Nghiên Cứu Đặc Điểm Hình Ảnh Học Và Đánh Giá Hiệu Quả Của Kỹ Thuật Lấy Huyết Khối Cơ Học Ở Bệnh Nhân Nhồi Máu Não Cấp (Tt Anh).Pdf

MINISTRY OF EDUCATION AND TRAINING

MINISTRY OF HEALTH

THE THESIS HAS BEEN COMPLETED AT

HANOI MEDICAL UNIVERSITY
=======

HANOI MEDICAL UNIVERSITY

Supervisor: Prof. Pham Minh Thong
Reviewer 1:

NGUYEN QUANG ANH
Reviewer 2:

EVALUATING THE CHARACTERISTIC OF
DIAGNOSTIC IMAGING AND THE RESULTS OF
MECHANICAL THROMBETOMY IN THE TREATMENT
OF ACUTE ISCHEMIC STROKE PATIENTS

Reviewer 3:

The thesis will be present in front of board of university examiner and

Speciality
Code

: Radiology & Nuclear medicine
: 9720111

reviewer level at

on

ABSTRACT OF DOCTORAL THESIS

This thesis can be found at:
- National Library
- Hanoi Medical University Library
HA NOI - 2023


1
LIST OF PUBLISHED ARTICLES CONCERNING THE THESIS

1.

Nguyen Quang Anh. Effect of mechanical thrombectomy with vs.
without intravenous thrombolysis in acute ischemic stroke. Clin
Ter 2022; 173 (3): 257 - 264

2.

Nguyen Quang Anh. Results of mechanical thrombectomy in
acuted ischemic stroke patients due to large vessel occlusion at
Bach Mai hospital: sharing experiences from 227 cases. JMR 2022;
154 E10 (6): 28 - 36

INTRODUCTION

Cerebral stroke includes hemorrhage and infarction, in which
ischemic stroke accounts for 80-87% of cases. The consequences of
ischemic stroke, if not detected and treated in time, are severe, leaving
a double burden on both families and society. Similar to the world, in
Vietnam, with the shift of the disease pattern according to the
development of modern society, the number of patients with acute
ischemic stroke every year tends to increase rapidly while the number
of our stroke centers is not enough to meet both quantity and quality.
With all these reasons, a trial with large sample size in Vietnam to
have the overview of imaging characteristics and to analyze the
effectiveness of mechanical thrombectomy techniques are needed.
Therefore, we conduct a study “Evaluating the characteristic of
dianostic imaging and the results of mechanical thrombectomy
in the treatment of acute ischemic patients”, with two details
purposes:
1. Describe CT Scanner and MRI imaging charateristics in
patients with acute ischemic stroke due to large vessel
occlusion
2. Evaluating effects of mechanical thrombectomy in acute
ischemic stroke patient with large vessel occlusion
1. The need of thesis implementation:
The number of patients with acute ischemic stroke increased in all
hospitals at all levels. Timely diagnosis and treatment help to reduce
disability rates and improve the patient's chances of recovery and return
to a normal life. In Vietnam, a lot of studies about this topic were
reported in the past 5 years, which have been published both domestic
and international journals. However, there are still many controversial
issues that need to be clarified with a sufficiently large sample size. In
the diagnosis, CT Scanner is preferred because of its suitability for the
patient's urgent situation but the application of multiphase (evaluating

collateral circulation) and perfusion imaging (identifying the core,
penumbra volume) are still limited. In treatment, there have been many
trials were conducted and published after the success of 5 randomized
controlled trials in 2016 that proved the effect of mechanical


2

3

thrombectomy treatment. Some issues still need to be clarified such as:
the role of rt-PA, a different effect between mechanical devices or
factors affecting the rate of recanalization and clinical recovery post
treatmentn. Therefore, conducting a research at a comprehensive stroke
center with a large number of patients will help to partially solve some
of problems and improve treatment effectiveness in Vietnam.
2. Novel contributions of the thesis:
- ASPECTS was 7.76 ± 1.20 (median 8) and pc-ASPECTS was
7.55 ± 1.62 (median 8). The highest rate seen in M1 occlusion with
41% of patients. The average of collateral score was 3.32 ± 1.44,
which seen the most in tandem occlusion group. In the perfusion map,
the average volume of infarction was 23.5 ± 9.0 cm3, which is smaller
in groups with higher ASPECTS (9-10 point) and better collateral
score (4-5 point).
- Good recanalization (TICI 2b-3) was 84.6% while the rate of
successful first-pass was 47.6%. The hemorrhagic rate seen in 25.1%
of patients but only 3.1% was symptomatic intracranial hemorrhage.
There was 2.7% of severe complications. After 3 months, good clinical
recovery (mRS ≥ 2) was 65.2% and the mortality was 12.8%.
- There was no significant difference seen both in 3 groups of

mechanincal devices (stent, aspiration, solumbra) and in 2 kinds of
treatment in the first 4.5 hours (thrombectomy alone vs thrombectomy
+ intravenonus rt-PA).
- Procedural time ≤ 60 mins (OR 5,952; 95% CI 2,755 – 12,821, p
= 0,000) was an independent predictor to the successful recanalization
(TICI 2b-3). Age < 80 (OR 3,842; 95% CI 1,764 – 8,365; p = 0,011),
NIHSS baseline < 18 (OR 4,917; 95% CI 2,524 – 9,580; p = 0,000),
good collateral (OR 15,047; 95% CI 7,181 – 31,529; p = 0,000) and
good recanalization (OR 3,006; 95% CI 1,439 – 6,276; p = 0,005)
were both independent factors in predict the good outcome (mRS 0-2)
at 90 days after treatment.

