEG was performed with subdural electrodes, whereas the other 6 were evaluated with depth electrodes. MEG was performed preoperatively in all but one patient. An MEG focus was implemented in resection planning in 3 patients. FDG PET was performed in all, but only implemented in one patient. Histopathological evaluation revealed one non-lesional case, 4 cases of FCD and 2 cases with mild developmental malformation. All patients were free from permanent neurological deficits and presented with Engel 1A or 1B outcome on the last follow-up.

CONCLUSION

We demonstrate that extra-temporal MRI-negative epilepsy can be treated successfully provided an extensive preoperative planning is performed. The most important diagnostic was stereo-EEG, whereas additional data from MEG was helpful and FDG PET was rarely useful in our cohort.

Brain surface reformatted imaging (BSRI) for intraoperative neuronavigation in brain tumor surgery

BACKGROUND

For safe resection of lesions situated in or near eloquent brain regions, determination of their spatial and functional relationship is crucial. Since functional magnetic resonance imaging and intraoperative neurophysiological mapping are not available in all neurosurgical departments, we aimed to evaluate brain surface reformatted imaging (BSRI) as an additional display mode for neuronavigation.

METHODS

Eight patients suffering from perirolandic tumors were preoperatively studied with MRI and navigated transcranial magnetic stimulation (nTMS). Afterwards, the MRI was automatically transformed into BSR images in neuronavigation software (Brainlab, Brainlab AG, Feldkirchen, Germany). One experienced neuroradiologist, one experienced neurosurgeon, and two residents determined hand representation areas ipsilateral to each tumor on two-dimensional (2D) MR images and on BSR images. All results were compared to results from intraoperative direct cortical mapping of the hand motor cortex and to preoperative nTMS results.

RESULTS

Findings from nTMS and intraoperative direct cortical mapping of the hand motor cortex were congruent in all cases. Hand representation areas were correctly determined on BSR images in 81.3 % and on 2D-MR images in 93.75 % (p = 0.26). In a subgroup analysis, experienced observers showed more familiarity with BSRI than residents (96.9 vs. 84.4 % correct results, p = 0.19), with an equal error rate for 2D-MRI. The time required to define hand representation areas was significantly shorter using BSRI than using standard MRI (mean 27.4 vs. 40.4 s, p = 0.04).

CONCLUSIONS

With BSRI, a new method for neuronavigation is now available, allowing fast and easy intraoperative localization of distinct brain regions.

Evaluation of off-the-shelf displays for use in the American Board of Radiology maintenance of certification examination

PURPOSE

To prospectively compare high-, mid-, and low-resolution off-the-shelf displays currently employed by commercial testing centers, in terms of visibility of lesion features needed to render a diagnostic decision when possible diagnoses are provided in a multiple-choice format during a maintenance of certification (MOC) examination.

MATERIALS AND METHODS

The Psychometrics Division of the American Board of Radiology (ABR) approved the studies (human subjects and HIPAA compliant). One study compared 1280 x 1024 displays with 1024 x 768 displays; the second, 1600 x 1200 with 1280 x 1024 displays. Images from nine subspecialties were used. In each study, observers viewed images twice-once on each display. Diagnoses were provided, and observers rated visibility of diagnostic features.

RESULTS

Of 7977 data pairs analyzed in study 1, the 1024 and 1280 displays received the same ratings for 5726 data pairs (72% of the time), with the 1024 display receiving a higher rating for 679 data pairs (9% of the time) and the 1280 receiving a higher rating for 1572 data pairs (19% of the time) (P < .0001). When rating differences existed, all subspecialties except nuclear medicine had significantly more high-visibility ratings with the 1280 display. Of 1090 data pairs analyzed in study 2, the 1280 and 1600 displays received the same ratings for 689 data pairs (63% of the time), with the 1280 receiving a higher rating for 162 data pairs (15% of the time) and the 1600 receiving a higher rating for 239 data pairs (22% of the time) (P = .0001). When rating differences existed, only cardiopulmonary and musculoskeletal images had significantly more high-visibility ratings with the 1600 display.

CONCLUSION

For the ABR MOC examinations, 1280 x 1024 displays should be used, compared to 1024 x 768 displays; 1600 x 1200 displays may be necessary for some images. Good-quality images must be used on the examinations, so digital rather than digitized film images should be used to ensure high-quality images.