Basic Science Clinical

Vascular Headache Epilepsy Immunology Movement Neuromuscular Dementia Other Topics Neurosurgery

Presurgical Evaluation

What Do You Need to Know?

Drug-resistant epilepsy = failure of 2 appropriate ASMs; probability of seizure freedom drops to ~5% per subsequent agent — refer early
Phase I (noninvasive): video-EEG, 3T MRI epilepsy protocol, FDG-PET, ictal SPECT/SISCOM, MEG/MSI, neuropsychological testing
Phase II (invasive): SEEG has become the predominant intracranial modality in North America for most indications; subdural grids retain a role when high-density cortical functional mapping is the primary goal; Phase II required in 30–40% of surgical candidates
Language lateralization: fMRI has largely replaced Wada (>90% concordance); Wada still needed for memory lateralization
Concordance model: all modalities concordant → best outcomes; in mTLE/HS with full concordance Engel I reaches 70–80%; extratemporal concordant cases lower (typically 40–60% Engel I); discordance → poorer outcomes or need for Phase II
MRI-negative epilepsy: 20–40% of drug-resistant focal epilepsy (depending on protocol/field strength); 30–45% Engel I (vs. 60–70% for lesional TLE); PET, MEG, SEEG become critical
Best prognostic factors: identifiable MRI lesion, mTLE with HS, concordance across modalities, shorter epilepsy duration

🚩 Don’t Miss — Test-Day Priorities

Drug-resistant epilepsy = 2 failed ASMs: Refer EARLY — each subsequent ASM adds only ~5% chance of seizure freedom; do NOT wait for 5+ failures.
Concordance is everything: If video-EEG + MRI + PET + MEG all converge on one zone → resect; if discordant or non-lesional → Phase II SEEG.
Ictal SPECT timing: Tc-99m HMPAO must be injected within seconds of seizure onset (<20 sec ideal; <10–20 sec for extratemporal) — late injection captures propagation, not onset.
SISCOM = subtraction ictal SPECT co-registered with MRI: mismatch between ictal hyperperfusion and interictal hypoperfusion identifies seizure onset zone — gold standard for SPECT analysis.
fMRI replaces Wada for language in most centers; Wada (intracarotid amobarbital) still used selectively for memory lateralization and postoperative deficit risk assessment.
Neuropsych red flags for postop verbal memory decline: intact preoperative verbal memory + planned dominant temporal lobectomy = highest risk; bilateral hippocampal involvement and near-normal baseline also high risk.
SEEG vs subdural grids: SEEG = depth electrodes, deep/multilobar sampling, lower complication rate; subdural grids = high-density cortical mapping including ECS for eloquent cortex.
Engel I outcomes by substrate: MTLE-HS with ATL = 60–70%; FCD with concordant data = 40–60%; non-lesional = 25–40%.
RNS (responsive neurostimulation): bilateral mesial temporal foci, eloquent cortex pathology, or multifocal disease where resection is inadvisable.
NEVER resect eloquent cortex (motor, speech, memory) without ECS mapping — awake craniotomy with cortical stimulation preserves function.

🔍 Buzzwords & Pathognomonic FindingsPhase I modalities · Localization techniques · Phase II / pitfalls

Phase I (non-invasive) modalities

Video-EEG monitoring → capture ≥3 habitual seizures; gold standard for seizure-onset zone localization
3T MRI epilepsy protocol (HARNESS-MRI) → thin-slice coronal hippocampi, 3D FLAIR, 3D T1 IR for FCD detection
FDG-PET → interictal hypometabolism in epileptogenic zone (80–90% sensitivity in mTLE)
Ictal SPECT (Tc-99m HMPAO) → injected within seconds of seizure onset → SISCOM co-registered with MRI
MEG / magnetic source imaging → interictal magnetic dipole source localization; detects tangential sulcal dipoles missed by EEG
fMRI language/motor mapping → has replaced Wada in most centers (>90% concordance)

