Prof. François Nicoli
The story of the cerebral collateral circulation began a very long time ago, more than a century.1
As early as 1987, an angiographic study performed in patients with a subacute proximal MCA occlusion provided evidence of the clinical impact of the velocity of the retrograde collateral filling. Authors of this study found that if the conduction time of contrast medium, from the intracranial siphon to the insular portion of the MCA (M2) through the anterior cerebral arteries, was higher than 5 seconds, then an extensive CT hypodensity would develop.2 Two years later, Bozzao et al3 demonstrated that the presence of a good collateral circulation during the first 6 hours after a stroke reduced the size of the final parenchymal brain damage in patients with middle cerebral artery stem–trunk occlusion. In 2002, by using a multivariate analysis applied to a cohort of about 100 patients with MCA M1 occlusion mostly treated with intra-arterial thrombolysis, Kucinski et al4 demonstrated that the angiographic degree of collateralization was the only predictive factor of clinical outcome.
The in vivo visualization of the leptomeningeal arterial supply was only possible by using conventional angiography until 1997, the date of the first attempt to visualize the collateral circulation by using CT angiography.5 CTA clot burden score and collateral score were described only 12 years later by Tan et al.6
Thanks to this original method able to noninvasively assess the extent of the collateral circulation, to the growing availability of efficient acute stroke treatments like intravenous thrombolysis (IVT) and endovascular thrombectomy, and to the increasing number of comprehensive stroke centers, the awareness of physicians about the collateral status of patients with acute intracranial arterial occlusion was progressively raised after many years of oversight or indifference. The recent increasing number of publications relative to the impact of baseline collateral circulation on outcome of patients with acute ischemic stroke clearly demonstrates this evolution (Figure).
Key dates in the recent story of collaterals in ischemic stroke.5,6,11,17,27-30
Papers selected for this edition of the AJNR News Digest highlight the intensity of the recent research focused on cerebral collaterals and on either CT-based or MR-based methods able to noninvasively and accurately assess the collateral circulation at the acute phase of stroke.
CT Angiographic Assessment of Collaterals
In 2015, Fanou et al7 showed that the collateral circulation, as assessed by using CTA, was an independent predictor of radiologic and clinical outcomes. The magnitude of significance was greater in patients who did not recanalize versus those who did recanalize. In addition, this predictive value was often increased by using the combination of information relative to the baseline collateral status and the total ischemic volume.7
Interestingly, the same year, Cheng-Ching et al8 demonstrated that the CTA collateral status but not time from stroke onset to imaging was an independent predictor of the size of core infarct in patients with anterior circulation large vessel occlusion presenting within 6 hours from onset. Thus, as suggested by some studies9 and results of several recent positive randomized endovascular trials,10 both baseline parameters, collateral status and core volume, are important to take into consideration for a better selection of patients for endovascular procedures. The same reasoning was also recently formulated in conclusions of the Acute Stroke Imaging Research Roadmap III about imaging selection and outcomes in acute stroke reperfusion clinical trials11 (Figure).
However, results of the CTA collateral assessment depend on the technique used. Indeed, recent studies showed the higher accuracy of multiphase or dynamic CTA versus single-phase CTA for the assessment of collaterals in patients with acute stroke.12–14 In comparison with multiphase or dynamic CTA, single-phase CTA clearly underestimates the collateral circulation. The assessment of collaterals by using multiphase and dynamic CTA provides a more accurate prediction of clinical outcome compared with conventional single-phase CTA.12–14 In addition, dynamic CTA, particularly if assessed in the arteriovenous phase, is superior to conventional CTA in predicting the follow-up infarction volume.13 A great advantage of dynamic CTA over multiphase CTA is that dynamic CTA is constructed from CT perfusion data, which have additional value like the possible estimation of the core volume through the measure of the severely decreased cerebral blood flow and the computation of Tmax maps to evaluate the hypoperfused areas at risk of necrosis.14 Using dynamic CTA, van den Wijngaard et al14 also evaluated the velocity of collateral filling by calculating the duration in seconds of contrast arrival at the ICA until maximal contrast enhancement was achieved for each hemisphere separately. These authors proposed defining a fast filling as an optimal filling within 4.5 seconds after optimal filling in the unaffected hemisphere.
