Development of a nomogram to predict hemorrhage transformation after mechanical thrombectomy in patients with acute ischemic stroke caused by large vessel occlusion in the anterior circulation

doi:10.21203/rs.3.rs-3748348/v1

Background: Hemorrhage transformation (HT) is one of the serious complications after mechanical thrombectomy (MT) in patients with acute ischemic stroke (AIS) and is associated with poor prognosis. The aim of this study was to develop a nomogram to predict the risk of post-MT HT in AIS patients.

Methods: AIS patients treated with MT between January 2020 and June 2022 were enrolled in this study. The enrolled patients were randomly divided into training and validation cohorts, in a 2:1 ratio. Lasso regression and machine learning algorithms were used for feature selection. Multivariate logistic analysis was applied to identify the optimal predictors. A nomogram was developed to predict the risk of post-MT HT. Performance of the nomogram was determined by its discrimination, calibration, and clinical usefulness.

Results:A total of 205 patients were enrolled in the study, with 145 in the training cohort and 60 in the validation cohort. The common risk factors revealed by Lasso regression and random forest algorithm were blood glucose on admission, CRP, NIHSS score, ASPECTS and CRP-to-albumin ratio. Multivariate logistic analysis showed that blood glucose on admission (Odds Ratio (OR)=5.61, 95%CI:1.86-20.83, P=0.004) and CRP (OR=73.52, 95%CI: 25.06-276.77, P<0.001) were independent predictors of post-MT HT. The nomogram was developed based on blood glucose on admission and CRP. Moreover, the proposed nomogram showed good discriminative ability with an area under the curve of 0.924 and 0.846 in the training and validation cohort, respectively. The calibration plot showed good concordance between nomogram prediction and actual observation. Decision curve analysis indicated that the nomogram had favorable clinical application benefits.

Conclusion: This study proposed a nomogram based on CRP and blood glucose on admission to predict the risk of post-MT HT in anterior AIS patients. The nomogram showed reliable predictive performance and can help clinicians identify patients at high risk of HT.

acute ischemic stroke

mechanical thrombectomy

nomogram

hemorrhagic transformation

anterior circulation

Mechanical thrombectomy (MT) is one of the most effective reperfusion therapies for acute ischemic stroke (AIS) patients caused by large vessel occlusion(LVO)[1]. Despite the successful perfusion of intracranial large vessels, only a small number of patients have benefited from reperfusion therapy. Hemorrhagic transformation (HT) is one of the serious complications after MT and is associated with adverse outcomes in AIS patients[2]. Therefore, early identification of patients at high risk of HT after MT has great clinical significance.

Previous studies have demonstrated several risk factors for post-MT HT, mainly including old age, hypertension, hyperglycemia on admission, high baseline National Institutes of Health Stroke Scale (NIHSS), low Alberta Stroke Program Early Computed Tomography Score (ASPECTS), and long time from puncture to recanalization[3–5]. Several scoring systems for symptomatic intracranial hemorrhage(sICH) were developed[6, 7]. HT, including symptomatic intracranial hemorrhage(sICH) and asymptomatic intracranial hemorrhage (aICH) after MT, is correlated with poor prognosis of AIS patients[8]. However, few studies have established models to predict the occurrence of HT after MT in AIS patients.

Machine learning (ML) is capable of modeling information based on causal and/or statistical data, processing a large number of variables, and producing more accurate results than traditional statistical methods[9]. ML is currently being used as a powerful and reliable tool for clinical practice. Nomograms can visualize the results of regression models and have been used for disease diagnosis, severity assessment and prognostic prediction[10].

In this study, we aimed to develop a nomogram to predict the occurrence of post-MT HT in AIS patients. The nomogram can help clinicians identify patients at high risk of post-MT HT and optimize management strategies.

Study design

We retrospectively reviewed patients who received MT for large vessel occlusions of anterior circulation within 6 hours of symptom onset at the First Affiliated Hospital of Dalian Medical University between January 2020 and June 2022. The inclusion criteria of this study were as follows: (1) treated with MT within 6 hours of symptom onset and reperfusion was successful; (2) age ≥ 18 years; (3) had occlusion of the internal carotid artery or middle cerebral artery by computed tomographic angiography, magnetic resonance angiography, or digital subtracted angiography; (4) baseline NIHSS score ≥ 6; (5) ASPECTS ≥ 6; (7) without any type of intracranial hemorrhage on non-contrast computed tomography (NCCT). Patients with malignant tumor, autoimmune disease, severe renal insufficiency, hepatic disease, or heart failure were excluded from this study.

All patients underwent follow-up CT 24 hours after MT and immediately in case of a worsening condition. Two double-blinded neurologists reviewed the images to confirm the presence of HT and its subtype.

All enrolled patients were randomly assigned to the training or validation cohort based on a ratio of 2:1.

Data Collection

Demographic data, medical history, clinical factors, laboratory indices, radiological data and procedural characteristics were collected for analysis. Stroke severity at baseline was assessed using the NIHSS. Pretreatment infarction core was quantified using the Alberta Stroke Program Early CT Score (ASPECTS)[11].

Definition

HT was classified into hemorrhagic infarction 1(HI1), hemorrhagic infarction 2(HI2), parenchymal hematoma 1(PH1) and parenchymal hematoma 2 (PH2) according to ECASS II classification (European Cooperative Acute Stroke Study). HI1 manifests as small hyperdense petechiae occupying less than one third of the ischemic territory. HI2 is a more confluent hyperdensity throughout the infarct zone involving greater than one-third of the ischemic territory but without mass effect. PH1 is a homogeneous hyperdensity occupying < 30% of the infarct zone with mass effect while PH2 is a homogeneous hyperdensity occupying > 30% of the infarct zone with significant mass effect[12].