Localization concepts

Concordance model → all modalities converge on same zone → resect; discordance → Phase II or poorer outcome
SISCOM mismatch → ictal hyperperfusion + interictal hypoperfusion identifies seizure onset zone
Wada test (intracarotid amobarbital) → transient hemispheric anesthesia tests contralateral memory + language
Verbal memory deficit on neuropsych → language-dominant (usually left) temporal lobe pathology
Visuospatial memory deficit → non-dominant (usually right) temporal lobe pathology
Engel classification → I = seizure-free; II = rare disabling (<3/yr); III = worthwhile improvement; IV = no benefit

Phase II / pitfalls

SEEG (stereo-EEG) → depth electrodes for deep/multilobar sampling; lower complication rate vs subdural grids
Subdural grids → high-density cortical mapping including ECS for eloquent cortex preservation
Electrical cortical stimulation (ECS) during awake craniotomy → language/motor cortex mapping before resection
RNS (responsive neurostimulation) → bilateral mesial temporal, eloquent cortex, or multifocal foci where resection inadvisable
Late SPECT injection → captures propagation, NOT seizure onset — misleading localization
Resection of eloquent cortex without ECS mapping → postoperative motor/language deficit — AVOID

When to Refer for Epilepsy Surgery

Drug-resistant epilepsy (ILAE definition): failure to achieve seizure freedom after adequate trials of 2 tolerated, appropriately chosen ASMs
Probability of seizure freedom drops to ~5% with each subsequent ASM after 2 failures
Earlier referral = better outcomes: average delay from drug resistance to surgery is 10–20 years in many series
ERSET (Engel JAMA 2012): small RCT (n=38, terminated early); 73% of surgical arm vs. 0% medical arm seizure-free during year 2
ILAE recommends referral as soon as drug resistance is identified, regardless of epilepsy type
Comprehensive epilepsy center care reduces premature mortality — even in patients who do not undergo surgery

💎 Board Pearl

Drug-resistant epilepsy = failure of 2 ASMs. After 2 failures, each additional ASM adds only ~5% chance of seizure freedom. Do NOT wait for 5+ ASM failures before referring. Boards test this threshold repeatedly.

Phase I (Noninvasive) Evaluation

Modality	What It Shows	Sensitivity / Specificity	Key Points
Video-EEG monitoring	Capture habitual seizures; identify seizure-onset zone; classify semiology	Gold standard for seizure localization	Typically 5–14 days; capture ≥3–5 habitual seizures; ASMs often reduced; interictal IEDs lateralize irritative zone
3T MRI epilepsy protocol	Structural lesion identification (HS, FCD, tumors, vascular malformations)	1.5T misses ~20% of lesions detected at 3T	ILAE HARNESS-MRI protocol (Bernasconi et al., Epilepsia 2019); key sequences: 3D T1 (1 mm), 3D FLAIR, coronal T2 perpendicular to hippocampus, SWI; NOT a “routine brain MRI”
FDG-PET	Interictal hypometabolism in epileptogenic zone	80–90% sensitivity for mTLE; 45–60% extratemporal	More sensitive than MRI for some subtle lesions; concordance with EEG strengthens surgical candidacy
Ictal SPECT (SISCOM)	Ictal hyperperfusion at seizure-onset zone	70–90% for TLE; lower for extratemporal	Inject as early as possible — optimal <20 sec; <45 sec for TLE; for extratemporal/frontal seizures the window is much narrower (<10–20 sec) because of rapid propagation; SISCOM = subtraction ictal SPECT coregistered to MRI
MEG / MSI	Magnetic source imaging of interictal epileptiform discharges	Complementary to EEG; better for neocortical foci	Most useful in MRI-negative cases; detects tangential dipoles (sulcal cortex) better than EEG; guides SEEG placement
Neuropsychological testing	Baseline cognitive function; lateralization of language/memory	Supports lateralization; predicts postop deficits	Verbal memory deficit = language-dominant (usually left) temporal lobe; visuospatial memory deficit = non-dominant (usually right) temporal lobe; establishes preoperative baseline

MRI Epilepsy Protocol — Key Sequences

ILAE HARNESS-MRI protocol (Bernasconi et al., Epilepsia 2019): minimum core sequences = 3D T1 millimetric, 3D FLAIR, high-resolution 2D coronal T2 perpendicular to hippocampal long axis.

