01Core principlesThe concepts and mechanisms needed to understand the subject.
Epilepsy surgery assessment asks two separate questions: can the epileptogenic network be localised sufficiently, and can it be treated with an acceptable functional risk? A structural lesion may support the hypothesis but can be incidental; conversely, MRI-negative focal epilepsy can sometimes be localised by converging electroclinical and functional evidence.
Resective or ablative treatment aims at seizure freedom when a removable focus can be defined. Vagus-nerve, deep-brain or responsive stimulation usually aims to reduce seizure burden when resection is unsafe or localisation is distributed. Outcomes vary with pathology, concordance, procedure and duration, so counselling uses centre-specific evidence and individual uncertainty.
Key points
- Refer drug-resistant epilepsy for consideration of resective-surgery assessment even when routine MRI is normal; referral opens evaluation and does not guarantee an operation.
- Drug resistance generally means failure of two tolerated, appropriately chosen and used antiseizure medication schedules to achieve sustained seizure freedom.
- NICE also supports early tertiary assessment when MRI abnormalities predict a high risk of drug resistance; years of additional seizures are not a prerequisite.
- Presurgical work-up builds a concordant hypothesis from semiology, video EEG, epilepsy-protocol MRI, neuropsychology and, when needed, functional or metabolic imaging and intracranial recording.
- The team weighs chance of seizure control against language, memory, motor, visual and psychosocial harms, then incorporates the person’s goals and tolerance of uncertainty.
- A normal MRI, bilateral EEG abnormality, psychiatric comorbidity or learning disability should not be used casually to block referral; each changes the evaluation and support required.
- If a safe resection is not available, the outcome may be further diagnostic work, medication optimisation, dietary therapy or palliative neuromodulation rather than no treatment.
02Mechanisms and patternsImportant relationships and how to distinguish them.
Failure of two tolerated, appropriately selected and adequately used medication schedules to achieve sustained seizure freedom should trigger tertiary surgical evaluation without an arbitrary further delay.
Absence of a lesion on routine MRI does not exclude a surgically remediable focus and must not prevent referral for specialised imaging and electroclinical review.
An MRI abnormality associated with likely drug resistance can justify early tertiary review before multiple years of recurrent seizures.
Referral is for multidisciplinary assessment; resection requires sufficiently concordant localisation, acceptable functional risk and informed patient preference.
Status epilepticus, new persistent neurology or serious injury is treated immediately and should not wait for an elective presurgical pathway.
03Interpreting evidenceInformation, measurements and their limitations.
Consider the information, its meaning and its limitations before deciding what follows.
- 01
Prolonged video-EEG telemetry - Why
- Capture habitual events with simultaneous behaviour and electrical activity, confirm whether they are epileptic and contribute to onset localisation.
- Interpretation and limitations
- Scalp ictal onset is integrated with semiology and imaging; deep or rapidly propagating seizures may be poorly localised and a captured non-epileptic event does not exclude coexisting epilepsy.
- 02
Epilepsy-protocol MRI reviewed by specialist neuroradiology - Why
- Detect subtle focal cortical dysplasia, hippocampal abnormality, tumour, vascular lesion or acquired scar and define surgical anatomy.
- Interpretation and limitations
- A lesion is persuasive when it matches semiology and EEG; a normal scan does not close assessment and an incidental lesion should not dictate resection.
- 03
Neuropsychological assessment - Why
- Characterise memory, language, executive function, mood, learning and lateralising patterns while establishing a functional baseline.
- Interpretation and limitations
- Results contribute to localisation and predict cognitive risk, but are not a stand-alone truth test or a reason to deny supported patients referral.
- 04
Functional mapping and advanced localisation - Why
- Map language and motor networks and refine discordant or MRI-negative hypotheses using fMRI, PET, SPECT, MEG or intracranial EEG when indicated.
- Interpretation and limitations
- Each modality has spatial and temporal limits; invasive recording is justified only when a specific testable hypothesis can change treatment.
04Applied reasoningWorked examples connecting principles to decisions.
01Worked caseFocal seizures after two appropriate regimensDisabling focal seizures persist after two tolerated, appropriately chosen and adequately used medication schedules, with or without an MRI lesion.+
- 1Verify seizure diagnosis, habitual semiology, adherence, dose and duration, medication appropriateness, adverse effects and the person’s goals; address immediate safety and rescue planning.
- 2Refer promptly to a tertiary epilepsy service for consideration of surgical assessment rather than waiting for additional routine drug failures or a visible MRI lesion.
- 3Build a concordant hypothesis from video EEG, epilepsy-protocol MRI and neuropsychology, adding functional imaging or intracranial recording only when it answers a defined decision.
- 4At conference, compare resection or ablation, neuromodulation and non-surgical options using seizure, cognitive, neurological and psychosocial outcomes.
02Discordant dataMRI, EEG and semiology do not alignA lesion is absent, multiple abnormalities exist or electrophysiology points away from structural imaging.+
- 1Re-review habitual events and source imaging with epilepsy specialists; check whether the lesion could be incidental or the scalp onset misleading.
- 2Use targeted advanced tests to resolve a hypothesis, not an indiscriminate battery that accumulates conflicting findings.
- 3If uncertainty prevents safe definitive treatment, communicate it clearly and choose further observation, invasive testing or palliative therapy according to likely value and burden.
03Functional-risk routePossible focus near eloquent cortexThe suspected epileptogenic zone overlaps language, memory, motor or visual networks.+
- 1Map function with the appropriate non-invasive and, if necessary, invasive methods while establishing baseline cognition.
- 2Compare the predicted seizure benefit with specific deficits and the person’s work, education, caring and independence priorities.
- 3Consider tailored resection, ablation, neuromodulation or continued medical care rather than framing the decision as total resection versus abandonment.
05Checking understandingVerify the reasoning, revisit uncertainties and apply feedback.
- While awaiting evaluation, track seizure type, frequency, duration, recovery, injuries, triggers, adherence and rescue-medication use using a practical diary.
- Review driving, bathing, heights, machinery, occupation, pregnancy plans, contraception interactions and sudden-unexpected-death risk according to the individual’s seizures.
- After surgery, monitor seizures and auras, new neurological deficit, language, memory, mood, medication effects, wound and functional reintegration; do not stop medicines abruptly.
- Assess outcomes over an appropriate interval because early postoperative seizures, medication taper and long-term recurrence have different meanings.
06Special situationsVariants, exceptions and circumstances that change the usual approach.
Refer before deciding
The tertiary team determines candidacy; non-specialists should not exclude assessment because the scan is normal or the case appears complex.
Concordance carries weight
No single test identifies an epileptogenic zone reliably in every patient, so converging independent evidence improves confidence.
Function defines resectability
A well-localised focus may still be unsafe to remove if expected language, memory, motor or visual harm outweighs benefit.
Non-resection is an outcome
Neuromodulation, targeted medicines, diet or clarified non-epileptic events may make the evaluation valuable even without resection.
07Common pitfallsFrequent interpretation and management errors.
- 01
Requiring an MRI lesion before referral contradicts NICE and can deny MRI-negative patients a useful specialist assessment.
- 02
Equating two prescriptions with drug resistance without checking diagnosis, selection, adherence and adequate use misclassifies potentially treatable pseudo-resistance.
- 03
Treating a conference referral as consent for resection collapses localisation, functional-risk and preference decisions into one step.
- 04
Promising seizure freedom or medicine withdrawal ignores procedure-specific uncertainty and the need for supervised postoperative management.