Article In Brief
In the phase 2 SELECT study, investigators reported that CT together with CT perfusion helped predict which patients would benefit from endovascular thrombectomy after acute ischemic stroke. Independent experts raised questions about the study design and said they looked forward to more information to come from the SELECT2 trial.
CT imaging and CT perfusion (CTP) accurately predicted which stroke patients would benefit from endovascular thrombectomy, but when used together they were even more accurate, according to a report of a multicenter study published in the March issue of Annals of Neurology.
“The SELECT trial is the first large prospective cohort study evaluating the influence of imaging modalities on thrombectomy treatment decisions and outcomes. There is currently no randomized evidence on which imaging modality is superior or should influence one treatment decision more than the other,” lead author Amrou Sarraj, MD, associate professor of neurology at McGovern Medical School at UTHealth told Neurology Today.
“Our results also showed that in a phase 2 study, there is a potential benefit from thrombectomy in patients with discordant imaging profiles (favorable on one modality and unfavorable on the other), which has not been shown before, since those patients were not enrolled in prior thrombectomy trials,” he added.
Study Details, Findings
As part of the phase 2 SELECT trial, Dr. Sarraj and colleagues analyzed data on 361 patients with ischemic stroke and anterior circulation large vessel occlusion who presented to nine US study sites within 24 hours from the time they were last known to be well between January 2016 to February 2018.
Patients received a pre-specified and unified imaging evaluation which consisted of CT, CT-angiography, and CTP. The treatment decision, endovascular thrombectomy (EVT) vs medical management, was non-randomized and left to the physicians' discretion.
In addition, the researchers used pre-set parameters for CT and CTP that if met, meant patients had favorable imaging profiles. Dr. Sarraj explained that a favorable profile demonstrates limited ischemic changes and a small stroke size, thus a higher likelihood of benefit following thrombectomy. The favorable profiles can be observed on both non-contrast CT, defined as an ASPECTS score of 6 or more, and on CTP imaging, defined as ischemic core volume of less than 70 cc with the presence of mismatch volume of at least 15 cc and a mismatch ratio of ≥1.2, he noted. The definition of favorable profiles are based on the imaging inclusion criteria from prior trials that proved thrombectomy efficacy and safety.
Dr. Sarraj said that an unfavorable profile means the imaging showed significant ischemic changes and a large infarct size (ASPECTS score of 0-5 on CT and ischemic core >70cc on CT-perfusion), findings that are usually associated with worse outcomes, even after thrombectomy. The authors noted that unfavorable profiles were largely excluded from prior trials, thus untested in a prospective manner as was the case in SELECT.
Of the 361 patients enrolled in the SELECT study, a significant proportion of patients had favorable imaging results on both the CT and CTP scans. Those patients also had significantly higher odds of receiving EVT and higher 90-day functional independence rates after recovery (58 percent).
Favorable profiles on the two modalities correlated similarly with 90-day functional independence rates (p=0.41). Having a favorable profile on both modalities also significantly increased the odds of receiving thrombectomy as compared with having discordant profiles (p<0.001).
Fifty-eight percent of the patients with favorable profiles on both modalities achieved functional independence compared with 38 percent of those with discordant profiles, and 0 percent when both were unfavorable (p<0.001 for trend). In favorable-CT/unfavorable-CTP profiles, EVT was associated with high symptomatic intracerebral hemorrhage (24 percent) and mortality rates (53 percent), according to the report.
Among patients who received EVT, 87 percent had a favorable CT profile as compared to 91 percent with a favorable CTP profile. In 81 percent (n=231) of the thrombectomy patients, there was an agreement between CT and CTP profiles, with both having a favorable profile. In all, just 3 percent of the patients (n = 9) had unfavorable profiles on both modalities and 16 percent (n = 45) had a discordant profile. In patients treated with medical management alone, 29 percent (n = 22) had a discordant profile. All images were assessed by an independent imaging core-lab that was blinded to the treatment decision and clinical outcomes.
Dr. Sarraj said that the SELECT 2 trial began enrollment in September of 2019 and is on course for completion by June of 2022. “The trial will provide randomized data regarding thrombectomy efficacy and safety in patients with large ischemic core infarcts on CT or perfusion imaging. It will also provide evidence about whether there is heterogeneity of thrombectomy treatment effects based on different imaging modalities on non-contrast CT and perfusion which will help to guide stroke clinicians on how to best utilize imaging findings in their treatment decision making and outcome prognostication,” he noted.