3D T1 (1 mm isotropic): cortical thickness, gray-white junction blurring (FCD), volumetric analysis
3D FLAIR (1 mm isotropic): hippocampal signal abnormalities, FCD, gliosis
Coronal T2 (2–3 mm): perpendicular to long axis of hippocampus — essential for HS detection
SWI/GRE: cavernous malformations, calcifications, hemosiderin deposits

PET vs. SPECT — Key Distinctions

FDG-PET: interictal study; shows hypometabolism; hypometabolism extends beyond epileptogenic zone (localizing but not precise)
Ictal SPECT: inject as early as possible — optimal <20 sec; <45 sec acceptable for TLE; for extratemporal/frontal seizures the window narrows to <10–20 sec due to rapid propagation; late injection = propagation, NOT onset; SISCOM increases accuracy

💎 Board Pearl

FDG-PET = interictal hypometabolism. Ictal SPECT = ictal hyperperfusion. These are OPPOSITE findings, both localizing to the epileptogenic zone. SPECT should be injected as early as possible — optimal <20 sec; <45 sec acceptable for TLE; <10–20 sec for extratemporal/frontal seizures due to rapid propagation. Late injection shows propagation, not origin.

Language & Memory Lateralization

Language Dominance

Left hemisphere dominant: 95% of right-handers; ~70–80% of left-handers (with ~15% bilateral and ~10% right-dominant); bilateral/right language more common with early left-hemisphere lesions

fMRI vs. Wada Test

Feature	fMRI	Wada Test
Language lateralization	>90% concordance with Wada in typical right-handers with clear lateralization; preferred first-line; concordance drops to ~70–80% in left-handers, bilateral representation, and lesional cases near language cortex — Wada preferred in those scenarios	Gold standard but invasive; declining use
Memory lateralization	Hippocampal fMRI protocols exist but not yet validated as Wada replacement	Remains the standard for memory lateralization
Invasiveness	Noninvasive; repeatable	Invasive (catheter angiography); transient neurologic complications ~0.6%, permanent stroke ~0.05–0.1%
Availability	Widely available	Limited to select centers; declining expertise

When Is the Wada Test Still Needed?

fMRI shows atypical or bilateral language representation
Memory lateralization before temporal lobe resection (esp. left TLE with concern for memory decline)
fMRI nondiagnostic due to motion artifact, poor task performance, or technical failure
Concern for contralateral hippocampal insufficiency (bilateral hippocampal abnormalities)

Activation Agents (Wada and Intraoperative)

Historical Wada agent: sodium amobarbital (Amytal)
Current alternatives due to amobarbital supply shortages:
- Methohexital (short-acting barbiturate) — used for both Wada-equivalent intracarotid testing and intraoperative ECoG activation
- Etomidate — Mayo Clinic Etomidate Speech and Memory protocol (eSAM); also used intraoperatively because etomidate activates interictal epileptiform discharges, helping identify the irritative zone

💎 Board Pearl

Wada test classically used amobarbital but, due to amobarbital availability, methohexital and etomidate are increasingly the standard agents at many centers. Etomidate has the added advantage of activating interictal epileptiform discharges intraoperatively.

💎 Board Pearl

fMRI has replaced Wada for LANGUAGE lateralization (>90% concordance). Wada is still needed for MEMORY lateralization — especially before left temporal resection. Left hemisphere = language dominant in 95% of right-handers, 70–80% of left-handers.

Phase II (Invasive) Evaluation

Indications for Intracranial EEG

Discordant/inconclusive Phase I data; MRI-negative with lateralized EEG
Seizure onset from or near eloquent cortex; bilateral independent temporal onsets
Extratemporal epilepsy with broad localization; 30–40% of surgical candidates require Phase II

