Article In Brief
Researchers say the Thrombectomy for Stroke with Large Infarct on Non-contrast CT (TESLA) trial offers true-to-life findings since patients were selected using simpler and lower-cost imaging that is available just about everywhere, compared with other modalities, such as the pricier and less-accessible CT perfusion imaging. Regardless of the imaging modality or inclusion criteria, however, the trial and others like it have found that thrombectomy tends to produce better functional outcomes in patients with stroke.
Patients with a large infarct stroke evaluated exclusively with a commonly available imaging modality—non-contrast CT—and for whom a relatively longer time elapsed before randomization still showed promising results from thrombectomy compared with patients who received usual medical care, according to a Sept. 23 report in JAMA.
The findings, based on the Thrombectomy for Stroke with Large Infarct on Non-contrast CT (TESLA) trial, didn’t meet the formal threshold for superiority of thrombectomy, but researchers and independent experts told Neurology Today they offer the latest encouraging data on the efficacy of thrombectomy even for patients deemed to have large-core infarcts. They noted the data offer perhaps the most true-to-life findings, since patients were selected using a simpler and lower-cost imaging that is available just about everywhere compared with other modalities, such as the pricier and less-accessible CT perfusion imaging.
“The primary difference amongst these large-core trials is in the way that the patients were selected, particularly the imaging approach that was used,” said Albert Yoo, MD, PhD, a co-primary investigator and medical director at the Texas Stroke Institute. “Our trial used a simple non-contrast CT scan to identify these patients. And just by virtue of that selection, the trial is more generalizable because non-contrast CT imaging is what’s done mostly in clinical practice at stroke centers, and it’s available at every stroke center.”
Other trials have tended to employ a combination of imaging modalities in their selection of patients, including CT perfusion, non-contrast CT, and MRI. The recent TENSION trial, which stopped early because of the efficacy of thrombectomy over usual care, used primarily—but not exclusively—non-contrast CT, and thrombectomy had to be completed within 12 hours.
Regardless of the imaging modality or inclusion criteria, the trials have found that thrombectomy tends to produce better functional outcomes. The SELECT2, TENSION, and ANGEL ASPECT trials were all stopped early to because of efficacy, and the RESCUE Japan trial also found superiority for thrombectomy.
The TESLA trial included patients with an ASPECTS score of 2 to 5, which is a somewhat wider inclusion criteria compared to the other trials, in which patients were eligible if they had scores of 3 to 5 or a qualifying infarct volume. Researchers randomized 152 patients to receive thrombectomy and 148 to usual care. Patients had to be treated within a 24-hour time window, and the median was 10.9 hours for the thrombectomy group, the longest duration of the trials that have examined thrombectomy for large infarcts.
The 90-day utility-weighted modified Rankin Scale (UW-mRS) score was 2.93 for the thrombectomy group and 2.27 for the control group, with an adjusted difference of 0.63 (95 percent credible interval, -0.09 to 1.34). The posterior probability for superiority of thrombectomy was 96 percent, just below the pre-specified threshold of 97.5 percent.
The UW-mRS score assigns varying weights to each mRS score disability level to capture patient and clinician judgments regarding the relative value of each level. The statistical power of the UW-mRS and mRS is similar, but researchers said the UW-mRS score was used in the trial to incorporate a patient-centered outcome. Also, they said, an established minimum clinically important difference for UW-mRS helps with the interpretation of treatment effects.
Researchers found no significant difference in 90-day mortality—35.3 percent for the thrombectomy and 33.3 percent for controls—or in 24-hour symptomatic intracranial hemorrhage—4 percent vs. 1.3 percent.
Dr. Yoo noted that the 95 percent confidence interval for the change in functional score excludes a clinically relevant harm from thrombectomy but shows results indicative of benefit.
Osama Zaidat, MD, MS, FAAN, co-primary investigator for TESLA and neuroscience and stroke medical director at St. Vincent Mercy Hospital in Ohio, said that even if the sample were expanded—such as by pooling the non-contrast CT patients with those from the TENSION trial—the message would likely remain the same.
“We don’t think it’s going to change what TESLA showed,” he said. “We know it’s going to be either stronger or the same trajectory that there is a benefit in large-core infarct.”
“Our study, despite the lack of formal statistical significance, I think provides reasonable support for those centers that wanted to select large-core patients with CT alone through 24 hours,” Dr. Yoo said. He said the next step for the research team will be to assess the role of brain edema as a measure of the tissue damage severity “to perhaps better predict the treatment response in these large infarct patients.”
Dr. Zaidat added, “When you have more brain to save, you’ll have a higher return with patient outcome. However, when you have large [infarct], you’re still better than not treating, so it becomes resource-related, it becomes a patient and individual family decision. But at least you would have some evidence to back you up.
