Stroke Imaging Treatment Selection
Stroke Imaging Treatment Selection
Volunteer speaker Michael Lev, MD FAHA FACR, walks through a short video on Stroke Imaging Treatment Selection.
Content
1.14 -> Hi I'm Michael Lev and I'll be talking in this AHA
6.42 -> educational series about Stroke
Imaging for treatment selection.
14.46 -> These are my disclosures.
18.78 -> This is a summary of our stroke Imaging algorithm which has been used at my Institution for over two
26.34 -> decades. And it is designed to answer three basic questions. Question one is there a hemorrhage or
34.62 -> large, well-established infarct that would be a contraindication to IV or endovascular treatment.
41.34 -> Question two using CTA is there a proximal large vessel occlusion that is a target for
49.98 -> endovascular clot retrieval or endovascular thrombectomy. And question three is there a
57.42 -> large well-established infarct core that would be a contraindication to either IV or endovascular
66.48 -> treatment. Of course IV treatment could be with alteplase or tenecteplase within 4.5 hours
73.32 -> and IV could also be used for a wake-up stroke but uh that is um more of a off-label
82.2 -> indication and for triaging late IV treatment we
typically use flare DWI mismatch. You want to see
92.52 -> a relatively normal flare MRI with a positive but
less than 50 to 70 ml diffusion weighted lesion.
102.48 -> And for endovascular therapy in
the under six to eight hour range
107.04 -> the big clinical trials which we'll discuss
momentarily have used infarct core of 70 ml.
114.66 -> For the up to 24 hour endovascular
therapy we've used infarct core up to 50 ml.
125.22 -> Now I should add that endovascular therapy
has been studied in several large clinical
134.16 -> trials, prospective randomized trials
that have showed its benefit. In fact
139.5 -> in 2015 there were five major New England
Journal of Medicine studies -- Mr Clean, Escape,
146.64 -> Extend IA, Swift Prime and Revascat --that not
only established CTA as critical to Patient
157.14 -> selection for the endovascular therapy for
catheter-based stroke treatments but also
163.92 -> the Dawn and Diffuse trial in uh published
in the New England Journal in late 2017
170.22 -> and early 2018 respectively extended
that time window up to 16 to 24 hours.
181.08 -> I think the big take-home message therefore
is that as of early 2018 CTA became the new
190.44 -> required imaging standard in selection of
patients for endovascular thrombectomy and
196.8 -> up through then although we had been doing
CTA, CTA collateral imagingy, for over two
203.4 -> decades the standard of care was really just
a non-contrast CT to rule out a bleed since
210.78 -> that would be an absolute contraindication
to either IV or endovascular thrombectomy.
218.58 -> Now I'd like to start with an example this was
actually a patient who we were able to treat
224.76 -> very quickly because they were visiting another
patient in uh in our hospital when they became
231.6 -> acutely hemiporetic and as you could see here the
symptoms were on the left because the lesion is on
240.12 -> the right and we see a right hyperdense vessel
sign at the top of the internal carotid artery
246.96 -> extending into the proximal middle cerebral artery
MCA N1 segment. Now importantly the non-contrast CT
257.7 -> parenchymal image is not only does not show
a bleed but is normal with respect to gray-
264.3 -> white differentiation in other words there's no
early signs of infarction or vasogenic edema here.
272.22 -> Now the next imaging test in our stroke Imaging
algorithm CT angiography - CTA - confirms that there's
280.02 -> good filling of the left of the basilar
artery posterior cerebral arteries in the
285.6 -> posterior circulation but there is a top of ICA
and proximal M1 middle cerebral artery occlusion.
296.34 -> Now CTA is valuable not only to show those
proximal large vessel occlusions, which are
302.1 -> targets for clot retrieval thrombectomy, but they
show collateral flow which is very important
309 -> prognostically and in decision making as
well as parenchymal flow which is part of
315.54 -> perfusion imaging, CT perfusion imaging. And here
you could see what's called a malignant pattern
321.24 -> where not only are there very poor collaterals
over more than half the territory at risk but
329.16 -> there is also just very low blood flow to that
region with absent capillary filling and more
337.2 -> than half the MCA territory showing severe
ischemic change with very low blood flow.
345.48 -> Now in part this was a good outcome
because the patient was found so early
351.06 -> but we do show here that we had successful
intra-arterial or endovascular thrombectomy.
359.34 -> So that MCA reopened that top of ICA occlusion
was also reopened. But interestingly because
370.32 -> the patient was found so early and because
the non-contrast CT was stone-cold normal,
379.2 -> um despite the fact that there was very poor
perfusion because we caught this patient so
384.6 -> early you could see that the diffusion weighted
imaging post thrombectomy is nearly normal as
391.38 -> was the T2 and flare Imaging so this was almost
a complete save even though there was a large
398.52 -> area of malignant severe ischemia because that
non-controcity showed tissue was still normal
406.2 -> and because we caught the patient so early. Now
importantly in the state of Massachusetts where
413.76 -> I'm from there are only a relatively small number
of hospitals that are comprehensive stroke centers
419.64 -> capable of performing endovascular thrombectomy,
whereas there are over 72 primary stroke centers
427.5 -> that receive stroke patients and are capable of
giving IV Therapy but not endovascular therapy.