3. Thesis layout:
The thesis consists of 128 pages. Apart from the introduction (2
pages), the conclusion (2 pages), the recommendations (1 page) and
the limitation (1 page), it also has four chapters include: Chapter 1:
Overview 43 pages; Chapter 2: Materials and methods 18 pages;
Chapter 3: Results 25 pages; Chapter 4: Discussion 36 pages. The
thesis consists of 26 tables, 26 pictures, 10 charts, and 182 references
(Vietnamese: 8, English: 174).
Chapter 1
OVERVIEW
1.1. Literature review in the world
In terms of diagnosis, according to the recommendations of the
American Heart Association and Stroke, computed tomography is still
preferred while magnetic resonance imaging is recommended for the
diagnosis of vertebro - basilar occlusion or wake-up stroke. Evaluation
of collateral circulation was studied by Menon since 2014 on 140
patients and then showed that this method has good reliability in
assessing collateral in ischemic areas (n = 30, k = 0.81 , p<0.01),

reduce uncertainty in treatment decisions and have better predictive
value of clinical recovery. After the success of the DAWN and
DEFUSE III trials in 2018, the therapeutic intervention window was
extended up to 24 hours with strictly criteria selection based on
perfusion imaging. In treatment, the breakthrough improvement of
thrombectomy devices has resulted in superior results in the group
using the new generation stents compare to intravenous rt-PA group.
Meta-analysis based on data of 5 randomized trials in 2016 (MR
CLEAN, EXTEND IA, ESCAPE, SWIFT PRIME, REVASCAT)
including 1287 patients performed by the Hermes Collaboration
showed a significant reduction in disability after 90 days in patients
treated by thrombectomy compared to control group (OR 2.49, 95%
CI 1.76-3.53; p < 0.0001). Thanks to endovascular thrombectomy, the
number need to treat (NNT) was only 2.6 to obtain a good recovery
(mRS≥2). Efficacy relative to some sub-group including patients with
age ≥ 80 years (OR 3.68, 95% CI 1.95-6.92), time from onset to
administration ≥ 300 minutes (1.76, 1.05-2.97) and non indication for


4

5

IV rt-PA (2.43, 1.30-4.55). Mortality rate after 3 month and
symptomatic hemorrhagic transformation were not significant
difference when comparing between all groups. This analysis once
again confirms the benefit of mechanical intervention with newgeneration stents in patients with acute cerebral ischemia due to
anterior circulation occlusion. Meanwhile, the introduction of large
bore catheter (with inner diameter from 0.060" - 0.072") has brought
another effective choice in thrombectomy treatment. In 2015, the study

by Almandoz et al showed a better clinical recovery rate at 90 days
using the ADAPT technique (55.6%) compared with the solumbra
group (30.9%). The results also suggest that the ADAPT technique is
considered an independent predictor of good clinical outcome in
patients with anterior circulatory occlusion. However, in 2017, the
ASTER randomized clinical trial of 381 patients comparing direct
aspiration (192 patients) with stents retriever (189 patients) for major
artery occlusion did not show a significant difference in clinical
recovery (45% vs 50%), where the rate of complete recovery (mRS 0)
was lower in the thrombectomy group (24% vs 40%). The “rescue
treatment” applied 33% in the aspiration group and 24% in the stent
group both increased the good recanalization rate (TICI 2b-3) from
63% and 68% to 85.4% and 83.1%, respectively (p = 0.53).
Additionally, rt-PA treatment was previously using in “rescue”, where
patients were infused intravenously and wait 60 minutes for clinical
improved observe. If there is no clinical change and NIHSS score ≥ 8
were considered as a treatment failure, then endovascular intervention
is implemented. These limitations were also shown in the design of 2
studies IMS III and Synthesis. By the end of 2016, this method had
changed from “rescue” to “combined” meaning indicated patients will
receive rt-PA right on the computerized tomography table, then
immediately transferred to the intervention room to conduct
mechanical thrombectomy. This takes advantage of both methods and
saves time for the recanalization. This change is also reflected in the
treatment recommendations from 2018 of the American Heart
Association or European Stroke Association in patients with major
embolism within the first 4.5 hours.

1.2. Literature review in Vietnam
Procedural thrombectomy in Vietnam have been conducted since 15

years ago in which the second generation device (stent Solitaire) was first
used at Bach Mai Hospital in 2012. However, the intervention procedure
was still no consensus because of the limited results of studies at this time.
Thanks to the success of 5 major international studies in 2016 and the
change in recommendations of the American Heart Association and
Stroke during 2015 - 2018, many studies in Vietnam in the past 5 years
were conducted with a good design showing many encouraging results in
which 3 major intervention centers at Bach Mai Hospital, People's
Hospital 115 and Military Hospital 103 took part in DIRECT-SAFE
study. Nguyen Hoang Ngoc et al evaluated 138 patients with acute
ischemic stroke due to large vessel occlusion at Military Central
Hospital 108 from July 2016 to June 2017. The results showed that
45.7% occlusion of the middle cerebral artery, 36.2% of the internal
carotid artery and 17.4% of the basilar artery, ASPECTS > 6 was 95%.
The good recanalization rate (TICI 2b-3) was 79.7% and the good
outcome (mRS 0-2) was 58.7%. Another research by Vu Viet Lanh et
al (2019) evaluated 104 patients with large vessel occlusion treated by
Solitaire stents at People's Hospital 115. As a result, there were 55.8%
occlusion of middle cerebral artery, 37.5% occlusion of internal carotid
artery and 6.7% occlusion of basilar artery. The rate of good
recanalization seen in 73.9% of patientns while 12.5% of symptomatic
intracranial hemorrhage discovered. Follow-up after 3 months, good
clinical recovery rate (mRS 0-2) reached 50% and mortality was 23.1%.
The study also showed factors related to good neurological recovery
including: age < 70, good revascularization, good collateral circulation,
asymptomatic hemorhage and NIHSS baseline ≤ 15 (p < 0.05). At Bach
Mai Hospital, Dao Viet Phuong et al (2019) conducted a study on 86
patients with the anterior circulation occlusion within the first 6 hours
which was treated by using a combined therapy (intravenous rt-PA +
thrombectomy). The results showed that the good recanalization (TICI

