Endovascular Therapy for Acute Stroke Still of Unproven Benefit: The Quest Continues

Abstract & Commentary

By Dana Leifer, MD, Associate Professor of Clinical Neurology, Weill Cornell Medical College

Dr. Leifer reports no financial relationships relevant to this field of study.

Synopsis: CT and MRI parameters were identified that appeared to predict final infarct volume and clinical outcome of acute stroke patients. The parameters were used to develop a model to identify patients most likely to benefit from acute treatment with endovascular arterial therapies. However, several very recent trials, reported at the 2013 International Stroke Conference, failed to find benefit for endovascular treatment using this model and other protocols.

Source: Kidwell C, et al. Multiparametric MRI and CT models of infarct core and favorable penumbral imaging patterns in acute ischemic stroke. Stroke 2013;44:73-79.

Intravenous (iv) or intra-arterial treatment with tissue plasminogen activator (tPA) and mechanical clot extraction can limit the damage that some ischemic strokes cause, but they also can cause complications, including hemorrhage into established infarcts. The length of the interval between the time when a patient was last known to be normal and the time of treatment is currently a key factor in patient selection, but a more reliable way to identify patients who would benefit from treatment is needed, especially for endovascular procedures. Some patients, within a given time window, have completed infarcts while others outside of the window may have a large penumbra — an area of ischemic tissue that is not infarcted and that potentially could be salvaged.

To identify imaging parameters predicting response to therapy, Kidwell and colleagues studied acute stroke patients with intracranial occlusions of large arteries in the anterior circulation on CT or MR angiography. Prior to treatment, the patients also underwent CT perfusion scans or MRI with diffusion and perfusion-weighted imaging to identify regions of infarction and of salvageable ischemia. Follow-up imaging was performed to define the final infarct. The investigators derived models to predict the fate of individual voxels and combined these to generate estimates of final infarct volume and to predict the clinical outcome. Good outcome was defined as a modified Rankin Scale of 2 or less.

The MRI model predicted infarction at the voxel level with 71% accuracy and the CT model with 79% accuracy. Analysis of the models suggested the optimum cutoff points for predicting good clinical outcomes were an infarct volume of < 90 mL and a ratio of predicted infarct to tissue at risk of < 0.7. Using these cutoffs, the MRI model predicted good outcomes with 85% accuracy and the CT model with 78% accuracy in the cohorts used to derive the model.

Kidwell and colleagues then tested these models in the Mechanical Retrieval and Recanalization of Stroke Clots Using Embolectomy (MR RESCUE) study. They presented the results recently at the International Stroke Conference (ISC), where two other randomized trials of endovascular treatment of acute stroke also were presented.

MR RESCUE enrolled patients with intracranial internal carotid or middle cerebral artery occlusions who could undergo CT perfusion imaging or MRI and be treated by embolectomy within 8 hours from the time when they were last known normal. Patients who were eligible for IV tPA were treated with it prior to enrollment and underwent imaging to look for a penumbra after completion of the IV tPA.


Despite the promising imaging results that provided the foundation for the trial as discussed above, MR RESCUE failed to show any benefit for mechanical embolectomy over medical management even in patients with a large penumbra. Patients with a penumbra did better than those without one whether treated medically or by mechanical embolectomy. One possible explanation is that embolectomy was done relatively late with a mean time of 6.4 hours from when patients were last known to be normal to the start of endovascular procedures.

The Interventional Management of Stroke-III (IMS-III) trial, which was also just reported at the ISC, treated patients more quickly. Endovascular procedures were completed in an average of 5.4 hours from the time patients were last known to be normal. Nevertheless, IMS-III still found no overall benefit for combined IV tPA and endovascular treatment compared to IV tPA alone. Secondary analysis, however, showed that there was a trend toward benefit from endovascular treatment if IV tPA was started within 2 hours and if the time from starting IV tPA to groin puncture was < 90 minutes. In addition, there was an estimated 14% relative risk reduction per 30-minute decrease in time to completion of the endovascular procedure, so faster treatment may make a difference.

The Local versus Systemic Thrombolysis for Acute Ischemic Stroke (SYNTHESIS Expansion) study, also just reported at the ISC, did not find an overall difference between combined IV tPA and endovascular treatment vs IV tPA treatment alone, even though endovascular treatment was begun a median of only 3.75 hours from stroke onset — only 1 hour later than the median time for starting IV tPA in the study.

Taken together, the results of these three newly reported studies suggest that very rapid treatment will be needed to demonstrate a significant benefit from endovascular treatment. In view of these results and nonrandomized experience suggesting dramatic benefits for some patients from embolectomy, additional randomized studies are needed with newer devices such as stent retrievers that can open arteries more quickly and with higher success rates than the devices used in the trials that were just reported.