Successful reperfusion was defined as a modified Thrombolysis in Cerebral Infarction score of 2b or 3[13].

Statistical analysis

Continuous variables were expressed as mean ± SD or median (interquartile range, IQR) and categorical variables as frequency (percentage). Differences in baseline characteristics between groups for continuous variables was assessed using the Student t test (normal distribution) or non-parametric Mann-Whitney U test (biased distribution), and the Chi-square test or Fisher's exact test for categorical variables.

Least absolute shrinkage and selection operator (LASSO) regression was applied to minimize the potential collinearity and overfitting of variables. The most predictive covariates were selected by the minimum (λ min). The random forests (RF) model is a popular ensemble method used for classification or regression[14]. Subsequently, the common variables identified by LASSO regression analysis and RF model were entered into multivariable logistic regression analysis using a stepwise method. Variables with p < 0.05 were selected for the development of nomogram. The sample size met the requirement of 10 events per variable to reduce bias and variability of the predictive model. The discriminative ability was measured using the area under the ROC curve (AUC). A calibration curve with bootstraps of 1000 resample was used to describe the the consistency between the predicted and observed probabilities. To estimate the clinical usefulness of the nomogram, decision curve analysis (DCA) was applied by calculating the net benefit for a range of threshold probabilities.

Statistical analysis was performed using the statistical software IBM SPSS (version 26.0) and R software (http://www.R-project.org). The following packages in R were used: glmnet, boruta, rms, pROC, and ggDCA. P value < 0.05 (2-sided) was considered statistically significant.

Baseline characteristics

A total of 205 patients were enrolled in the study, including 102 HT patients and 103 non-HT patients. Among the HT patients, the corresponding proportions of HI1, HI2, PH1 and PH2 were 36.27%, 26.47%, 14.71% and 22.55%, respctively.

The training cohort consisted of 145 patients(72 HT cases, non-HT 73 cases), and the validation cohort consisted of 60 patients (HT 30 cases, non-HT 30 cases). Flow diagram of patient inclusion and exclusion is provided in Fig. 1.

Table 1 showed the demographic,clinical information and laboratory parameters of all enrolled patients. The HT incidence was comparable between the training and validation cohorts. The baseline characteristics of patients in the training and validation cohorts are listed in Tables 2 and 3, respectively.