2b-3) achieved in 91% of patients, of which the complete
recanalization rate (TICI 3) was 52%. Symptomatic hemorrhagic
transformation only accounted for 5.8%. The rate of good clinincal
recovery (mRS 0-2) was 69.8% and the rate of complete recovery
(mRS 0-1) was 53.5%. Most recently, a trial by Tran Anh Tuan et al
(2020) on 22 patients with basilar artery occlusion showed that the rate


6

7

of good recanalization and clinical recovery was 68.2% and 50%,
respectively. The rate of major bleeding was 13.6% and the mortality rate
was 36.4%. According to the authors, low parenchymal injury (pcASPECTS ≥ 7) and good revascularization rate (TICI 2b-3) were
predictors of good clinical outcome (mRS 0-2). Another study by Vu
Dang Luu et al (2020) including 17 complicated cases of tandem
occlusion recorded a good recanalization rate (TICI 2b-3) reached 82.4%.
There were 10/17 patients undergoing acute stenting, but there was no
difference in treatment outcomes between the two groups (retrograde vs
antegrade thrombectomy, p > 0, 05). The rate of symptomatic
hemorrhagic transformation post treatment was 11.8% and good outcome
(mRS 0-2) after 3 months was 47.1%.

2.2. Methodology
2.2.1. Methodology
Prospective clinical intervention study, pre and post treatment
evaluation, non-randomized, non-blind and no control group.
2.2.2. Number of patient
Estimated 227 patients

2.2.3. Data analysis
The data were analyzed using SPSS 22.0 software. Algorithms used in
the study include::
- General descriptive statistics of research variables.
- Qualitative variables are described by frequency and percentage,
using the X2 test (if the standard variable) or the "Fisher exact
test" (when the non-standard variable with any expected
frequency has a value < 5).
- Quantitative variables are described by mean and standard
deviation (if standard variable) or median value (if non-standard
variable). When comparing 2 means: use T-test with standard
variables and Mann-Whitney test with non-standard variables.
For multiple means, use the ANOVA test with the standard
variable and the Kruskal–Wallis test with the non-standard
variable.
- Perform binary comparison to identify significant clinical,
imaging, and interventional factors affecting the good
revascularization rate (TICI 2b-3) and good clinical recovery
(mRS 0-2) after 3 months.
- Perform multivariable regression analysis to variables with
independent prognosis to predict the good recanalization (TICI
2b-3) and good clinical recovery (mRS 0-2) after 3 months
follow-up
- p was considered as a significant difference when its value ≤
0,05.

Chapter 2
MATERIALS AND METHODS
2.1.Research subjects
2.1.1. Inclusion criteria

Based on AHA/ASA 2018 guidelines and recommendation:
- Age ≥ 18; NIHSS ≥ 6
- ASPECTS ≥ 6 with anterior occlusion; for posterior occlusion:
pc-ASPECTS ≥ 7 or no pons lesion idenntified
- Eveidence of large vessel occlusion (ICA, M1, M2, basilar
artery) showed in diagnostic imaging
- Time from onset to administration not more than 16 hours. For
late window 6 – 16 hours: criteria based on DEFUSE 3 (core
volume ≤ 70ml and ratio of penumbra/ core ≥ 1,8)
- Patient’s familly understand about the procedure and agree to
sign in a commitment to treatment
2.1.2. Exclusion criteria
- Wake-up stroke.
- Any hemorrhage shown in image
- Chornic occlusion (Moya Moya disease…)
- Pre - mRS ≥ 2 (before stroke)
- Severe condition with other diseases (kidney failure, invasive
cancer…) and could not be followed up


8

9

Chapter 3
RESULTS
The study was conducted on 227 patients (N) with acute ischemic
stroke who underwent mechanical thrombectomy at the Radiology
Center of Bach Mai Hospital: 205 cases (N1) with anterior occlusion
and 22 cases (N2) ) with basilar occlusion. There were 178 patients

(N3) who came to the hospital in the first 4.5 hours and 32 cases came
within the 6-16 hour window (that was controlled by perfusion
nimaging).
3.1. General characteristics of the patients
3.1.1. Age and gender
- Mean was 65±13 (22-90). The age group < 45 years accounted
for 5.7%, the middle-aged group (45-69) accounted for 55.5% and the
elderly group (>70) accounted for 38.8%. Male/female ratio = 1.2%.
3.1.2. Clinical and time characteristic at hospital admission
- The rate of hypertension was 59.5%, hyperglycemia was 58.6%
and atrial fibrillation was 24.7%. NIHSS score at admission was
14.27 ± 4.8.
- The mean time from onset to hospital admission was 203 ± 153
minutes; from hospital admission to first image was 39 ± 37 minutes;
from hospital admission to femoral puncture was 98 ± 55 minutes.
3.2. Imaging characteristic of large vessel occlusion
3.2.1. Occlusion site
- The rate of M1 middle cerebral artery occlusion was 41.9% (95
patients), internal carotid occlusion was 23.8%, M2 segment occlusion
was 13.2% and tandem occlusion was 11.5%. There were 9.7% cases
(22 patients) of basilar artery occlusion.
3.2.2. Non contrast imaging charateristic
- The mean ASPECTS and pc-ASPECTS was 7.76 ± 1.20 and 7.55
± 1.62, respectively. The time from onset to the first scan of anterior
and posterior occlusion was 241 ± 148 minutes and 243 ± 181
minutes, respectively.
- For anterior occlusion, there was a significant difference in
ASPECTS at different sites (p=0.01): M2 occlusioin had the least
parenchymal damage with an average ASPECTS score of 8 .33 ± 1.09

and the time from onset to the first imaging was shortest: 148 ± 130
minutes (p= 0.05). Occlusion of the internal carotid artery and the M1
had more parenchymal damage, respectively the average ASPECTS
of 7.69 ± 1.23 and 7.54 ± 1.13. Time from onset to hospital imaging
of the tandem group was the longest: 279 ± 227 minutes, p = 0.02.
- The group of patients with more severe clinical condition (higher
NIHSS) had more parenchymal damage (ASPECTS 6-7) and vice
versa, but the difference was not statistically significant (p = 0.09).
3.2.3. Collateral characteristic in CT Scanner multiphase
- The mean collateral score was 3.32 ± 1.44 (for anterior
occlusion). Tandem group had the best collateral score (3.5 ± 1.3) but
the difference was not statistically significant when compared with
others (p = 0.95).
- The group of patients with good clinical status (lower NIHSS)
has a better collateral score (13.1 ± 4.3) and vice versa. There was a
statistically significant difference between the study groups (p=0.04).
3.2.4. Perfusion imaging characteristic
- 32 patients admitted to the hospital with time window > 6 hours
underwent cerebral perfusion imaging. The mean time was 505 ± 134
(min) and the mean core volume was 23.50 ± 9.00 (cm3)
- Less parenchymal lesions (ASPECTS 8-10) have a smaller core
volume (20.85 ± 9.35) compared with more parenchymal lesions
(ASPECTS 6-7) corresponding to a larger core volume (26.58. ±
7.81) but the difference was not significant (p = 0.07). Time from
onset to hospital admission between groups was not significant also (p
= 0.61).
- Core volume decreased gradually from the group with poor
collateral circulation (29.00 ± 5.52) to moderate (23.74 ± 7.76) and
good (21.81 ± 10.15) while penumbra volume gradually increased,
2.50 ± 0.75 (poor); 3.57 ± 1.86 (moderate) and 5.11 ± 3.42 (good),

respectively. The difference was not statistically significant (p = 0.30
and 0.14). The time from onset to hospital admission was different
between groups, but not statistically significant (p = 0.27).


10
3.3. Endovascular mechanical thrombectomy effect
3.3.1. Characteristic and general results
- Of the 227 patients undergoing mechanical thrombectomy,
80.6% received endotracheal anesthesia and 19.4% received local
anesthesia. The average intervention time was 40 ± 27 minutes (9 –
150) with an average number of pass was 2.73 ± 1.58. There were 178
patients admitted to the hospital in the first 4.5 hours of which 64
cases (accounting for 28.2%) were treated with rt-PA.
- The good recanalization (TICI 2b-3) was 84.6%. There were 6
patients with severe complications related to the procedure: 4 cases of
dissection (1.8% - of which 1 case died, accounting for 0.4%) and 2
cases of perforation (0.9% - leading to both mortalities). The mean
number of thrombectomy pass was 1.87 times (median 1, range 1-10).
- 24 hours follow-up, the mean NIHSS was 10.68 ± 7.76, lower
than that at admission with an average reduction of 3.59 ± 4.0 points.
130 patients decreased ≥ 4 points and 61 patients decreased ≥ 8 points
with 86.9% of them having good clinical recovery after 3 months.
- 148 patients had a good clinical recovery (mRS 0-2), reaching
the rate of 65.2% while 29 cases of death (mRS 6) accounted for
12.8%. 7 patients had symptomatic transformation, accounting for
3.1%. Asymptomatic hemorrhage seen in 50 patients, accounting for
22%.
- Good clinical recovery after 3 months (mRS 0-2) seen in
occlusion group of internal carotid artery, M2, tandem and M1 were:

59.3%, 63.3%, 69.2% and 73.7%, respectively. Occlusion of the
middle cerebral artery (M1 and M2 segments) had a lower mortality
rate, 6.3% and 6.7%, respectively. With posterior circulation, the
good clinical recovery rate was only 40.9% while the mortality was
highest (45.5%).

11

Figure 3.1. Rate of good recanalization and clinical recovery by number of
thrombectomy pass

- 47.6% had good revascularization (TICI 2b-3) and 35.2% had
good clinical recovery (mRS 0-2) with only 1 pass. The cumulative
rate of good recanalization increased to 66.1% and 77.1% after the 2nd
and 3rd pass. After the 4th, the increase was not significant (<1%). The
rate of good clinical recovery at the cumulative 90 days also increased
to 51.1% and 59.5% after 2nd and 3rd pass. After the 4th, the increase
was not significant (<1%).
3.3.2. Comparison in groups of treatment
3.3.2.1. Based on mechanical device
- Characteristics at hospital admission between 3 groups of devices
(stent, aspiration and solumbra) did not show statistically significant
differences (p > 0.05) in terms of clinical factors (age, gendner,
NIHSS), imaging (ASPECTS and pc-ASPECTS) and time from onset
to hospital admission. The time from hospital admission to femoral
puncture in the solumbra group (stent + aspiration) was 85 ± 38
minutes, shorter than the stent group (107 ± 58 minutes) and the
aspiration group (96 ± 58 minutes). The difference was statistically
significant (p=0.04).

12
Table 3.1. Results comparison between 3 groups of mechanical devicie

13
- The rate of symptomatic intracranical hemorrhage post treatment
was not recorded in the stent group, while the other 2 groups had no
difference (5.6% and 4.5%, p > 0.05). Good clinical recovery and
mortality after 3 months in 3 intervention groups did not show a
statistically significant difference (p = 0.32).
3.3.2.2. Based on treatment in the first 4.5 hours
- Characteristics at hospital admission between the 2 treatment
groups did not differ (p > 0.05) in terms of imaging and most of the
clinical features. The time from onset to hospital admission in the
combined group (IV rt-PA + thrombectomy) was significantly shorter
(p < 0.05), but the time variables from hospital admission to femoral
artery puncture did not differ between the 2 groups (p > 0.05).
Table 3.2. Results comparison between 2 groups of treatment
in the first 4.5 hours

- Between the 3 groups of device, the occlusion rate at M1, basilar
artery and tandem was similar but the stent group had a lower rate in
the internal carotid artery (12.9%) and higher in the M2 (22.6%), the
difference was significant (p < 0.05). The rate of using combined
intravenous rt-PA before thrombetomy in all 3 groups did not differ:
26.7% - 30.1% with p = 0.83.
- The mean intervention time in aspiration group was shortest: 35 ±
23 minutes, the difference was not statistically significant (p = 0.21).
The average number of pass in the stent group was least (1.61) with
significant difference (p = 0.02). The stent group had a significantly

lower rate of rescue therapy (using a second device) than the
aspiratiion group: 7.52% versus 33.33%, p < 0.05. The rate of good
recanalization (TICI 2b-3) achieved at first pass was highest in the
stent group: 61.3%, p = 0.04 but the general recanalization was all
good post procedure in all 3 groups without statistical difference (p =
0.55).

- In the group of thrombectomy alone, the occlusion rate in the
internal carotid artery was higher (28.1%) while in the M2 and
tandem group, this rate was lower (12.3% and 7.9%) compared with
combined group, the difference was statistically significant (p < 0.05).


14

15

Other parameters related to the procedure including time of
intervention, number of pass, good recanalization and hemorrhagic
transformation rate, there was no difference (p > 0.05). The good
clinical outcome after 3 months in the 2 groups of treatment did not
show a statistically significant difference (p = 0.60).
3.3.3. Independent factors affecting the good recanalization and
good clinical recovery rate after 3 months
- Factors affecting the good recanalization were: no extracranial
stenosis (p = 0.022) and intervention time ≤ 60 minutes (p = 0.000).
In multivariate regression analysis, intervention time ≤ 60 min was
the most independent factor affecting the good revascularization rate
post treatment (OR 5,952; 95% CI 2.755 – 12,821, p = 0.000).
- Factors correspondinig to the rate of neurological recovery after

3 months were: age < 80, hypertension, atrial fibrillation, NIHSS
admission < 18 points, degree of collateral circulation, degree of
revascularization and symptomatic transformation post thrombectomy
(p < 0.05). In multivariable regressionn analysis, age <80 (OR 3.842;
95% CI 1.764 – 8.365; p = 0.011), NIHSS admission < 18 points (OR
4.917; 95% CI 2.524 – 9,580; p = 0.00), good collateral score (OR
15,047; 95% CI 7.181 – 31,529; p = 0.000) and good
revascularization (OR 3.006; 95% CI 1.439 – 6,276; p = 0.005) were
the independent factors, having the most influence on good clinical
outcome (mRS 0-2) at 90 days after treatment.
Chapter 4
DISCUSION
4.1. General characteristics of the patients
4.1.1. Age and gender
The mean age in our study was 65 ± 13, in which the youngest
patient was 22 and the oldest was 90. This result was higher than some
domestic results of Nguyen Huy Thang (60.5 ± 12.2) and Dao Viet
Phuong (61.9 ± 11.8). Compared with some international studies,
patients in our study had the same mean age with MR CLEAN and
SWIFT PRIME but lower than ESCAPE. The increase in age
adversely affects clinical recovery as in the study of Weinar et al. In
our study, the elderly group (> 70 years old) accounted for 38.8%, of

which the group > 80 years old accounted for 14.1% (32 patients).
According to the current data, older age is not a contraindication to
mechanical thrombectomy. It should be noted that the rate of stroke in
young people tends to increase, especially in the 40-45 year old group.
Our study recorded that young patients with cerebral ischemic stroke
accounted for 5.7%, similar to the statistics in the world (from < 5% to
20%). Additionally, the research results showed that men predominate

in terms of gender with 55%, 1.2 times higher than women, lower than
Vu Viet Lanh study (60.4% male, male/female ratio was 1.5).
4.1.2. Clinical and time characteristic at hospital admission
Our study recorded the mean value of NIHSS at the time of
admission was 14.27 ± 4.8 (range 8-20), which were both lower than
other international studies. In an analysis of 1281 patients, Adam et al
showed that NIHSS was valuable in predicting clinical outcomes after
treatment, with a score of < 6 having a good prognosis while a score of
> 16 was often related with high mortality and disability rates. In our
study, the proportion of patients with hypertension and hyperglycemia
predominated, 59.5% and 58.6%, respectively. The rate of
hypertensive patients was similar to the study by Toyoda (61%) and
significantly higher than the study of Mai Duy Ton and Dao Viet
Phuong (27.9%). In this study, we discovered atrial fibrillation by
electrocardiogram in 56 patients, accounting for 24.7%. This result is
lower than that of Vu Viet Lanh (26%) and MR CLEAN (28.3%). In
our study, the median value of time from onset to hospital admission
as 165 minutes. Compared with international studies, this was still a
long time (1.5-2 times) and would reduce the patient's chance of
getting treatment. It was noteworthy that the time from imaging to
femoral puncture was shortened with a median value of 51 minutes,
equivalent to the value in the ESCAPE study and shorter when
compared with EXTEND-IA and SWIFT PRIME trials.
4.2. Imaging characteristic of large vessel occlusion
4.2.1. Occlusion site
The results in our study showed that the distribution of occlusion
site was consistent with the general trend of large studies in the world
when the anterior circulation accounts for the majority (90.3%) with

