1. Dual Energy CT Angiogram of the Head: Quantifying Iodine Concentration in Acute Ischemic Stroke
Mohammed F. Mohammed, Adam Min, Olivia Marais, Rawan Abu Mughli, David Ferguson, Tim O’Connell, Axel Rohr, Savvas Nicolaou
University of British Columbia, Faculty of Medicine, Department of Radiology
RSNA 2017, “Emergency Radiology: Neuroradiology”, SSJ06, Room N22B , November 28th, 2017

2. Disclosures
None

3. Perfusion imaging in acute stroke
Perfusion imaging enables assessment of function and physiology of brain tissue
Can differentiate between core infarct and penumbra1
Enhances accuracy of CT stroke diagnosis1
Determining the amount of salvageable tissue is crucial in the decision to reperfuse2
Currently, CTP or MR-PWI
Lack of standardized thresholds for core, penumbra
Potentially time consuming post-processing
Hopyan et al. Radiology (2010)
Campbell et al. J of Stroke (2014)
Heit et al. Stroke (2016)

4. Dual-Energy CT as a novel perfusion imaging technique
Has shown promise in other neuroimaging applications1
Identifying hemorrhage after thrombectomy
Visualization of brain edema after acute infarct
Cardiac, thoracic, and abdominal studies have shown that iodine concentrations measured using DECT correlate strongly with CTP perfusion parameters2-4
Can DECT provide reliable perfusion data in acute stroke?
Fewer scans, more rapid diagnosis and treatment, less center-to-center variability
Postma et al. Curr Radiol Rep (2015)
Dana et al. Expert Rev Cardiovasc Ther (2015)
Hansmann et al. World J. Radiol (2013)
Stiller et al. Invest Radiol (2015)

5. Objective
Is there a significant difference in iodine concentrations of normal versus ischemic brain parenchyma, which can be reliably measured using DECT?

6. Materials and Methods Retrospective, IRB-approved study
34 consecutive patients labelled as “Stroke Code Activations”
All <4h since symptom onset
Excluded patients with posterior circulation or lacunar infarcts
We excluded patients with PCA, infratentorial or non-territorial infarcts.

7. Materials and Methods
Contrast-enhanced DECT angiogram of head and neck
Dual-Source, Dual Energy Scan
Tube A: 90 kV
Tube B: 150 kV with tin filter

8. Normal
We used a proprietary software platform that analyzes the dual energy datasets.
3 material decomposition performed with base materials air, water and iodine.
Iodine maps reconstructed from the arterial phase image of multiphase CT angiography (arterial, peak venous, and late venous phases according to Calgary Stroke Protocol)
ROIs over insular cortex (GM), basal ganglia, and external capsule (WM) on both hemispheres to measure iodine concentration
Here’s what the Iodine Maps look like
3 images taken from the same patient with suspected acute infarct
ROI’s drawn over the External Capsule, , the basal ganglia, and the insular cortex, measuring iodine concentration and percent iodine uptake
Top, Normal hemisphere
Bottom, corresponding ROI on the abnormal hemisphere.
Light orange = more iodine, dark orange = less iodine
Grossly, can appreciate that affected areas look hypodense on Iodine map compared to normal hemisphere.
Looking at iodine quantification, we see that indeed, Iodine concentrations are significantly lower in ischemic regions.
Blow up the images, add arrows to each
Maybe cut it in the middle and fuse images so just one image
There might be difficulty on just visual assessment alone – more delayed images might demonstrate these changes better on a purely visual assessment
Ischemic

9. Materials and Methods
Reviewed by a fellowship-trained neuroradiologist
CTA was the reference standard to confirm acute infarct
Compared normal vs ischemic brain parenchyma
Iodine concentration
Percent iodine density (normalized to precentral gyrus)
Percent iodine uptake normalized to the normal precentral gyrus cortex
100% iodine uptake represents normal grey matter

10. Right MCA Stroke Normal Ischemic
Say “iodine density” instead of “uptake”. “Ioinde density was significantly decreased in the affected hemisphere)
ROI’s drawn over the External Capsule and the basal ganglia measuring iodine concentration and percent iodine uptake
We see that iodine concentration and percent iodine uptake are significantly decreased in the ischemic
:the quantity of iodine vs uptake
ROI’s.
Ischemic

11. Normal Ischemic Here’s what the Iodine Maps look like
3 images taken from the same patient with suspected acute infarct
ROI’s drawn over the External Capsule, , the basal ganglia, and the insular cortex, measuring iodine concentration and percent iodine uptake
Top, Normal hemisphere
Bottom, corresponding ROI on the abnormal hemisphere.
Light orange = more iodine, dark orange = less iodine
Grossly, can appreciate that affected areas look hypodense on Iodine map compared to normal hemisphere.
Looking at iodine quantification, we see that indeed, Iodine concentrations are significantly lower in ischemic regions.
Blow up the images, add arrows to each
Maybe cut it in the middle and fuse images so just one image
There might be difficulty on just visual assessment alone – more delayed images might demonstrate these changes better on a purely visual assessment
Ischemic