Table 1

Baseline characteristics of HT and non-HT patients after MT in AIS patients caused by LVO in the anterior circulation
Variables	Total (n = 205)	HT (n = 102)	non-HT (n = 103)	p
Subtypes of HT, n (%)				< 0.001
HI1	37 (18.05)	37 (36.27)	0 (0)
HI2	27 (13.17)	27 (26.47)	0 (0)
non-HT	103 (50.24)	0 (0)	103 (100)
PH1	15 (7.32)	15 (14.71)	0 (0)
PH2	23 (11.22)	23 (22.55)	0 (0)
Collateral score, n (%)				< 0.001
0	2 (0.98)	2 (1.96)	0 (0)
1A	88 (42.93)	61 (59.8)	27 (26.21)
1B	81 (39.51)	29 (28.43)	52 (50.49)
2	32 (15.61)	9 (8.82)	23 (22.33)
3	2 (0.98)	1 (0.98)	1 (0.97)
Sex, n (%)				0.598
female	75 (36.59)	35 (34.31)	40 (38.83)
male	130 (63.41)	67 (65.69)	63 (61.17)
Age, Mean ± SD(year)	68.67 ± 10.6	66.54 ± 10.82	70.79 ± 9.98	0.004
BMI, Median (Q1,Q3)(Kg/m²)	25.7 (22.8, 26.8)	25.75 (22.5, 26.7)	25.6 (23.1, 27.1)	0.944
HBP, n (%)				0.184
0	101 (49.27)	45 (44.12)	56 (54.37)
1	104 (50.73)	57 (55.88)	47 (45.63)
DM, n (%)				0.512
0	123 (60)	64 (62.75)	59 (57.28)
1	82 (40)	38 (37.25)	44 (42.72)
CHD, n (%)				1
0	118 (57.56)	59 (57.84)	59 (57.28)
1	87 (42.44)	43 (42.16)	44 (42.72)
Hyperlipemia, n (%)				0.806
0	136 (66.34)	69 (67.65)	67 (65.05)
1	69 (33.66)	33 (32.35)	36 (34.95)
AF, n (%)				0.954
0	122 (59.51)	60 (58.82)	62 (60.19)
1	83 (40.49)	42 (41.18)	41 (39.81)
Stroke TIA, n (%)				0.642
0	159 (77.56)	81 (79.41)	78 (75.73)
1	46 (22.44)	21 (20.59)	25 (24.27)
Drinking, n (%)				0.724
0	146 (71.22)	71 (69.61)	75 (72.82)
1	59 (28.78)	31 (30.39)	28 (27.18)
Smoking, n (%)				1
0	136 (66.34)	68 (66.67)	68 (66.02)
1	69 (33.66)	34 (33.33)	35 (33.98)
Anticoagulants, n (%)				0.524
0	177 (86.34)	86 (84.31)	91 (88.35)
1	28 (13.66)	16 (15.69)	12 (11.65)
Antiplatelets, n (%)				0.498
0	166 (80.98)	85 (83.33)	81 (78.64)
1	39 (19.02)	17 (16.67)	22 (21.36)
Baseline SBP, Median (Q1,Q3)(mmHg)	145 (135, 160)	149 (138, 160)	143 (135, 159)	0.141
Baseline DBP, Median (Q1,Q3)(mmHg)	94 (86, 102)	95 (87, 100)	90 (85, 102.5)	0.369
Glu, Median (Q1,Q3)(mmol/L)	6.55 (5.9, 8.14)	7.19 (6.28, 10.11)	6.17 (5.46, 7.36)	< 0.001
WBC, Mean ± SD(109/L)	8.47(7.02,10.61)	8.86 (7.07, 11.12)	8.19 (6.86, 10.03)	0.05
NEUT, Mean ± SD(109/L)	5.42 (4.06, 7.32)	6.27 (4.36, 8.69)	5 (3.6, 6.5)	< 0.001
LYMPH, Median (Q1,Q3)(109/L)	1.48 (0.97, 2.08)	1.59 (0.93, 2.24)	1.41 (1.02, 2.01)	0.303
MONO, Mean ± SD(109/L)	0.43 (0.36, 0.61)	0.44 (0.35, 0.61)	0.42 (0.36, 0.55)	0.637
RBC, Mean ± SD(1012/L)	4.42 (4.02, 4.85)	4.47 (4.08, 4.95)	4.41 (4.01, 4.82)	0.296
PLT, Median (Q1,Q3)(109/L)	198 (170, 247)	191 (166, 237.25)	209 (176, 259)	0.038
RDW, Median (Q1,Q3)(%)	0.13 (0.12, 0.13)	0.13 (0.12, 0.13)	0.13 (0.12, 0.13)	0.36
PDW, Median (Q1,Q3)(%)	12.4 (10.8, 14.4)	12.2(10.83,14.32)	12.4(10.75,14.25)	0.956
NLR, Median (Q1,Q3)	3.78 (2.41, 6.64)	4.08 (2.3, 8.84)	3.58 (2.45, 4.86)	0.074
MLR, Median (Q1,Q3)	0.29 (0.2, 0.44)	0.28 (0.18, 0.4)	0.29 (0.2, 0.52)	0.466
PLR, Median (Q1,Q3)	143.83(102.7,205.34)	128.63(93.8,203.56)	152.63(114.18,211.24)	0.03
CRP, Mean ± SD(mg/L)	7.9 (5.6, 9.8)	9.8 (8.8, 10.8)	5.6 (3.9, 7.15)	< 0.001
ALB, Mean ± SD(g/L)	38.7 (34.7, 41.2)	38.15 (33.23, 40.48)	39.2 (35.35, 42.6)	0.012
CRP/ALB, Mean ± SD	0.21 (0.15, 0.26)	0.26 (0.22, 0.3)	0.15 (0.1, 0.19)	< 0.001
PT, Mean ± SD(s)	11.6 (11.1, 11.9)	11.6 (10.95, 11.9)	11.6 (11.15, 11.95)	0.813
APTT, Mean ± SD(s)	24.8 (22.6, 27.5)	24.6 (22.22, 27.28)	25.1 (22.85, 27.7)	0.487
Fig, Median (Q1,Q3)(g/L)	2.78 (2.52, 3.4)	2.74 (2.5, 3.18)	2.91 (2.54, 3.52)	0.109
ALT, Median (Q1,Q3)(U/L)	26 (22, 33)	27 (22, 36.75)	25 (22, 29.5)	0.08
AST, Median (Q1,Q3)(U/L)	28 (25, 36)	29 (26, 40)	26 (23.3, 34.5)	0.002
ALP, Median (Q1,Q3)(U/L)	73 (59, 89)	80 (62.25, 90)	67 (57.5, 85)	0.059
r-GT, Median (Q1,Q3)(U/L)	32 (22, 53)	32 (21.25, 54.75)	34 (23, 49.5)	0.977
TBIL, Mean ± SD(umol/L)	14.2 (11.5, 18.7)	15.75 (11.72, 19.05)	13.8 (11.4, 18.5)	0.233
Cr, Median (Q1,Q3)(ug/L)	65 (56, 76)	66 (56.25, 79)	64 (55, 74.5)	0.633
BUN, Median (Q1,Q3)(mmol/L)	5.16 (4.3, 6.52)	5.44 (4.42, 7.12)	5.06 (4.1, 6.3)	0.022
UA, Median (Q1,Q3)(umol/L)	319 (269, 389)	326(277.75, 389.75)	306 (265.5, 380)	0.504
Ca²⁺, Median (IQR) (mmol/L)	2.19 (2.12, 2.27)	2.21 (2.11, 2.32)	2.18 (2.13, 2.25)	0.274
TC, Mean ± SD(mmol/L)	4.48 ± 0.76	4.51 ± 0.78	4.45 ± 0.73	0.537
TG, Mean ± SD(mmol/L)	1.72 (1.54, 1.85)	1.75 (1.56, 1.89)	1.69 (1.5, 1.85)	0.18
HDL-C, Mean ± SD(mmol/L)	1.08 (0.99, 1.18)	1.07 (0.98, 1.21)	1.08 (1.02, 1.18)	0.526
LDL-C, Median (Q1,Q3)(mmol/L)	3.09 (2.96, 3.21)	3.14 (2.98, 3.24)	3.06 (2.95, 3.18)	0.061
MONO/HDL-C, Mean ± SD	0.4 (0.31, 0.53)	0.41 (0.3, 0.54)	0.39 (0.32, 0.52)	0.643
TOAST, n (%)				0.717
CE	87 (42.44)	44 (43.14)	43 (41.75)
LAA	92 (44.88)	47 (46.08)	45 (43.69)
Others	26 (12.68)	11 (10.78)	15 (14.56)
Occlusion artery, n (%)				0.966
ACA	5 (2.44)	3 (2.94)	2 (1.94)
ICA	108 (52.68)	53 (51.96)	55 (53.4)
MCA	92 (44.88)	46 (45.1)	46 (44.66)
Occlusion site, n (%)				0.87
A1	5 (2.44)	3 (2.94)	2 (1.94)
ICA	108 (52.68)	53 (51.96)	55 (53.4)
M1	70 (34.15)	36 (35.29)	34 (33.01)
M2	21 (10.24)	9 (8.82)	12 (11.65)
M3	1 (0.49)	1 (0.98)	0 (0)
NIHSS, n (%)				< 0.001
10	17 (8.29)	0 (0)	17 (16.5)
11	17 (8.29)	0 (0)	17 (16.5)
12	35 (17.07)	3 (2.94)	32 (31.07)
13	67 (32.68)	39 (38.24)	28 (27.18)
14	29 (14.15)	23 (22.55)	6 (5.83)
15	21 (10.24)	18 (17.65)	3 (2.91)
16	16 (7.8)	16 (15.69)	0 (0)
17	3 (1.46)	3 (2.94)	0 (0)
ASPECT, n (%)				< 0.001
7	15 (7.32)	15 (14.71)	0 (0)
8	23 (11.22)	22 (21.57)	1 (0.97)
9	44 (21.46)	33 (32.35)	11 (10.68)
10	115 (56.1)	32 (31.37)	83 (80.58)
11	8 (3.9)	0 (0)	8 (7.77)
OTD, Median (Q1,Q3)(min)	180 (170, 210)	180 (180, 210)	180 (170, 200)	0.591
DTP, Median(Q1,Q3)(min)	48 (45, 51)	48 (45, 51)	48 (45, 54)	0.138
OTP, Median (Q1,Q3)(min)	231 (219, 250)	228(221.25, 251.25)	231 (215.5, 249.5)	0.929
PTR, Median (Q1,Q3)(min)	52 (48, 56)	52 (48, 56.75)	50 (47, 56)	0.384
OTR, Median (Q1,Q3)(min)	283 (268, 302)	283 (270, 303.75)	283 (264, 301.5)	0.735
Times of MT, n (%)(h)				0.805
1	106 (51.71)	51 (50)	55 (53.4)
2	90 (43.9)	47 (46.08)	43 (41.75)
3	9 (4.39)	4 (3.92)	5 (4.85)
thrombolysis, n (%)				0.416
0	161 (78.54)	83 (81.37)	78 (75.73)
1	44 (21.46)	19 (18.63)	25 (24.27)
mTICI, n (%)				< 0.001
2b	43 (20.98)	35 (34.31)	8 (7.77)
3	162 (79.02)	67 (65.69)	95 (92.23)
HT = hemorrhage transformation;BMI = body mass index;HBP = hypertension;DM = diabetes; CHD = coronary heart disease;AF = atrial fibrillation;TIA = transient ischemic attack;SBP = systolic blood pressure; DBP = diastolic blood pressure;Glu = glucose;WBC = white blood cell count; PLT = Platelets count; NEUT = neutrophil count;LYMPH = lymphocyte count;MONO = monocyte count;RBC = red blood cell count;RDW_CV = coefficient of variation of red blood cell distribution width;PDW = platelet distribution width;NLR = NEUT/LYMPH;MLR = MONO/LYMPH;PLR = PLT/LYMPH;CRP = C-reactive protein;ALB = albumin;PT = prothrombin time; APTT = activated partial thromboplastin time;Fig = fibrinogen;ALT = alanine amino-transferase; AST = aspartate amino-transferase; ALP = alkaline phosphatase; γ-GT = γ-glutamyl transpeptadase; T-BIL = total bilirubin;Cr = creatinine;BUN = blood urea nitrogen;UA = uric acid;Ca2 + = calcium;TG = triglyceride; TC = total cholesterol; LDL-C = Low-density lipoprotein cholesterol; HDL-C = High-density lipoprotein cholesterol;TOAST = toast classification;NIHSS = national Institute of Health stroke scale;ASPECT = alberta stroke program early CT score;OTD = time from morbidity to go to hospital;DTP = time from morbidity to puncture;OTP = time from morbidity to puncture;PTR = time from puncture to recanalization;OTR = time from morbidity torecanalization;MT = mechanical embolectomy;mTICI = modified thrombolysis in cerebral infarction;IQR = inter-quartile range