“I think that subgroup analysis across our TESLA trial and other trials will focus on the threshold when not to treat, because it’s becoming hard to exclude patients from thrombectomy. If there is a strong trend like in TESLA, with 96 percent probability of improvement in patient outcome, versus not doing anything, then when do you stop?”
Independent Experts Weigh In
In an accompanying editorial, three Cooper University Health Care clinicians—Jane Khalife, MD, MS, assistant professor of neurology; Daniel Tonetti, MD, MS, director of the acute stroke and neurointerventional suite; and Tudor Jovin, MD, chair of neurology—applauded the pragmatism of the TESLA criteria.
“The findings of TESLA leave open the possibility that a simple imaging selection strategy may be sufficient over the entire spectrum of 0 to 24 hours,” they wrote, adding, “The benefit of thrombectomy during the extended 12- to 24-hour window remains not fully established.”
They also said the findings leave open questions of whether more advanced imaging is required in the later time window and whether octogenarians with large infarcts should be considered for thrombectomy, as they have not been well-represented in recent trials.
Andrew Demchuk, MD, director of the Calgary Stroke Program for Alberta Health Services and professor of clinical neurosciences at the University of Calgary, said that although TESLA was a neutral trial in the strictest sense, the findings don’t shift the trajectory of the message from the recent large-core infarct trials.
“This takes nothing away from the guideline that will clearly say large-core stroke thrombectomy is efficacious,” he said. “This doesn’t take away from that at all. What it does, however, is shed a bit of light into where the issues still lie around understanding this entire patient selection better. On the later time window and the bigger strokes, especially, we need more data. And the effects here were smaller probably because the strokes were a bit bigger and more importantly the strokes were treated later.”
“Our study, despite the lack of formal statistical significance, I think provides reasonable support for those centers that wanted to select large-core patients with CT alone through 24 hours.”—DR. ALBERT YOO
He said the criteria and methods add value to these findings, since advanced software was not used.
“Many centers around the world cannot afford that licensed software, so to use automated software is a luxury,” he said. “And many of the other trials did use some of that information for patient selection. So the fact that this didn’t require that is more real-world, pragmatic, and realistic when it comes to a global viewpoint on the subject.”
He said the findings suggest there are questions around “time, size, and severity of change that we need more understanding on to better zero in on treatment decisions.”
At his center, thrombectomy tends to be offered to patients if they have low to medium ASPECTS scores and are pre-morbidly functioning well, preferably in an earlier time window—following the general criteria in the trials.
“Where it becomes really challenging is when patients have some significant comorbidities or when we’re getting into a later time window and the extent of changes are quite large,” he said. “Then it becomes a more difficult decision. I’m very much looking forward to seeing some of these secondary analyses of the pooled data over the next year or so. I think that will shed light into this and help us with those decisions.”
“This trial is appreciated for reliance on one of the most accessible imaging modalities—a non-contrast CT. And the implications of these findings may have greater impact on particular patient populations.”—DR. AMANDA JAGOLINO-COLE
He added that the ASPECTS is only a semi-quantitative score, with some regions that are small in volume being assigned a point, while other regions with larger volume also assigned a point. That has resulted in a wide range of infarct volume—in the 30 cc range to well over 100 cc—to be considered large infarcts across the recent trials, he said.
“There must be a point of no benefit where the changes on the CT are severe—quite hypodense—or when you have a very large infarct on baseline imaging—200 cc or more—that could be an ASPECTS of 3 still, in some circumstances. So I’ll be very interested to see how efficacious the intervention is when they’ve been able to break down severity of ischemic change, size of initial infarct, and time from onset across the large population of patients over multiple trials.”
Amanda Jagolino-Cole, MD, FAAN, associate professor of neurology with McGovern Medical School at UTHealth Houston, said her center all too frequently sees patients with large infarct due to large vessel occlusion, often because of delays in presentation of hospital transfers. Her center uses a “case-by-case approach” in determining whether thrombectomy is appropriate, incorporating data from trials and what they see on clinical presentation and imaging, she said.
She noted the findings from TESLA, for which two of the more than 70 authors were from UTHealth, support further consideration of thrombectomy for these patients.
“This trial is appreciated for reliance on one of the most accessible imaging modalities—a non-contrast CT,” she said. “And the implications of these findings may have greater impact on particular patient populations,” such as those with prolonged transfer times because of geography, weather, or system-based delays.
She hopes future research can yield better understanding of optimal time windows and optimal workflow for these patients.
“I look forward to incorporation of large infarct thrombectomy trials into guidelines,” she said. “This will help clinicians more broadly apply trial findings into their individual practice settings.”