16

17

95 cases of M1 occlusioin (41.9%) while the vertebro-basilar
occlusion only account for 9.7%.
4.2.2. Non contrast imaging charateristic
We found that the group of patients with large vessel occlusion
(internal carotid artery, M1 and basilar artery) had a higher degree of
parenchymal damage, corresponding to an average ASPECTS/ pcASPECTS of 7.69 ± 1.23 , 7.54 ± 1.16 and 7.55 ± 1.62. The M1 group
had lower ASPECTS values compared to other sites of the anterior
circulation. The reason was that the patients in this group had a late
admission time (215 ± 136 minutes), so there was more parenchymal
damage, the difference was statistically significant (p < 0.05). A
special feature was that the tandem group had the time from onset to
hospital admission and from onset to imaging was longest (242 ± 221
minutes and 279 ± 227 minutes, respectively), however, the
parenchymal lesions according to ASPECTS were not much (only
inferior to M2 occlusion group). This was consistent with the
progression of chronic vascular stenosis that helped to created good
collateral anastomosis. Our study did not have a low ASPECTS
subgroup (< 6), but the analysis results also showed a concordance
between imaging and clinical in patients with ASPECTS 6-7
(moderate parenchymal lesions) had worse clinical status in cobination
with higher NIHSS (14.3 ± 4.6) at the time of admission. This was
similarly to the study of Mai Duy Ton and Dao Viet Phuong.
4.2.3. Collateral characteristic in CT Scanner multiphase
When assessing collateral circulation in patients with anterior
occlusion (205 cases) by CT Scanner multiphase, we recorded an
average collateral score of 3.32 ± 1.44, in which 119 cases with good

circulation (58%), 60 cases with moderate (29.3%) and 26 cases with
poor collaterals (12.7%). The M1 occlusion group had the lowest
collateral score (3.27 ± 1.41) while the tandem group had the highest
collateral score (3.5 ± 1.3), which was suitable thanks to the presence
of surface collateral. The study results also showed that, in the group
of patients with good collateral circulation, the mean NIHSS at
baseline was lower (13.1 ± 4.3) compared with the other two groups
(> 14.5), a statistical significant difference. This was similar to the

conclusion of Hwang when evaluating 86 patients with middle
cerebral artery occlusion, noting that the group with poor collateral
circulation had a severe clinical condition and a higher risk of atrial
fibrillation compared with the another group without these factors.
4.2.4. Perfusion imaging characteristic
There were 32 cases admitted to hospital beyond 6th hour needed
to undergo computed tomography perfusion. The results showed that
although the mean time to hospital admission of this group was 505 ±
134 minutes (equivalent to 6.5 hours), the average ischemic core
volume recorded was only 23.5 ± 9.0ml. The ratio of the volume
between the penumbra and the core was 4.23, consistent with the
selection criteria of DEFUSE 3. When comparing in the subgroups
based on ASPECTS and collateral score, the results showed that good
ASPECTS (8-10) and good collateral score (4-5 points)
corresponding to the lower core volume on cerebral perfusion map
(20.85 ± 9.35 vs 26.8 ± 21.08 or 21.81 ± 10.15 vs 23.74 ± 7.76 and
29.0 ± 5, respectively). EXTEND-IA was one of the pioneering
studies using quantitative assessment based on infarction volume
(measured on perfusion map) <70ml as a criteria for patient selection
without applying ASPECTS like most other studies.
4.3. Endovascular mechanical thrombectomy effect

4.3.1. Characteristic and general results
In our study, the rate of endotracheal anesthesia was 80.6% while
local anesthesia accounted only for 19.4%. A meta-study by Feil et al
on 6635 stroke patients who underwent mechanical thrombectomy in
Germany from 2015 to 2019 showed that the rate of endotracheal
anesthesia accounted for the majority (67.1%) but resulted in slowing
down time to femoral puncture compared to sedation alone (71 versus
61 minutes, p < 0.001). In multivariable regression analysis, the
endotracheal anesthesia was also a predict factor in reducing the good
recovery post thrombectomy (OR = 0.82, CI 0.71 - 0.94, p = 0.004).
The average number of thrombectomy pass in our study was 1.87
(median 1) with 47.6% having good revascularization at first-pass.
The results in Figure 3.1 showed that within 4 passes (219 patients),
the cumulative revascularization rate increased rapidly, from 47.6% to


18

19

82.8%, corresponding to the good clinical outcome follow-up after 3
months, from 35.2% to 64.3%. From the 5th pass, these rates increase
very slowly (1-5%), meaning thrombectomy was no longer effective
at thihs point. A recent study by Garcial-Tornel on 459 patients also
showed that the treatment outcome (84% good recanalization of TICI
2b-3) was inversely proportional to the number of pass, especially in
the group > 4 passes, which was similarly to our results. In our study,
the average intervention time was 40 ± 27 minutes. In the large
European studies, ESCAPE had the prominence of early
recanalization (median 30 minutes, range 18 - 45.5). This yielded

positive results when comparing the clinical recovery after 90 days
between the intervention group (52%) and the rt-PA alone group
(29%).
The good recanalization rate in our study was 84.6%, in which the
level of TICI 3 accounted for 53.3% (121 patients) and TICI 2b
accounted for 31.3% (71 patients). This was higher when compared
with the well-known studies conducted in 2016 with using 2nd
generation devices such as MR CLEAN (58%), ESCAPE (72%) or
REVASCAT (66%) and thus, outperforming other older studies using
1st generation in 2013 (IMS III: 38% or MR RESCUE: 27%). We
recorded 6 cases (2.6%) of severe complications related to the
intervention, of which 3 cases resulted in death, accounting for 1.3%.
This complication rate was similar to other trials such as SWIFT
PRIME (4/98 patients with subarachnoid hemorrhage) or EXTENDIA (1/70 patients with perforation). Follow-up after 24 hours, we
noticed 50 cases of asymptomatic hemorrhage (22%) and only 7 cases
of symptomatiic transformation (3.1%). The rate of transformation in
general accounted for 25.1%, higher than that of the REVASCAT
(21.4%) but lower than that results from Dao Viet Phuong (29%). The
rate of symptomatic intracranial hemorrhage was also higher in the
results of Vu Viet Lanh (12.5%) and Dao Viet Phuong (5.8%)
compared to ours (3.1%). The mean NIHSS after 24 hours of
intervention in our study was 10.68 ± 7.76 (median 8) with an average
reduction of 3.59 ± 4.0 compared to the time of hopital admission.
This rate of Dao Viet Phuong was 70.9%, higher than our result of