“The authors are to be commended for tackling a burning question in the field—should patients be selected for mechanical thrombectomy with CT ASPECTS score, CTP core-infarct volumes, or both, and should this be influenced by the time window from stroke-onset (less than or more than six hours)?” said Pooja Khatri, MD, MSc, FAAN, professor of neurology at the University of Cincinnati and co-director of the University of Cincinnati Stroke Team.
Dr. Khatri said that currently, the standard of care is to make treatment decisions based on CT alone within six hours of stroke-onset and add CTP beyond six hours from onset.
“It has been assumed that the vast majority of patients within six hours will have favorable CTP parameters, but that additional CTP parameters are needed for patient selection beyond six hours. These assumptions were the basis of trials of thrombectomy to date,” she said, noting that automated post-processing software required for CTP interpretation is very expensive, and “most hospitals in low-middle income countries as well as many in high income countries have not invested in it.”
Dr. Khatri added that The SELECT investigators show that the two imaging modalities match up 80 percent of the time. “In this non-randomized study, those 20 percent with discordant imaging did worse with treatment than those with concordant imaging, especially if the CTP was the unfavorable imaging modality. The big question now is, regardless of how well the patient did or did not do based on imaging prediction, would they have been better or worse off with treatment? Are these imaging biomarkers prognostic or treatment effect modifiers? Only a randomized trial can answer that question, and the subsequent, ongoing SELECT2 trial, which is randomizing patients with unfavorable imaging on CT, CTP or both, seems poised to address this question. It will be important for that study to do prespecified analyses that consider whether time from stroke-onset affects the answer,” she said.
Lee H. Schwamm, MD, professor of neurology at Harvard Medical School, noted that the study seeks to compare selection “between simple and readily available, inexpensive CT of the brain, and use of contrast enhanced imaging with sophisticated image processing software that approximates the amount of tissue with reduced blood flow but that is still salvageable if flow is rapidly restored.
“The design of the study is not ideal to answer this question since the individual sites had access to both types of imaging on every patient before they decided to treat with thrombectomy,” Dr. Schwamm told Neurology Today.
Dr. Schwamm explained that as tissue death progresses, it becomes visible on CT. “The CT is divided into 10 regions in the distribution of the middle cerebral artery, which is commonly affected in strokes needing thrombectomy. Each area that is injured and whose damage is visible on CT loses one point, so a scan with an ASPECTS score that is less than six means half of the regions have already been damaged, so chances of a good outcome with thrombectomy are lowered.” he explained.
The CTP scans can add critical information in patients whose diagnosis is less certain, he added. For instance, CTP scans can supplement the information provided by CT, including when the CT scan is blurry or the patient has prior minor, chronic damage that confounds interpretation, or in patients who arrive very late after stroke onset who have tissue to save. “I don't think the paper changes the approach very much, since all patients get a CT angiogram to look at the blood vessels and confirm the blockage, and a lot of information is also available on that image.”
Dr. Schwamm noted that the disadvantage of the CTP, however, is that it can also be artifactually abnormal, called a ghost infarct core, and in this case, “suggest widespread low flow when there is none, in which case the CT is helpful as it usually looks normal.”
Based on SELECT study findings, “if one relies on either CT or CTP alone, this may exclude patients who might benefit from EVT,” said May A. Kim-Tenser, MD, associate professor of clinical neurology and associate division chief of neurocritical care/stroke at the University of Southern California/Keck School of Medicine.
“This study raises the possibility that the number of successfully-treated patients might be maximized if both imaging modalities are used and patients are treated if either modality shows a favorable profile. Even though patients with discordant imaging had lower rates of functional independence following EVT, the rates are potentially more favorable than patients treated with medical management alone. Randomized, controlled trials are needed to assess the safety and efficacy of EVT in patients with large baseline ischemic core lesions on either CT or CTP and evaluate whether EVT is safe and efficacious in large core infarcts on either modality,” Dr. Kim-Tenser said.
She added that the findings may impact clinical practice in that perhaps even the large-core, low-CT ASPECT patients in the large-core subpopulation can undergo EVT and have higher rates of functional independence in comparison to medical management alone.
The SELECT trials are funded by grants from Stryker Neurovascular.
Dr. Sarraj reports grants from Stryker Neurovascular. Dr. Schwamm serves on the national steering committee of the TIMELESS trial of late window tenecteplase and thrombectomy, sponsored by Genentech for which one of the study authors is PI, reports consulting with Genentech and Medtronic on stroke prevention in atrial fibrillation, and is co-principal investigator of Stroke AF sponsored by Medtronic. He has also presented or organized CME for Boehringer Ingelheim and Medscape on thrombolysis and stroke systems of care. Dr. Kim-Tenser is on the speaker's bureau of Chiesi USA, Inc.