Conference Coverage

Radiosurgery found not superior to open surgery for temporal lobe epilepsy


AT AES 2016

HOUSTON – Despite enrollment difficulties that limited the study, a recently completed randomized trial comparing radiosurgery with open lobectomy to treat temporal lobe epilepsy offers some guidance for patients and their physicians.

Radiosurgery’s noninferiority to open lobectomy couldn’t be shown from the ROSE (Radiosurgery or Open Surgery for Epilepsy) trial, but language deficits were similar – and quite small – by 3 years after either procedure. Expected visual field deficits were similar in each procedure as well. However, since the trial didn’t reach its target enrollment, several primary outcome measures could not be fully assessed.

Sign for surgery center at a hospital EyeMark/Thinkstock
In order to try to enroll the 234 cases needed to show noninferiority, the ROSE trial involved a total of 18 sites: 14 in the United States, 2 in India, and 2 in the United Kingdom. In the end, though, a total of 58 patients aged 18 years and older were enrolled and completed the study; 31 received lobectomy, and 27 had radiosurgery.

On the face of it, radiosurgery has significant appeal. Although open resective surgery is effective, there’s still some risk of infection and blood loss, and neuropsychological changes as well as other focal neurologic deficits are seen. Still, the study saw many challenges, but the largest, according to the investigators, was in recruitment. “Patients like to choose,” said Nicholas M. Barbaro, MD, chair of the department of neurosurgery at Indiana University, Indianapolis. Dr. Barbaro, one of several ROSE coinvestigators who presented the study findings at the annual meeting of the American Epilepsy Society, noted that if patients felt that lobectomy was the best choice, then there would be no incentive to enter a trial where they might be randomized to radiosurgery. Also, he said, some patients might be reluctant to be irradiated, fearing short-term or long-term toxicity.

Trial hypotheses and protocols

The ROSE trial aimed to show that stereotactic radiosurgery (SRS) would not be inferior to anterior temporal lobectomy (ATL) in achieving a seizure-free state by months 25-36 post procedure. The lag to response after radiosurgery is about 1 year; seizure freedom, defined as 12 consecutive months with no seizures, was assessed from months 25 to 36 of the study for the primary outcome of seizure freedom.

Investigators also hypothesized that fewer SRS patients would have significant reductions in measures of language function; further, they predicted that patients in both treatment arms would experience improvements in quality of life (QOL), and that QOL would improve as seizure freedom increased. Finally, the trial sought to show that SRS was cost effective, compared with ATL, with the marginal cost-utility ratio dropping below $50,000 per quality-adjusted life-year (QALY).

Patients in the ATL arm received a standard “Spencer” ATL, with adequacy of resection assessed by MRI performed 3 months after surgery. An inadequate resection would have been classified as an adverse event, but all ATL patients had an adequate resection by study criteria, and all those whose histopathology was available (n = 20) had some hippocampal sclerosis.

Patients in the SRS arm had the amygdala and anterior 2 cm of the hippocampus, as well as the adjacent parahippocampal gyrus, irradiated. This resulted in a total treatment volume ranging from 5.5 to 7.5 cc. Patients received 4 Gy to the 50% isodose line, and treatment could involve an unlimited number of isocenters. The brain stem could receive no more than 10 Gy and the optic nerve and chiasm no more than 8 Gy. All treatment plans were cleared by the ROSE steering committee. The SRS patients had some variation in dose and volumes treated, but all were within the approved limits of the study.

Trial outcomes

As expected, the surgery arm achieved rapid seizure remission, while the SRS arm saw a steady increase in seizure-free numbers beginning at about 12 months after surgery. During study months 25-36, 78% of the ATL arm and 52% of the SRS arm were seizure free. “The null hypothesis of inferiority of SRS was not rejected,” said Mark Quigg, MD, professor of neurology at the University of Virginia, Charlottesville.

Most patients in both groups had no or minimal changes in verbal memory, with no significant differences between the groups at 36 months after treatment.

QOL measures improved rapidly for those who received open surgery, and more slowly for those in the radiosurgery arm, a pattern “consistent with the known association between improved seizure control and quality of life,” said John Langfitt, PhD, a neuropsychologist and professor of neurology and psychiatry at the University of Rochester (N.Y.). However, the study was underpowered to show noninferiority of SRS for QOL measures at 36 months.

“There was a preliminary trend toward reduced health care use over time in the open surgery arm,” said Dr. Langfitt, again noting that the earlier seizure control achieved in surgery reduced health care utilization for that group sooner than for the SRS group. “The power may be limited by sample size and the tendency of utilization to be highly skewed,” he said.

Also as expected, the ATL arm saw early surgery-related adverse events such as scalp wound infections, subdural hematomas, and deep vein thromboses. These were infrequent overall. In contrast, the SRS group saw more cerebral edema–related adverse events during months 9-18, with headaches, new neurologic deficits, and transient seizure exacerbation.

All but three patients received postoperative visual field testing. Of the patients receiving SRS, 34% (10 of 29) had an upper superior quadrant visual field defect, as did 42% (11 of 26) of patients in the ATL arm.

Since the primary treating surgeon and neurologist could not be blinded as to study arm, another neurologist who was blinded was responsible for assessing the outcome measures, and also could identify adverse events. The trial’s steering committee was also blinded to ongoing outcomes.

Pilot study results

A pilot study had previously found that SRS was comparable to the efficacy that had been seen in larger, prospective trials of open surgery, with about two-thirds of patients seizure free at 36 months. Although most patients experienced brief exacerbation of auras or complex partial seizures after radiosurgery, visual field defects were similar to those experienced by patients undergoing standard ATL. Overall, neuropsychological outcomes for those undergoing SRS in the pilot were good, with a low incidence of declines in language and verbal memory function of the dominant hemisphere, and no short-term affective changes were seen. SRS patients who were seizure free after the procedure experienced a significant improvement in QOL.

The promising pilot results contrasted with the limited findings of the ROSE study. In regard to seizure freedom in ROSE, said Dr. Quigg, “The data appear to show that radiosurgery is inferior to ATL, but the low power of the study means that we cannot conclude this with sufficient confidence. Nor can we conclude that the two treatments are noninferior.”

The study was partially funded by Elekta, the manufacturer of the Gamma Knife radiosurgery device used in the study. Dr. Barbaro reported no other disclosures. Dr. Langfitt reported being a consultant for Monteris. Dr. Quigg reported being an investigator for several antiepileptic drug trials sponsored by pharmaceutical companies.

On Twitter @karioakes

Next Article: