Imaging prognostic markers for surgical seizure control in medically intractable temporal lobe epilepsy

Epilepsy is the most common serious brain disorder, affecting over 750,000 people in the UK. The condition can be characterised by devastating seizures, which severely affects a person's quality of life. Epilepsy frequently affects a patient's cognitive and mental health, and the disorder contributes to elevated propensity for depression and suicide compared to the general population. These effects are most serious in patients who have intractable epilepsy (i.e. people who continue to suffer from seizures despite antiepileptic drug treatment). Intractable epilepsy constitutes 30% of all patients with epilepsy, but accounts for 75% of the immense health care costs of the disorder.

If certain criteria are met, many patients with intractable focal epilepsy (where epileptic seizures originate from one region of the brain) may be offered neurosurgery with the goal of removing the part of the brain causing epileptic seizures. This procedure may substantially improve a patient's quality life - and potentially save it. In patients with temporal lobe epilepsy (TLE), which is the most common type of intractable focal epilepsy, neurosurgery will completely stop debilitating seizures after one year of the operation in 70% of patients. This figure reduces as time goes on, and can be as low as 40% 10 years after surgery. Furthermore, surgery may reduce the frequency and severity of seizures, but not eliminate them. It is unknown why one patient seemingly suitable for neurosurgery will have a poor postsurgical outcome and another will have a successful outcome. Undergoing neurosurgery in an attempt to stop epileptic seizures is likely to be the most important decision a patient will make in their lifetime. It is therefore extremely important to have reliable information before surgery that will inform the patient and their doctors of the likelihood of seizure control after surgery so that patient and doctor can make an informed decision on whether to pursue neurosurgery and weigh the potential benefits against risks.

It is the goal of the proposed research to understand the reasons for why some patients deemed suitable for neurosurgery continue to have disabling postsurgical seizures, and to identify the aspects of brain structure and brain impairment that predict whether a patient will respond favourably or unfavourably to surgery (i.e. the development of 'prognostic markers'). A structural brain abnormality suggests that this part of the brain may be causing the seizures. Identifying such an abnormality using MRI scanning will offer reasonable confidence that the surgical removal of the abnormality will stop the seizures from starting. However, the MRI scans routinely acquired during clinical practice in hospitals may not detect subtle brain abnormalities that support epileptic seizures. The brain scanning methods used in this proposal are state-of-the-art and are extremely sensitive to very subtle alterations in normal brain structure, which may support the start and/or spread of epileptic seizures throughout the brain. It is anticipated that the suite of sophisticated analysis tools applied to the high quality brain imaging data obtained before surgery will build a detailed picture for why some patients continue to have seizures after surgery, and that this information can be used to predict whether a patient is likely or unlikely to be surgically rendered seizure free.

This research will have substantial scientific and clinical impact, with a fundamental aim to improve human health and quality of life.
Research objectives will be delivered in the short-to-medium term (3 years).

The aim of this proposal is to identify the neuroanatomical factors that are associated with postsurgical epileptic seizures using high quality presurgical T1-weighted MRI and diffusion tensor imaging (DTI) in a large sample of people with temporal lobe epilepsy and hippocampal sclerosis, and to develop prognostic markers on the basis of these factors to predict postsurgical seizure control.

Several hypotheses will be tested in patients who have undergone amygdalohippocampectomy, which are based on the concepts of (i) aberrant anatomy supporting epileptogenic activity that is not surgically resected, and which remains intact and functionally active postsurgically and (ii) distinct clinicopathological subtypes of patients with TLE. State-of-the-art MRI morphometric and DTI tractography techniques will be used that focus on the structure and connectivity of the affected and contralateral temporal lobes, extratemporal structures, and of the hippocampus deemed presurgically to be sclerotic.

Standardised postsurgical outcome classification at one and three years will be used to compare brain structure, connections and networks between patients (i) rendered free from all seizures, (ii) rendered free from complex partial seizures but with continuation of simple partial seizures (i.e. reduced severity), and (iii) with no worthwhile postsurgical improvement. Comparison will also be made to brain structure, connections and networks in a large sample of healthy controls.

Presurgical neuroimaging data will be combined with patient demographic, clinical and presurgical electrophysiological data to develop the most reliable predictors of postsurgical seizure outcome. In addition to the identification of neuroanatomical factors associated with poor postsurgical seizure control, the validity of a new hippocampal subfield mapping technique will be assessed in relation to quantitative histological assessment of resected specimens obtained from the same patients.

In accordance with MRC guidance, this section emphasises impact for non-academic beneficiaries.

This research proposal directly targets the improvement of human health and quality of life by providing a fuller understanding of neuropathology in epilepsy and predicting patient response to neurosurgery. Epileptologists and neurosurgeons will benefit from understanding the reasons for why a proportion of patients who are deemed suitable for brain surgery to relieve intractable epilepsy continue to suffer from debilitative seizures after surgery. It is in this context that knowledge generated from the proposed work will also benefit the person with epilepsy; patient quality of life after surgery is at its lowest when neurosurgery has failed to stop seizures and has lead to cognitive impairment. Even if cognitive impairment results from surgery, surgical control of seizures will lead to a significant improvement in patient quality of life. Therefore, by providing clinicians and patients with refined information on the likelihood of postsurgical outcome, both can make better-informed judgements regarding the patient's welfare after surgery. This is particularly beneficial for patients who - based on the work generated from this proposal - would be deemed unlikely to respond well to brain surgery, and for whom alterative therapies may be considered (e.g. vagus nerve and deep brain stimulation).

The demonstration of the sensitivity of diffusion tensor imaging (DTI) for determining microstructural alterations of biological tissue that underlie pathophysiology will encourage its application for surgical planning in other neurological and non-neurological diseases. We will also promote the use of quantitative metrics for surgical planning, which is a supplementary diagnostic approach that is underutilised in clinical contexts. Furthermore, the power of imaging to generate predictive markers of human disease and response to therapeutic intervention will be showcased in this research, which are concepts that permeate the whole of medical and surgical practice.

The translation of scientific data to clinical practice may occur soon after the completion of the proposed three-year project. Of particular relevance will be the incorporation of high-quality DTI protocols into UK presurgical epilepsy evaluation programmes, which is not routinely performed. The proposed work will demonstrate the diagnostic utility of DTI, and it's superiority over conventional clinical MRI. Moreover, given the dependence of quantitative DTI approaches on post-acquisition methodology, the proposed work will hopefully stimulate enhanced collaboration between neuroradiologists and basic scientists in presurgical contexts. In particular, sophisticated image analysis work - typically conducted by scientists - will be encouraged to supplement routine neuroradiological diagnostic information - provided by doctors - to provide the most informative presurgical imaging protocol available.

The proposed work may also have a potential benefit for the UK health care economy. Surgeons are more reluctant to operate on patients without an identifiable brain abnormality. If convincing and reliable imaging predictive markers of postsurgical outcome are developed, more patients with no abnormality on clinical MRI may be offered surgery. Given that brain surgery is the most effective method of seizure control, an increasing number of intractable patients undergoing surgery would save NHS costs dedicated to the long-term prescription of antiepileptic drugs (AEDs), which constitutes the main costs associated with treating epilepsy. Almost all presurgical patients with intractable epilepsy are prescribed multiple AEDs, whereas monotherapy is typically used postsurgically. The increasing and earlier use of surgery would therefore save the increased costs and risk of comorbidities associated with polytherapy.