Table 2

Baseline characteristics of HT and non-HT patients in the training cohort
Variables	Total (n = 145)	HT (n = 72)	non-HT (n = 73)	p
Collateral score, n (%)				< 0.001
0	2 (1)	2 (3)	0 (0)
1A	60 (41)	41 (57)	19 (26)
1B	62 (43)	24 (33)	38 (52)
2	19 (13)	4 (6)	15 (21)
3	2 (1)	1 (1)	1 (1)
Sex, n (%)				0.564
female	58 (40)	31 (43)	27 (37)
male	87 (60)	41 (57)	46 (63)
Age, Mean ± SD(year)	68.94 ± 10.78	66.92 ± 11	70.95 ± 10.24	0.024
BMI, Median (Q1,Q3)(Kg/m²)	25.7 (22.4, 26.7)	25.6 (21.87, 26.6)	25.7 (23.1, 27.2)	0.174
HBP, n (%)				0.456
0	74 (51)	34 (47)	40 (55)
1	71 (49)	38 (53)	33 (45)
DM, n (%)				1
0	81 (56)	40 (56)	41 (56)
1	64 (44)	32 (44)	32 (44)
CHD, n (%)				0.944
0	84 (58)	41 (57)	43 (59)
1	61 (42)	31 (43)	30 (41)
Hyperlipemia, n (%)				1
0	91 (63)	45 (62)	46 (63)
1	54 (37)	27 (38)	27 (37)
AF, n (%)				1
0	84 (58)	42 (58)	42 (58)
1	61 (42)	30 (42)	31 (42)
Stroke TIA, n (%)				1
0	110 (76)	55 (76)	55 (75)
1	35 (24)	17 (24)	18 (25)
Drinking, n (%)				0.957
0	102 (70)	50 (69)	52 (71)
1	43 (30)	22 (31)	21 (29)
Smoking, n (%)				0.678
0	100 (69)	48 (67)	52 (71)
1	45 (31)	24 (33)	21 (29)
Anticoagulants, n (%)				1
0	123 (85)	61 (85)	62 (85)
1	22 (15)	11 (15)	11 (15)
Antiplatelets, n (%)				0.871
0	115 (79)	58 (81)	57 (78)
1	30 (21)	14 (19)	16 (22)
Baseline SBP, Median (Q1,Q3)(mmHg)	145 (135, 160)	146.5 (136, 160)	143 (135, 158)	0.409
Baseline DBP, Median (Q1,Q3)(mmHg)	90 (85, 102)	93.5 (85.75, 100)	90 (85, 102)	0.506
Glu, Median (Q1,Q3)(mmol/L)	6.6 (5.79, 8.69)	7.3 (6.29, 10.31)	6.13 (5.13, 7.3)	< 0.001
WBC, Mean ± SD(109/L)	8.46 (6.82, 10.73)	8.64 (7, 12.37)	8.08 (6.6, 9.75)	0.044
NEUT, Mean ± SD(109/L)	5.2 (3.94, 7.26)	6.26 (4.24, 9.76)	4.9 (3.6, 6.4)	< 0.001
LYMPH, Median (Q1,Q3)(109/L)	1.54 (1.02, 2.18)	1.66 (1.08, 2.37)	1.44 (0.94, 2.03)	0.102
MONO, Mean ± SD(109/L)	0.43 (0.36, 0.59)	0.44 (0.36, 0.61)	0.42 (0.36, 0.56)	0.514
RBC, Mean ± SD(1012/L)	4.37 (3.97, 4.84)	4.39 (3.96, 4.93)	4.27 (3.97, 4.79)	0.297
PLT, Median (Q1,Q3)(109/L)	197 (171, 243)	195 (169.75, 238.5)	203 (171, 244)	0.305
RDW, Median (Q1,Q3)(%)	0.13 (0.12, 0.13)	0.13 (0.12, 0.13)	0.13 (0.12, 0.13)	0.323
PDW, Median (Q1,Q3)(%)	12.5 (10.9, 14.8)	12.25 (10.88, 13.8)	12.6 (10.94, 15.3)	0.495
NLR, Median (Q1,Q3)	3.63 (2.3, 6.53)	3.89 (2.26, 8.16)	3.46 (2.43, 5.19)	0.315
MLR, Median (Q1,Q3)	0.29 (0.18, 0.44)	0.28 (0.17, 0.4)	0.29 (0.18, 0.55)	0.541
PLR, Median (Q1,Q3)	136.81 (100.82, 210.84)	125.23 (93.41, 201.69)	151.22 (103.85, 227.27)	0.034
CRP, Mean ± SD(mg/L)	8.4 (5.9, 9.8)	9.8 (8.8, 10.8)	5.9 (3.9, 7.4)	< 0.001
ALB, Mean ± SD(g/L)	37.76 ± 4.89	36.61 ± 5	38.89 ± 4.53	0.005
CRP/ALB, Mean ± SD	0.21 ± 0.09	0.28 ± 0.06	0.15 ± 0.05	< 0.001
PT, Mean ± SD(s)	11.6 (11.1, 12)	11.55 (10.9, 11.9)	11.6 (11.2, 12.1)	0.439
APTT, Mean ± SD(s)	24.98 ± 3.31	24.6 ± 3.51	25.35 ± 3.08	0.173
Fig, Median (Q1,Q3)(g/L)	2.81 (2.49, 3.38)	2.75 (2.49, 3.19)	2.91 (2.54, 3.49)	0.237
ALT, Median (Q1,Q3)(U/L)	26 (22, 33)	27 (22, 34.5)	25 (22, 29)	0.087
AST, Median (Q1,Q3)(U/L)	29 (25, 35)	29 (26, 38)	26 (23, 35)	0.028
ALP, Median (Q1,Q3)(U/L)	71 (59, 88)	78.5 (60.75, 90)	65 (59, 84)	0.131
r-GT, Median (Q1,Q3)(U/L)	32 (22, 48)	31.5 (21, 47.25)	34 (23, 53)	0.375
TBIL, Mean ± SD(umol/L)	13.8 (11.3, 18.6)	13.75 (10.9, 18.58)	13.9 (11.5, 18.6)	0.945
Cr, Median (Q1,Q3)(ug/L)	64 (54, 76)	66 (55.75, 79)	63 (54, 74)	0.536
BUN, Median (Q1,Q3)(mmol/L)	5.13 (4.3, 6.6)	5.48 (4.3, 7.16)	5.06 (4.38, 6.3)	0.09
UA, Median (Q1,Q3)(umol/L)	319 (266, 389)	326 (279.25, 389.25)	306 (262, 376)	0.522
Ca²⁺, Median (IQR) (mmol/L)	2.21 ± 0.11	2.22 ± 0.12	2.19 ± 0.1	0.106
TC, Mean ± SD(mmol/L)	4.51 ± 0.72	4.53 ± 0.75	4.49 ± 0.69	0.741
TG, Mean ± SD(mmol/L)	1.73 (1.56, 1.87)	1.75 (1.59, 1.89)	1.72 (1.52, 1.85)	0.275
HDL-C, Mean ± SD(mmol/L)	1.07 (0.99, 1.18)	1.06 (0.98, 1.21)	1.08 (1.02, 1.18)	0.678
LDL-C, Median (Q1,Q3)(mmol/L)	3.12 (2.96, 3.22)	3.15 (3.01, 3.24)	3.06 (2.95, 3.18)	0.019
MONO/HDL-C, Mean ± SD	0.39 (0.31, 0.52)	0.4 (0.32, 0.52)	0.38 (0.31, 0.52)	0.489
TOAST, n (%)				0.852
CE	65 (45)	32 (44)	33 (45)
LAA	64 (44)	33 (46)	31 (42)
Others	16 (11)	7 (10)	9 (12)
Occlusion artery, n (%)				0.88
ACA	2 (1)	1 (1)	1 (1)
ICA	73 (50)	35 (49)	38 (52)
MCA	70 (48)	36 (50)	34 (47)
Occlusion site, n (%)				0.681
A1	2 (1)	1 (1)	1 (1)
ICA	73 (50)	35 (49)	38 (52)
M1	53 (37)	29 (40)	24 (33)
M2	16 (11)	6 (8)	10 (14)
M3	1 (1)	1 (1)	0 (0)
NIHSS, n (%)				< 0.001
10	10 (7)	0 (0)	10 (14)
11	11 (8)	0 (0)	11 (15)
12	28 (19)	2 (3)	26 (36)
13	47 (32)	28 (39)	19 (26)
14	22 (15)	17 (24)	5 (7)
15	12 (8)	10 (14)	2 (3)
16	12 (8)	12 (17)	0 (0)
17	3 (2)	3 (4)	0 (0)
ASPECT, n (%)				< 0.001
7	11 (8)	11 (15)	0 (0)
8	13 (9)	12 (17)	1 (1)
9	29 (20)	24 (33)	5 (7)
10	85 (59)	25 (35)	60 (82)
11	7 (5)	0 (0)	7 (10)
OTD, Median (Q1,Q3)(min)	180 (170, 210)	180 (177.5, 210)	180 (165, 200)	0.71
DTP, Median(Q1,Q3)༈min༉	48 (45, 54)	48 (45, 51.75)	51 (45, 54)	0.512
OTP, Median (Q1,Q3)(min)	231 (219, 250)	228 (221, 252)	231 (215, 249)	0.998
PTR, Median (Q1,Q3)(min)	52 (48, 56)	53.5 (48, 57)	50 (47, 56)	0.427
OTR, Median (Q1,Q3)(min)	283.48 ± 29.47	284.51 ± 30.26	282.45 ± 28.84	0.675
Times of MT, n (%)(h)				0.819
1	77 (53)	37 (51)	40 (55)
2	61 (42)	32 (44)	29 (40)
3	7 (5)	3 (4)	4 (5)
thrombolysis, n (%)				0.253
0	112 (77)	59 (82)	53 (73)
1	33 (23)	13 (18)	20 (27)
mTICI, n (%)				0.01
2b	22 (15)	17 (24)	5 (7)
3	123 (85)	55 (76)	68 (93)
HT = hemorrhage transformation;BMI = body mass index;HBP = hypertension;DM = diabetes; CHD = coronary heart disease;AF = atrial fibrillation;TIA = transient ischemic attack;SBP = systolic blood pressure; DBP = diastolic blood pressure;Glu = glucose;WBC = white blood cell count; PLT = Platelets count; NEUT = neutrophil count;LYMPH = lymphocyte count;MONO = monocyte count;RBC = red blood cell count;RDW_CV = coefficient of variation of red blood cell distribution width;PDW = platelet distribution width;NLR = NEUT/LYMPH;MLR = MONO/LYMPH;PLR = PLT/LYMPH;CRP = C-reactive protein;ALB = albumin;PT = prothrombin time; APTT = activated partial thromboplastin time;Fig = fibrinogen;ALT = alanine amino-transferase; AST = aspartate amino-transferase; ALP = alkaline phosphatase; γ-GT = γ-glutamyl transpeptadase; T-BIL = total bilirubin;Cr = creatinine;BUN = blood urea nitrogen;UA = uric acid;Ca2 + = calcium;TG = triglyceride; TC = total cholesterol; LDL-C = Low-density lipoprotein cholesterol; HDL-C = High-density lipoprotein cholesterol;TOAST = toast classification;NIHSS = national Institute of Health stroke scale;ASPECT = alberta stroke program early CT score;OTD = time from morbidity to go to hospital;DTP = time from morbidity to puncture;OTP = time from morbidity to puncture;PTR = time from puncture to recanalization;OTR = time from morbidity torecanalization;MT = mechanical embolectomy;mTICI = modified thrombolysis in cerebral infarction;IQR = inter-quartile range