57.3% partly due to the sooner window (only 4.5 hour) in the criteria
selection. If we used the criteria with NIHSS decline 8 points, we had
26.9% of cases at 24 hours while Dao Viet Phuong recorded a rate of
55.8%, all significantly lower compared with the results of the
EXTEND-IA study (80%). After 3 months of follow-up, out of 227

patients receiving thrombectomy treatment at Bach Mai hospital, we
recorded that 148 cases (65.2%) recovered well with mRS 0-2, 50
cases (22%) ) recovered slowly with mRS 3-5 and 29 cases (12.8%)
died (mRS 6). This result was only inferior to that published by Dao
Viet Phuong and EXTEND-IA study and similar to SWIFT PRIME
but better than most of the ESCAPE, REVASCAT or MR CLEAN
trials.
4.3.2. Comparison in groups of treatment
4.3.2.1. Based on mechanical device
Among 227 patients, 90 cases of initial treatment with stent
retriever, 90 cases of aspiration (by large bore catheter), and 44 cases
of solumbra technique (the combination of both devices). When
comparing the clinical and imaging characteristics at the baseline,
there were similarities between the 3 treatment groups in terms of age,
gender, comorbidities, NIHSS, ASPECTS/pc-ASPECTS (p > 0.05).
Time table anaylysis recorded a trend towards earlier hospital
admission in the stent group (193 ± 146 minutes versus 214 ± 154 and
203 ± 265 minutes, p = 0.41) while the time from admission to
femoral puncture in the solumbra group was shorter (85 ± 38 min vs.
107 ± 58 and 96 ± 58 min, p = 0.04). Intervention time in the
aspiration group was shortest (35 ± 23 min) while the stent group and
the solumbra group lasted longer, 41 ± 26 min and 46 ± 34 min,
respectively, p = 0, 21. This was consistent with the study of Turk or
Procházka, when the aspiration tube only need to be in contact at the
proximal part of thrombus. However, the number of pass to remove
thrombus was the most in our group using aspiration alone (2.11
passes, 1–10) compared to other two groups (stent: 1.61 passes, 1–10)
and Solumbra: 1.93 passes, 1–6), p=0.02. This inconsistency could be
explained by the distribution of occlusion sites in the treatment
groups in Table 3.1. The impressive final good recanalization resulted

20

21

in the aspirationn group could be explained by the rate of rescue
therapy (switching from treatment alone to combination) up to
33.33%, 4 times higher compared with only 7.52% in the initial stent
group, p = 0.00. This was similar to that in the ASTER study, with a
resuce rate of 33% in the aspiration group but 24% in the stent group,
which helped to increase the total good recanalization rate in the
aspiration group from 63% → 85.4% and stent group from 68% →
83.1%, p = 0.53. All our 3 groups recorded a relatively high
successful outcome rate (mRS 0-2): the stent group was 66.7%, the
aspiration group was 65, 6% and the solumbra group was 61.4%, no
statitically significant difference, p = 0.32. The mortality rate in the
stent group was only 7.6%, much lower than the other 2 groups
(respectively 15.5% and 18.2%) which was thought to be consistent
with the causes mentioned above: early admission time, lower rates of
both internal carotid occlusion and symptomatic intracraninal
hemorrhage post treatment. This result was similar to the ASTER
study where the good clinical recovery rate between the stents and
aspiration group was not different (50% and 45.3%, OR 0.83 [95%
CI, 0.54-1.26], p = 0.38) while Procházka recorded a better clinical
recovery in the 2 groups using stents and aspiration tubes compared to
the solumbra group (p < 0.05).
4.3.2.2. Based on treatment in the first 4.5 hours
Our comparison of 178 patients admitted at a 4.5-hour window
received thrombectomy with intravenous rt-PA (64 patients) versus

thrombectomy alone (112 patients). Although not randomized, the
characteristics of age, sex, NIHSS or ASPECTS at the baseline were
similar between these groups. Time from hospital admission to
femoral puncture (103 ± 56 minutes vs 95 ± 61 minutes, p = 0.37)
was not prolonged in the combined group, and even more urgently
with the rate of procedure time ≤ 90 minutes was 57.8% compared to
50.9%, p = 0.37. This suggested that the concern that the use of
intravenous rt-PA may slow down the interventional treatment does
not seem to be really suitable. The rate of good recanalization after rtPA in combination with mechanical thrombectomy was similar to that
of intervention alone: 90.6% vs. 87.7%, p = 0.56 (table 3.2). This

showed that the use of thrombolysis was not necessarily the key
factor affecting the outcome of recanalization in large vessel
occlusion. Additonally, we found that the endovascular time in the
group of thrombectomy alone tended to be shorter than the combined
group (37 ± 27 minutes and 41 ± 26 minutes, respectively, p = 0.19)
was consistent with the less number of pass: 1.79 ± 1.26 times
(median 1, range 1 - 6) versus 2.06 ± 1.28 times (median 2, range 1 10), p = 0.11, respectively. The rate of intervention time within 45
minutes also showed a better value in the thrombectomy alone group:
71.1% of cases compared with 68.8%, p = 0.74. The similarity in our
results with the study of Weber (good recanalization rates in the two
groups were 73.8% and 73.1%, p = 0.95; the time reports around the
procedure were similar but shorter in the thrombectomy group alone)
again supported the hypothesis that the use of rt-PA tended to
fragment the intracranial thrombus, leading to intervention migration.
The overall rate of intracerebral hemorrhage within 24 hours tended
to be higher in the combinend group (28.1% vs 24.6%, p = 0.6) while
the rate of symptomatic transformation was higher in the group of
thrombectomy alone (5.3% vs 1.6%, p = 0.42) but the difference was
not statistically significant. Our study and most recent international

studies did not show any evidence of association between intravenous
rt-PA and the risk of hemorrhage after treatment, which was
consistent with the results of Mai Duy Ton and Dao Viet Phuong. The
rates of good clinical recovery (mRS 0-2) and mortality (mRS = 6)
observed after follow-up in the two groups of patients were similar
(65.6% and 17.2% in the combined group versus 66.7% and 12.2% in
the thrombectomy alone group, p = 0.60).
4.3.3. Independent factors affecting the good recanalization and
good clinical recovery rate after 3 months
- When analyzing factors related to recanalization level, we found
that clinical factors such as: age (< 80), gender (male), atrial
fibrillation, occlusion site did not affect the recanalization efficiency
(p > 0.05). More interestingly, intravenous rt-PA did not play any
siginificant role in this results also (p > 0.05). Of the two factors