12. Results Mean iodine concentration (mg/mL) ROI within Normal Hemisphere
ROI within Affected Hemisphere
Insula
1.0
0.26
White matter
0.35
0.13
Basal ganglia
0.7
0.28
When we look at mean data from 34 patients, we see that Iodine uptake and concentration is significantly decreased in ischemic brain compared to normal brain, in all three regions
Mean Iodine Density
ROI within Normal Hemisphere
ROI within Affected Hemisphere
Insula
93%
23%
White matter
30%
10%
Basal ganglia
65%

13. Limitations Small cohort No MRI follow-up
Iodine map images not standardized
Other scanners or post-processing software may yield different results
Add to conclusions slide

14. Conclusion
Iodine concentrations were significantly decreased in ischemic brain tissue compared normal tissue suggesting DECT may be a reliable alternative perfusion imaging study
Further study to determine whether iodine quantification can be used to determine core and penumbra. Compare iodine quantifications in the peak venous and delayed venous phases.
A single DECT study may potentially replace NCCT, CTA, and CTP for acute stroke imaging
Are we working on delayed phase study?
Cut the intro

15. Thank you!
Thank you!

16. Perfusion imaging in acute stroke
Conventional imaging only provides anatomic information and lacks accuracy in identifying infarct in the hyperacute stage of stroke1
Perfusion imaging enables assessment of function and physiology of brain tissue
Can differentiate between core infarct and penumbra2
Enhances accuracy of CT stroke diagnosis3
Currently, CTP or MR-PWI
Characterization of core vs penumbra based on perfusion parameters: MTT, CBF, CBV, Tmax
Wardlaw et al. Radiology (2005)
Bivard et al. AJNR (2014)
Hopyan et al. Radiology (2010)
Campbell et al. J of Stroke (2014)

17. Perfusion imaging of acute stroke
Salvage of ischemic penumbra is the mechanism of recovery from stroke1,2
EXTEND-IA: “Patients with proximal occlusion and salvageable tissue on CTP had improved reperfusion, early neurologic recovery, and functional outcome after endovascular thrombectomy3”
Determining the amount of salvageable tissue is a crucial factor in the decision to reperfuse3
Time is brain, but timelines differ from patient to patient
Allows individualization of therapy; potentially extending therapeutic time window
Identification of at-risk, but salvageable brain tissue
Salvage of ischemic penumbra results in reduction of infarct growth, and clinical benefit.
Trials that incorporated CTP had best outcomes (EXTEND-IA, SWIFT PRIME)
Brief.
Davis et al. Lancet Neuro (2008)
Campbell et al. NEJM (2015)
Campbell et al. J of Stroke (2014)

18. Difficulties with Current Functional Imaging Methods
Lack of standardized thresholds for core, penumbra
Potentially time-consuming post-processing
Increased radiation dose (CT)
Motion between repeat scans could cause motion artifact
Lack of accessibility (MRI)
Heit et al. Stroke (2016)

19. Normal Ischemic Here’s what the Iodine Maps look like
3 images taken from the same patient with suspected acute infarct
ROI’s drawn over the External Capsule, , the basal ganglia, and the insular cortex, measuring iodine concentration and percent iodine uptake
Top, Normal hemisphere
Bottom, corresponding ROI on the abnormal hemisphere.
Light orange = more iodine, dark orange = less iodine
Grossly, can appreciate that affected areas look hypodense on Iodine map compared to normal hemisphere.
Looking at iodine quantification, we see that indeed, Iodine concentrations are significantly lower in ischemic regions.
Blow up the images, add arrows to each
Maybe cut it in the middle and fuse images so just one image
There might be difficulty on just visual assessment alone – more delayed images might demonstrate these changes better on a purely visual assessment
Ischemic

20. Materials and Methods Datasets post-processed on syngo.via platform
3 material decomposition performed with base materials air, water and iodine.
Iodine maps reconstructed from the arterial phase image of multiphase CT angiography (arterial, peak venous, and late venous phases according to Calgary Stroke Protocol)
ROIs over insular cortex (GM), basal ganglia, and external capsule (WM) on both hemispheres to measure iodine concentration
(Vitrual unenhanced = iodine, water, air
Brain hemorrhage = iodine, CSF, blood
Mo: which one?
Not a prototype, we used the de application x.
Did we alter the paramters?