Table 3

Baseline characteristics of HT and non-HT patients in the validation cohort
Variables	Total (n = 60)	HT (n = 30)	non-HT (n = 30)	p
Collateral score, n (%)				0.006
1A	28 (47)	20 (67)	8 (27)
1B	19 (32)	5 (17)	14 (47)
2	13 (22)	5 (17)	8 (27)
Sex, n (%)				0.022
female	17 (28)	4 (13)	13 (43)
male	43 (72)	26 (87)	17 (57)
Age, Mean ± SD(year)	68.02 ± 10.19	65.63 ± 10.5	70.4 ± 9.45	0.07
BMI, Median (Q1,Q3)(Kg/m²)	25.7 (23.58, 26.9)	26.25(25.5,27.05)	25.1(23.18,26.67)	0.029
HBP, n (%)				0.299
0	27 (45)	11 (37)	16 (53)
1	33 (55)	19 (63)	14 (47)
DM, n (%)				0.159
0	42 (70)	24 (80)	18 (60)
1	18 (30)	6 (20)	12 (40)
CHD, n (%)				0.794
0	34 (57)	18 (60)	16 (53)
1	26 (43)	12 (40)	14 (47)
Hyperlipemia, n (%)				0.551
0	45 (75)	24 (80)	21 (70)
1	15 (25)	6 (20)	9 (30)
AF, n (%)				0.789
0	38 (63)	18 (60)	20 (67)
1	22 (37)	12 (40)	10 (33)
Stroke TIA, n (%)				0.505
0	49 (82)	26 (87)	23 (77)
1	11 (18)	4 (13)	7 (23)
Drinking, n (%)				0.77
0	44 (73)	21 (70)	23 (77)
1	16 (27)	9 (30)	7 (23)
Smoking, n (%)				0.429
0	36 (60)	20 (67)	16 (53)
1	24 (40)	10 (33)	14 (47)
Anticoagulants, n (%)				0.195
0	54 (90)	25 (83)	29 (97)
1	6 (10)	5 (17)	1 (3)
Antiplatelets, n (%)				0.472
0	51 (85)	27 (90)	24 (80)
1	9 (15)	3 (10)	6 (20)
Baseline SBP, Median (Q1,Q3)(mmHg)	148 (135.75, 160)	153 (139.25, 166)	141.5 (135, 159)	0.166
Baseline DBP, Median (Q1,Q3)(mmHg)	99.5 (90, 104)	100 (90, 103.75)	94 (90, 104)	0.501
Glu, Median (Q1,Q3)(mmol/L)	6.39 (5.97, 7.61)	6.78 (6.07, 7.67)	6.26 (5.91, 7.48)	0.277
WBC, Mean ± SD(109/L)	9.03 ± 2.31	9.18 ± 2.58	8.89 ± 2.04	0.637
NEUT, Mean ± SD(109/L)	6.01 ± 2.54	6.78 ± 2.4	5.25 ± 2.49	0.018
LYMPH, Median (Q1,Q3)(109/L)	1.31 (0.96, 1.86)	1.17 (0.84, 1.81)	1.38 (1.12, 1.89)	0.318
MONO, Mean ± SD(109/L)	0.47 ± 0.19	0.47 ± 0.23	0.47 ± 0.15	0.952
RBC, Mean ± SD(1012/L)	4.57 ± 0.58	4.59 ± 0.67	4.55 ± 0.49	0.785
PLT, Median (Q1,Q3)(109/L)	201(169.75,264.75)	189(157.75,228.5)	228 (180.75, 277)	0.034
RDW, Median (Q1,Q3)(%)	0.13 (0.12, 0.13)	0.13 (0.12, 0.13)	0.13 (0.12, 0.13)	0.876
PDW, Median (Q1,Q3)(%)	12.15 (10.7, 13.22)	12.2 (10.83, 14.5)	11.95 (10.45, 12.8)	0.318
NLR, Median (Q1,Q3)	4.22 (2.74, 6.76)	4.96 (2.96, 9.11)	3.73 (2.54, 4.62)	0.076
MLR, Median (Q1,Q3)	0.29 (0.21, 0.42)	0.29 (0.21, 0.4)	0.31 (0.21, 0.43)	0.722
PLR, Median (Q1,Q3)	153.61(120.03,203.12)	161.74(98.87,227.97)	153.18(131.64, 175.42)	0.796
CRP, Mean ± SD(mg/L)	7.59 ± 2.7	9.48 ± 1.95	5.7 ± 1.92	< 0.001
ALB, Mean ± SD(g/L)	38.61 ± 5.68	37.9 ± 6.46	39.31 ± 4.77	0.34
CRP/ALB, Mean ± SD	0.2 ± 0.08	0.26 ± 0.07	0.15 ± 0.05	< 0.001
PT, Mean ± SD(s)	11.5 ± 0.71	11.59 ± 0.64	11.41 ± 0.78	0.332
APTT, Mean ± SD(s)	25.18 ± 3.45	25.63 ± 3.67	24.72 ± 3.22	0.31
Fig, Median (Q1,Q3)(g/L)	2.77 (2.55, 3.47)	2.72 (2.53, 3.08)	2.87 (2.55, 3.56)	0.271
ALT, Median (Q1,Q3)(U/L)	26.5 (21.75, 36)	28 (20.25, 40.5)	26 (22.25, 29.75)	0.584
AST, Median (Q1,Q3)(U/L)	28 (25, 39.25)	31 (26, 51.75)	26.5 (24, 29.75)	0.021
ALP, Median (Q1,Q3)(U/L)	75.5 (59, 89.25)	84 (65.25, 89.75)	69 (57, 88.75)	0.29
r-GT, Median (Q1,Q3)(U/L)	33.5 (22, 55.5)	39.5 (22.25, 63.75)	31 (22.25, 44.5)	0.181
TBIL, Mean ± SD(umol/L)	15.76 ± 5.37	17.23 ± 5.95	14.29 ± 4.32	0.033
Cr, Median (Q1,Q3)(ug/L)	66 (58.75, 80.25)	66 (59, 79.75)	66.5 (58.5, 79)	0.947
BUN, Median (Q1,Q3)(mmol/L)	5.16 (4.28, 6.38)	5.3 (4.84, 6.62)	4.98 (3.9, 6.22)	0.114
UA, Median (Q1,Q3)(umol/L)	320.5(278.75,385.5)	324(273.25,389.5)	311.5(282.5, 381)	0.912
Ca²⁺, Median (IQR) (mmol/L)	2.17 (2.11, 2.26)	2.17 (2.1, 2.26)	2.17 (2.13, 2.26)	0.906
TC, Mean ± SD(mmol/L)	4.41 ± 0.85	4.48 ± 0.87	4.35 ± 0.84	0.559
TG, Mean ± SD(mmol/L)	1.68 ± 0.33	1.72 ± 0.32	1.64 ± 0.33	0.372
HDL-C, Mean ± SD(mmol/L)	1.1 ± 0.14	1.09 ± 0.15	1.11 ± 0.13	0.656
LDL-C, Median (Q1,Q3)(mmol/L)	3.08 (2.87, 3.18)	3.08 (2.81, 3.18)	3.03 (2.91, 3.19)	0.871
MONO/HDL-C, Mean ± SD	0.44 ± 0.19	0.45 ± 0.24	0.43 ± 0.15	0.739
TOAST, n (%)				0.748
CE	22 (37)	12 (40)	10 (33)
LAA	28 (47)	14 (47)	14 (47)
Others	10 (17)	4 (13)	6 (20)