SEEG vs. Subdural Grids

Feature	SEEG (Depth Electrodes)	Subdural Grids
Implantation	Stereotactic via twist-drill holes; robot-assisted	Open craniotomy
Spatial coverage	Deep structures (hippocampus, insula, cingulate); bilateral feasible	Cortical surface; limited deep access; difficult bilateral placement
Complication rate	Symptomatic hemorrhage ~1% (Mullin meta-analysis 2016); permanent neurologic deficit ~0.6%; mortality ~0.3%; infection 1–2%; overall lower than grids	Overall 10–15% (hemorrhage, infection, CSF leak, edema)
Cortical mapping	Limited by electrode geometry	Excellent for detailed motor/sensory/language mapping
Current trend	Predominant intracranial modality in North America for most indications since ~2015	Declining; reserved for specific cortical mapping indications
Seizure-free outcomes	Seizure-free outcomes after resection guided by either modality are comparable in propensity-matched analyses (Jehi 2021); SEEG's advantage is lower morbidity, not higher cure rate

SEEG Technical Points

Typically 8–16 depth electrodes (8–18 contacts each); hypothesis-driven placement based on Phase I data
Robotic-assisted implantation (ROSA, Neuromate) — target error <2 mm; monitoring 7–14 days
SEEG-guided thermocoagulation: can ablate small foci at electrodes — minimally invasive therapeutic option

💎 Board Pearl

SEEG is the predominant intracranial modality in North America for most indications. SEEG = lower complications (~1% symptomatic hemorrhage vs. 10–15% overall for grids), better for deep structures (hippocampus, insula, cingulate), and allows bilateral implantation. Subdural grids retain a role when high-density cortical functional mapping is the primary goal.

Concordance Model & Surgical Decision-Making

All modalities concordant (semiology + EEG + MRI + PET + neuropsych) → best outcomes; in mTLE/HS with full concordance, Engel I rates reach 70–80%; extratemporal concordant cases are lower (typically 40–60% Engel I)
Discordance among modalities → poorer outcomes; may need Phase II or may not proceed with surgery
No single test is sufficient — convergence of multiple independent data sources is the fundamental principle
Final decision at multidisciplinary epilepsy surgery conference (epileptologist, neurosurgeon, neuroradiologist, neuropsychologist)

Key Concordance Domains

Seizure semiology • Scalp EEG (interictal + ictal) • MRI (structural lesion)
Neuropsychological profile • FDG-PET (hypometabolism) • Ictal SPECT (hyperperfusion) • MEG

MRI-Negative Epilepsy

Definition: no identifiable lesion on 3T MRI; affects 20–40% of drug-resistant focal epilepsy patients (depending on imaging protocol and field strength)
Surgical outcomes: 30–45% Engel I (vs. 60–70% for lesional TLE) — counsel patients about lower probability
PET, MEG, SEEG become critical — concordance among these can substitute for a visible lesion
7T MRI + computational postprocessing detects subtle FCD and hippocampal subfield abnormalities missed on 3T; reveals lesions in 20–30% of previously MRI-negative cases
Most MRI-negative patients require SEEG before resection; if not feasible, consider neuromodulation (RNS, VNS, DBS)

Engel Classification of Surgical Outcomes

Engel Class	Outcome	Subclasses
Class I	Free of disabling seizures	Ia = completely seizure-free; Ib = only auras; Ic = some seizures postop but seizure-free ≥2 years; Id = only generalized seizures with drug withdrawal
Class II	Rare disabling seizures	Almost seizure-free; rare seizures after initial complete control
Class III	Worthwhile improvement	Meaningful seizure reduction but ongoing disabling seizures
Class IV	No worthwhile improvement	No significant change or worsening of seizures

💎 Board Pearl

Engel Ia = completely seizure-free (the goal). Engel Ib = only auras (still considered “free of disabling seizures”). Boards may describe a patient with only auras postop and ask you to classify — that is Engel Class Ib, NOT Class II.

Favorable Prognostic Factors for Surgery

Identifiable lesion on MRI — strongest single predictor (especially HS or low-grade tumor)
Concordance across all modalities (EEG, MRI, PET, semiology, neuropsych)
Mesial temporal lobe epilepsy with hippocampal sclerosis — 60–80% Engel I
Shorter duration of epilepsy before surgery — earlier surgery = better outcomes
Normal IQ (≥70)
Unilateral interictal epileptiform discharges concordant with MRI; single seizure type with consistent aura
Complete resection of the epileptogenic lesion; temporal lobe epilepsy (better outcomes than extratemporal)