22

23

related to good recanalization (TICI 2b-3), we concluded that the
procedural time within 60 minutes (90.6% vs 61.7%, p = 0.000)
played a more important role than non-extracranial stenosis factor
(86.6% vs 69.2%, p = 0.022). This was shown in multivariate
regression analysis, when intervention time ≤ 60 min was an
independent predictor of good recanalization (OR 5,952; 95% CI
2.755 – 12,821, p = 0.000) compared with non-extracranial stenosis
(OR 2.890; 95% CI 1.131 – 7.353, p = 0.068). The results were
consistent with Vanacker's study on 439 patients showing that
extracranial stenosis combined with intracranial thrombosis (tandem

occlusion) reduces > 50% of good revascularization (ROC value is
0.72) and mechanical thrombectomy was the most important factor
promoting revascularization in acute ischemic stroke.
- Regarding to the level of neurological recovery after 3 months,
we evaluated based on clinical factors (age, comorbidities, NIHSS,
time), imaging factors (ASPECTS, collateral grade, occlusion site)
and interventional-related factors (combined extracraninal stenosis,
recanalization grade, first-pass effect, symptomatic intracraninal
hemorrhage). Multivaria regression analysis results for 4 factors
including: age <80 years old (OR 3,842; 95% CI 1.764 - 8.365; p =
0.011), NIHSS admission < 18 points (OR 4.917; 95% CI 2,524 9,580; p = 0.000), good collateral ciruculation (OR 15,047; 95% CI
7.181 - 31.529; p = 0.000) and good revascularization (OR 3.006;
95% CI 1.439 - 6,276; p = 0.005) were independent predictors of
good clinical outcome (mRS 0-2) at 90 days after treatment. This was
similar with the results of domestic and foreign published studies.

1. Describe the characteristic of both CT Scanner and MRI in
acute ischemic stroke patients due to large vessell occlusion
- The mean of ASPECTS was 7.76 ± 1.20 (median 8) and pcASPECTS was 7.55 ± 1.62 (median 8). The highest rate seen in M1
occlusion with 41% of patients.
- The average of collateral score was 3.32 ± 1.44, which seen the
most in tandem occlusion group. In the perfusion map, the average
volume of infarction was 23.5 ± 9.0 cm3, which is smaller in groups
with higher ASPECTS (9-10 point) and better collateral score (4-5
point)
2. Evaluating the results of endovascular treatment using
mechanical thrombectomy devices
* Treatment results:
- The good recanalization rate (TICI 2b-3) was 84.6%. Good firstpass revascularization achived in 47.6%, the mean number of pass was
1.87 (median 1). There were 2.7% cases of severe complications

related to procedure and 1.3% was deaths. The overall hemorrhagic
transformation after 24 hour was 25.1%, of which symptomatic one
was 3.1%
- After 24 hours, 57.3% treated patients had their NIHSS decreased
≥ 4 points and 26.9% decreased ≥ 8 points. Post 3 months, the good
clinical recovery rate (mRS 0-2) was 65.2% and the mortality rate
(mRS 6) was 12.8%.
* Comaprison between group of treatments:
- Based on device selection: the aspiration group the shortest
intervention time (35 ± 23 minutes) while the stent group had the least
number of pass (1.61 passes). There was no significant difference in
the rate of good revascularization and clinical outcome or mortality
after treatment in all 3 groups (p > 0.05)
- Based on with versus without rt-PA in combination with
thrombectomy: there was no significant difference in the number of
pass, the good recanalization rate, the hemorrhagic rate post treatment
as well as the rate of neurological recovery or death at the time of
follow-up after 3 months (p > 0, 05).
* Evaluate relevant factors:

CONCLUSION
Through a study on 227 acute ischemic patients with large vessel
occlusion who underwent mechanical thrombectomy at the Radiology
Center of Bach Mai hospital from January 2018 to June 2019, we
concluded that:


24
- Intervention time ≤ 60 min (OR 5,952; 95% CI 2.755 – 12,821, p
= 0.000) was an independent predictor of good revascularization (TICI

2b-3). Meanwhile, age < 80 (OR 3.842; 95% CI 1.764 – 8.365; p =
0.011), NIHSS admission < 18 points (OR 4.917; 95% CI 2.524 –
9,580; p = 0.000), good collateral circulation (OR 15,047; 95% CI
7.181 – 31,529; p = 0.000) and good revascularization (OR 3.006;
95% CI 1.439 – 6.276; p = 0.005) were independent prognostic factors
for good clinical recovery (mRS 0). -2) at 90 days after treatment.
RECOMMENDATION
In the diagnosis of acute ischemic stroke, CT multiphase should be
done to evaluate the grade of collateral circulation. More studies with
a larger number of patients are needed to conduct in Vietnam in order
to confirm the role of cerebral perfusion imaging technique in the
selection of stroke treatment in late window time (≥ 6 hours).
Mechanical thrombectomy has been again shown to be effective
with good revascularization and clinical recovery rates, and very low
related -complications. The comparison showed no difference in
outcomes between groups of treatment with different device selections
(stent, suction, combination) or treatment methods (with/ without rtPA) within the first 4.5h. However, more studies with randomized
controlled design are needed to overcome the limitations of our study
LIMITATION
Despite the large sample size, we have not been able to perform a
randomized or controlled study as expected. Therefore, the subgroup
analyzes when comparing the treatment effectiveness between choices
of devices or intervention methods still had biases. This is also an
inherent weakness when conducting research in Vietnam compared to
international ones, but it is also an expectation that further studies will
have better design with best support and effort.