Thus, dynamic CTA is able to provide a noninvasive and accurate characterization of the extent and the velocity of the collateral filling in patients with large intracranial arterial occlusion.
MR-Based Methods for the Assessment of Collaterals
In spite of CT-based methods used to estimate the extent of the ischemic core, like dynamic CTA or the CT perfusion measurement of the area of severely decreased CBF, diffusion-weighted imaging remains the more accurate way to delimitate the ischemic core and diagnose stroke mimics.15 Thus, one must be able to noninvasively and accurately assess the extent and the vigor of the collateral circulation by using a either CT-based method or MR imaging.
With this in mind, different strategies based on MR imaging have been proposed. Some of the most interesting methods are presented here. As for CT, 3 MR-based methods can be used to evaluate the collateral circulation: perfusion-weighted imaging, conventional MR angiography, and dynamic MRA.
MR Perfusion-Weighted Imaging
In comparison with dynamic CTA, multimodal MRI combining DWI and PWI provides more accurate information about the status of brain tissue in patients with acute stroke. However, it is much more difficult to extrapolate information about the collateral flow from MR PWI data. The most interesting methods are those based on a Tmax ratio computed from brain tissue analysis,16–18 those based on a specific pial Tmax assessment,19 those based on an accurate measurement of the arterial-to-tissue delay (ATD) correlated to the retrograde collateral filling,17,20 and those based on subtracted dynamic MR perfusion source images.21
For a long time, a good collateral flow has been known to be able to preserve tissue from ischemia until recanalization occurs. Very interestingly, several studies based on MR PWI recently demonstrated that a good collateral flow is also an independent predictor of recanalization after intravenous thrombolysis or endovascular thrombectomy.9,17,19–21
In 1992, Ringelstein et al22 suggested, for the first time, that a better collateral flow could increase the recanalization rate after IVT thanks to retrograde delivery of rtPA to the distal end of the clot. The first scientific validation of this exciting hypothesis was performed just 20 years later by Nicoli et al using MR PWI.17 To achieve this goal, it was necessary to define an index derived from the baseline MR perfusion imaging in order to noninvasively and indirectly quantify the collateral circulation deficit. Using a PWI-derived collateral flow index based on Tmax maps at different time points, this work provided the first evidence that a good baseline supplying flow significantly increases the rate of full recanalization at 24 hours in patients with acute MCA M1 occlusion treated with IVT within 3 hours after stroke onset.17 This result was confirmed in the same patient cohort by using an accurate measurement of the volume of tissue with severely increased increased ATD.9,17 Similar to results obtained by Saito et al in 1987,2 a volume of tissue with an ATD higher than 6 seconds clearly discriminates patients with good collateral flow from others.9
Very recently, thanks to measurement of the ATD from MR PWI data, Zhang et al20 fully confirmed that the velocity of collateral filling is independently associated with the rate of MCA recanalization at 24 hours after IVT. It is well known that an early MCA recanalization after IVT is an independent predictor of good clinical outcome at 3 months after stroke onset. But, a reocclusion or a late MCA recanalization can also occur within 24 hours after IVT and negatively or positively influence the outcome.17 This probably explains why a full MCA recanalization within 24 hours post-IVT is also a strong predictor per se of good stroke outcome at 3 months after thrombolysis.17 Thus, methods able to predict MCA recanalization 24 hours after IVT, like those described here, provide important information about the individual benefit of IVT during this acute period. Consequently, an aggressive treatment such as endovascular thrombectomy could be more justified if one can guarantee, before any treatment, that acute MCA occlusion would not be recanalized, even 24 hours later, if IVT was performed alone.