ARTICLE IN BRIEF
In a new study, researchers reported that the number of abnormal connections in the brains of patients with temporal lobe epilepsy was inversely proportional to the likelihood of achieving seizure freedom after surgery.
The complete wiring diagram of a patient's brain — their connectome — can be used to predict with a high degree of specificity the success of surgery for temporal lobe epilepsy (TLE), according to a new study that appeared in the April 8 online edition of Neurology.
Researchers showed that the number of abnormal connections in the brain was inversely proportional to the likelihood of achieving seizure freedom, a result that, while suspected, has never before been demonstrated in patients.
The need for better prognostic indicators is strong, the lead study author Leo Bonilha, MD, PhD, an associate professor of neurology at the Medical University of South Carolina in Charleston, told Neurology Today. Even with the best current indicators, “we are still not able to predict seizure freedom in 30- to 40-percent of patients,” he said, a fact that contributes to the low utilization of what can be, for many patients, a cure for medically intractable seizures.
While the best outcomes are seen in TLE patients with identifiable lesions, “even some patients who appear to be optimal surgical candidates,” based on clinical criteria, “are not able to achieve seizure freedom, and we don't know why.”
Connectomics offers a possible answer, Dr. Bonhila said. With advances in diffusion tensor MR imaging, the brain's connections can be mapped in unprecedented detail, allowing the generation of extraordinarily fine maps of links between cortical and subcortical gray matter. Automated computer analysis can then compare the sets of connections between two brains, determining the number of connectional differences between them.
In 2013, Dr. Bonilha and colleagues reported in Neurology that as a group, patients with refractory TLE had more aberrant connections than a group of healthy controls, with the exact pattern of aberrant connections differing from one patient to the next. That study also uncovered some specific wiring differences that were more common in those whose surgeries were unsuccessful.
But group-level differences may not be as valuable for counseling the individual patient as an individualized assessment of risk, based on the patient's own set of abnormal connections, Dr. Bonilha hypothesized, so he set out to map those connections in 35 patients who had undergone anterior temporal lobectomy.
The essential technique, he explained, began with a standard high-resolution diffusion tensor image, which, when processed appropriately, revealed the number of “streamlines,” or white matter connections, between 82 cortical regions of interest. The data was then transferred to an 82x82 matrix, with the strength of connectivity between two regions is given as a value in the cell where those two regions connect on the matrix. In this form, the data can be compared between two individuals.
By comparing an individual patient's matrix to a matrix containing the average of a set of healthy controls, the location, number, and strength of aberrant connections could be determined. “For this study, we used the number of normal or abnormal connections as an index to predict the individual's ability to become seizure-free following surgery,” Dr. Bonilha said.
He observed that freedom from seizures for at least one year following surgery was most likely when patients exhibited fewer connection abnormalities, especially in a brain subnetwork composed of the ipsilateral hippocampus, amygdala, thalamus, superior frontal region, lateral temporal gyri, insula, orbitofrontal cortex, cingulate and lateral occipital gyrus.
Assessing sub-networks with 30 links, including connections among these structures, Dr. Bonilha found there was a dramatic effect on the likelihood of surgical success when the number of aberrant links rose. With no aberrant links, the probability of success was about 85 percent. With up to three aberrant links, it remained high at about 75 percent, but with six or more aberrant links, it fell to less than 20 percent.
The specificity of the aberrancy test alone was 90 percent. By combining those results with clinical data, including duration of disease and presence of hippocampal atrophy, the specificity rose to 94 percent.
“These findings resonate with the hypothesis that epilepsy is a systems-level network disease,” Dr. Bonilha said, “and that seizures arise because a patient's network is active in an abnormal way. This corroborates this idea.”
He cautioned, however, that this retrospective, single-center study will need to be replicated prospectively at other sites before it can be considered for clinical use. Assuming further testing bears out these initial findings, he said, “it would be relatively easy to implement” as a clinical tool, since the measurement of neural network abnormalities could be derived from standard presurgical imaging, with automated computational analysis.
Dr. Bonilha was emphatic about the fact that the purpose of such an analysis would be to inform patients about their likely prognosis, not to screen out poor surgical candidates. “Today, people are reluctant to have surgery,” he said, in part because of the uncertainty of success. Lessening that uncertainty is the goal. “This is a method that should help people gain access to a curative treatment sooner rather than later.”
“To have something to be able to tell a patient about their chances of a seizure-free surgical outcome is beneficial for patients who are trying to decide whether or not to have surgery,” said Jerome Engel, MD, PhD, FAAN, a professor of neurology and director of the Seizure Disorder Center at the David Geffen School of Medicine at the University of California, Los Angeles.
One of the features of Dr. Bonilha's method that distinguishes it from a simple history-based screening tool is that it is likely to be used only in an epilepsy center, not in general practice. And that's a good thing, Dr. Engel said. “Anything that can indicate who is a good or bad candidate that would be used in the community, before the patient comes into the epilepsy center, runs the risk of eliminating patients for referral who ought to be referred. It's important to recognize that less than 1 percent of patients with refractory epilepsy ever get referred to an epilepsy center.”
Whether or not a patient is ultimately a good candidate for surgery, he said, the expert consultation may offer other advantages, including confirming or changing the diagnosis, offering other treatments, or helping the patient deal with psychosocial issues.
The importance of the connectome approach would be even greater, he said, “if it could help to understand the reasons why patients are failing standard surgery for temporal lobe epilepsy.” That may come from a deeper look at the individual patient connectome.
“Ideally, this type of research ought to provide information about how to alter the resection for those patients who are predicted to not do well with standardized resection,” Dr. Engel said. “If we understood the networks involved, we might be able to tailor the resection differently.”
“This is an interesting and thought-provoking paper,” said Jerzy P. Szaflarski, MD, PhD, FAAN, a professor of neurology and director of the Epilepsy Center at the University of Alabama at Birmingham. The link between aberrant wiring and surgical outcome “was something to be expected,” he said, but this is the first study to show it in this way. “I hope they are able to replicate these findings, because if they can, it will be a major step forward for evaluation of patients with temporal lobe epilepsy.”
Dr. Szaflarski concurred that this analysis should not be used to disqualify patients from epilepsy surgery, but rather to counsel them on the likelihood of success. “This is not a tool to use for exclusion, but rather to evaluate the chances for seizure freedom. It's a great promise that we will be able to tell our patients with a little more certainty.”
EXPERTS: ON CONNECTOMICS AS A BIOMARKER FOR EPILEPSY SURGERY OUTCOMES