Unfavorable Factors / Relative Contraindications

Bilateral independent ictal onsets — precludes single resective approach; consider neuromodulation
Primary generalized epilepsy — not a resective surgical candidate
Active psychosis or unstable major psychiatric illness
Eloquent cortex involvement without sparing option — epileptogenic zone overlaps motor, language, or primary visual cortex with no safe resection margin
Severe bilateral cognitive dysfunction — limited functional reserve and unclear benefit

💎 Board Pearl

Drug-resistant epilepsy = failure of 2 ASMs (not 3, not 5) — ~5% chance of seizure freedom per additional agent after 2 failures
FDG-PET = interictal HYPOmetabolism; Ictal SPECT = ictal HYPERperfusion — both localize to the epileptogenic zone but at different times
SISCOM = subtraction ictal SPECT coregistered to MRI; inject as early as possible (optimal <20 sec; <45 sec for TLE; <10–20 sec for extratemporal/frontal due to rapid propagation)
fMRI replaces Wada for language in typical right-handers with clear lateralization (>90% concordance); concordance drops to ~70–80% in left-handers, bilateral representation, and lesional cases near language cortex — Wada still needed for memory lateralization and for atypical language cases
SEEG is the predominant intracranial modality in North America for most indications: lower complications, better for deep structures, allows bilateral sampling; subdural grids retain a role for high-density cortical functional mapping
Concordance = success: all modalities agree → best outcomes (Engel I 70–80% in mTLE/HS, 40–60% extratemporal); any discordance lowers expected outcomes

Clinical Pearls

A “normal MRI” does NOT mean the patient is not a surgical candidate. MRI-negative patients can achieve 30–45% seizure freedom with surgery. PET, MEG, and SEEG can localize the epileptogenic zone when MRI fails. Always use 3T with an epilepsy protocol — 1.5T misses ~20% of lesions.
The average delay from drug resistance to surgical referral is 10–20 years. Every year of ongoing seizures worsens cognitive outcomes, psychosocial disability, and SUDEP risk. If a patient has failed 2 appropriate ASMs, the conversation about surgery should begin immediately.

References

Friedman D, Engel J. Surgical treatments, devices, and nonmedical management of epilepsy. Continuum (Minneap Minn). 2025;31(1):165–186.
Kwan P, Arzimanoglou A, Berg AT, et al. Definition of drug resistant epilepsy: consensus proposal by the ad hoc Task Force of the ILAE Commission on Therapeutic Strategies. Epilepsia. 2010;51(6):1069–1077.
Jehi L, Jette N, Kwon CS, et al. Timing of referral to evaluate for epilepsy surgery: expert consensus recommendations from the Surgical Therapies Commission of the ILAE. Epilepsia. 2022;63(10):2491–2506.
Engel J Jr, McDermott MP, Wiebe S, et al. Early surgical therapy for drug-resistant temporal lobe epilepsy: a randomized trial (ERSET). JAMA. 2012;307(9):922–930.
Tandon N, Tong BA, Friedman ER, et al. Analysis of morbidity and outcomes associated with use of subdural grids vs stereoelectroencephalography in patients with intractable epilepsy. JAMA Neurol. 2019;76(6):672–681.
Jehi L, Morita-Sherman M, Love TE, et al. Comparative effectiveness of stereotactic electroencephalography versus subdural grids in epilepsy surgery. Ann Neurol. 2021;90(6):927–939.
Mullin JP, Shriver M, Alomar S, et al. Is SEEG safe? A systematic review and meta-analysis of stereoelectroencephalography-related complications. Epilepsia. 2016;57(3):386–401.
West S, Nolan SJ, Cotton J, et al. Surgery for epilepsy. Cochrane Database Syst Rev. 2015;(7):CD010541.
Wiebe S, Blume WT, Girvin JP, Eliasziw M. A randomized, controlled trial of surgery for temporal-lobe epilepsy. N Engl J Med. 2001;345(5):311–318.
Jobst BC, Cascino GD. Resective epilepsy surgery for drug-resistant focal epilepsy: a review. JAMA. 2015;313(3):285–293.

🔒

Continue reading — sign in

The full note has more clinical pearls, tables, and board-focused tips. Free account, no fee.