435.72 -> And so we've set up this hub and spoke network of
referrals where we will transfer patients to the
443.1 -> hub hospitals. Now again at the hub hospital
where endovascular thrombectomy is performed
449.94 -> it's important to make sure we're looking at
those infant core thresholds. So if you can
455.04 -> get it, diffusion weighted imaging is the gold
standard for measuring the size of the infarct
460.98 -> core. On all axial cross-sectional imaging
its length times width times height over 2.
469.2 -> Now most centers do not have immediate
emergent access to MRI diffusion scanning.
477.72 -> And so there are other ways to measure core
and indeed the 2015 New England Journal of
484.32 -> Medicine EVT trials although they all used
CTA to find the large vessel occlusion all
490.2 -> measured core differently. And at one extreme
the Mr Clean trial used only non-contrast CT
496.14 -> to estimate core. On the other extreme the
Extend-IA trial used CTP to estimate core
502.8 -> and that resulted in a core underestimate
or potential overtreatment versus a core
512.04 -> overestimate and potential undertreatment.
And the Escape trial out of Calgary um used
519.96 -> CTA collaterals to measure core and um basically
it really was kind of a compromise between these
527.4 -> two extremes of underestimating core and overtreating or overestimating core and undertreating.
538.26 -> Now because CTA collaterals to estimate the
critically ischemic region is something you
545.28 -> might not all be familiar with I do want to spend
just a couple of moments reviewing how that might
551.88 -> be achieved. And again there's a lot of literature
on this. I am highlighting a paper from our group
560.64 -> and you can see here that when there
are symmetric collaterals going out
566.22 -> to the distal collaterals, that not only
corresponds to a baseline small infarct
571.68 -> volume but there's very small
infarct growth up to 24 hours.
579.18 -> Whereas when there is a malignant infarct pattern
where more than half the MCA territory at risk
587.4 -> shows poor or absent collaterals and where there
is not good parenchymal level capillary perfusion
594.72 -> that not only corresponds to a large initial
infarct at admission but to infarct growth in a
602.4 -> very large infarct on 24-hour follow-up. And you
can have a look at the paper and see that that
609.84 -> really does correspond to this large group
of patients so that's a consistent pattern.
616.26 -> Now the message I want to leave you with
with regard to core imaging is that the
621.84 -> gold standard is diffusion weighted MRI and
if you could get that to measure the core
626.58 -> size accurately and with less variability you
should. However non-contrast CT is also a measure
634.32 -> of core although not as sensitive because it only
becomes positive in the later stages of vasogenic
640.68 -> edema not cytotoxic edema like DWI. In addition
collaterals and perfusion are probabilistic.
650.82 -> They do not measure core per se but they measure
critically ischemic tissue. So a good way to think
657.24 -> about it is that diffusion measures areas where
infarct is likely despite immediate reperfusion,
663.78 -> whereas collaterals and CTP measure areas where
infarct is likely without immediate reperfusion.
672.9 -> So in conclusion we modified our stroke imaging
algorithm that I showed you at the beginning to
680.34 -> take into account um early versus late IV and
endovascular treatment. We focused more on CT, CTA,
691.8 -> CTA collaterals and CTP to um make it useful in
centers for both referral and receiving hub in
702.9 -> stroke in all the different time windows
that have been suggested by the clinical
707.82 -> trials. And since we are a comprehensive stroke
center that does EVT if we do get a patient
715.68 -> that is transferred from an outside hospital we
typically will go to this part of the algorithm
722.7 -> where we will try to get MR diffusion to get a
more accurate measure of uh that core infarct
729.96 -> size before bringing the patient to treatment.
But of course there's a lot of flexibility
734.22 -> here and in patients who do not have large
vessel occlusions who are not candidates for
740.7 -> thrombectomy we will consider doing the flare
DWI mismatch to enroll patients in late IVT.
750.96 -> And I'd just like to conclude by saying that
studies are over way to actually see if for
758.16 -> measuring core non-contrast CT alone without CTA
collaterals or CT perfusion or MR perfusion or
767.52 -> DWI might be sufficient for triaging patients
to endovascular therapy in the late window.
776.04 -> And there was actually just recently in November
2022 from these star collaborators a JAMA Network
784.38 -> open paper that showed comparable outcomes
were observed in patients in the late window
790.2 -> irrespective of neuroimaging selection criteria
admission non-contrast CT scan matriage emergent
797.34 -> laws vessel occlusion in the late window. So that's
being studied. And it's especially exciting to us
803.04 -> because we have recently published in Nature
Scientific Reports in January 2023 a artificial
812.46 -> intelligence model that could actually help
detect stroke better than humans and not only
819.18 -> detect better than this dotted Red Line human
experts but can help delineate stroke and so
826.5 -> has 97 percent specificity and I'm sorry 97
sensitivity and 99 percent specificity
835.98 -> in finding strokes that are bigger
than the 50 to 70 ml threshold that was
843.18 -> in the clinical trials. And so I'll stop there.
Thank you all so much for your kind attention.
Source: https://www.youtube.com/watch?v=mmWctyc5Mcw