Occlusion artery, n (%)				0.824
ACA	3 (5)	2 (7)	1 (3)
ICA	35 (58)	18 (60)	17 (57)
MCA	22 (37)	10 (33)	12 (40)
Occlusion site, n (%)				0.794
A1	3 (5)	2 (7)	1 (3)
ICA	35 (58)	18 (60)	17 (57)
M1	17 (28)	7 (23)	10 (33)
M2	5 (8)	3 (10)	2 (7)
NIHSS, n (%)				< 0.001
10	7 (12)	0 (0)	7 (23)
11	6 (10)	0 (0)	6 (20)
12	7 (12)	1 (3)	6 (20)
13	20 (33)	11 (37)	9 (30)
14	7 (12)	6 (20)	1 (3)
15	9 (15)	8 (27)	1 (3)
16	4 (7)	4 (13)	0 (0)
ASPECT, n (%)				< 0.001
7	4 (7)	4 (13)	0 (0)
8	10 (17)	10 (33)	0 (0)
9	15 (25)	9 (30)	6 (20)
10	30 (50)	7 (23)	23 (77)
11	1 (2)	0 (0)	1 (3)
OTD, Median (Q1,Q3)(min)	180 (180, 210)	180 (180, 210)	180 (180, 200)	0.633
DTP, n (%)				0.532
35	1 (2)	0 (0)	1 (3)
36	1 (2)	1 (3)	0 (0)
39	5 (8)	4 (13)	1 (3)
40	1 (2)	0 (0)	1 (3)
42	4 (7)	3 (10)	1 (3)
45	17 (28)	9 (30)	8 (27)
48	14 (23)	7 (23)	7 (23)
51	6 (10)	3 (10)	3 (10)
54	10 (17)	3 (10)	7 (23)
60	1 (2)	0 (0)	1 (3)
OTP, Median (Q1,Q3)(min)	230.5(221.25,250.5)	228.5(222.75,249)	231(219.75,253.25)	0.876
PTR, Median (Q1,Q3)(min)	50 (47.75, 56)	51 (48, 56)	50 (47, 55)	0.733
OTR, Median (Q1,Q3)(min)	283 (270.75, 301)	281(271.5,300.75)	283 (267, 300.5)	0.994
Times of MT, n (%)(h)				1
1	29 (48)	14 (47)	15 (50)
2	29 (48)	15 (50)	14 (47)
3	2 (3)	1 (3)	1 (3)
thrombolysis, n (%)				1
0	49 (82)	24 (80)	25 (83)
1	11 (18)	6 (20)	5 (17)
mTICI, n (%)				< 0.001
2b	21 (35)	18 (60)	3 (10)
3	39 (65)	12 (40)	27 (90)
HT = hemorrhage transformation;BMI = body mass index;HBP = hypertension;DM = diabetes; CHD = coronary heart disease;AF = atrial fibrillation;TIA = transient ischemic attack;SBP = systolic blood pressure; DBP = diastolic blood pressure;Glu = glucose;WBC = white blood cell count; PLT = Platelets count; NEUT = neutrophil count;LYMPH = lymphocyte count;MONO = monocyte count;RBC = red blood cell count;RDW_CV = coefficient of variation of red blood cell distribution width;PDW = platelet distribution width;NLR = NEUT/LYMPH;MLR = MONO/LYMPH;PLR = PLT/LYMPH;CRP = C-reactive protein;ALB = albumin;PT = prothrombin time; APTT = activated partial thromboplastin time;Fig = fibrinogen;ALT = alanine amino-transferase; AST = aspartate amino-transferase; ALP = alkaline phosphatase; γ-GGT = γ-glutamyl transpeptadase; T-BIL = total bilirubin;Cr = creatinine;BUN = blood urea nitrogen;UA = uric acid;Ca2 + = calcium;TG = triglyceride; TC = total cholesterol; LDL-C = Low-density lipoprotein cholesterol; HDL-C = High-density lipoprotein cholesterol;TOAST = toast classification;NIHSS = national Institute of Health stroke scale;ASPECT = alberta stroke program early CT score;OTD = time from morbidity to go to hospital;DTP = time from morbidity to puncture;OTP = time from morbidity to puncture;PTR = time from puncture to recanalization;OTR = time from morbidity torecanalization;MT = mechanical embolectomy;mTICI = modified thrombolysis in cerebral infarction;IQR = inter-quartile range

Feature Selection

All variables were tested for their ability to predict the post-MT HT in anterior AIS patients. The Lasso regression model was performed to screen risk factors. The Lasso coefficient profiles of features and the optimal penalization coefficient lambda (λ) are shown in Fig. 2. The results of Lasso regression showed 9 variables, including blood glucose on admission, PDW, CRP, CRP/ALB, PT, occlusion site, basic NIHSS score, ASPECTS and Times of MT, were associated with post-MT HT.

Figure 3 and Fig. 4 showed the result of random forest algorithm. Seven variables, including blood glucose on admission, neutrophil count, CRP, ALB, CRP/ALB, basic NIHSS score and ASPECTS, could be used to predict post-MT HT.

Blood glucose on admission, CRP, CRP/ALB, basic NIHSS score and ASPECTS were common predictors of post-MT HT screened by the LASSO regression and random forest model.

Multivariate logistic regression

Multivariate logistic analysis were performed to identify the optimal predictors of post-MT HT. The results are presented in Fig. 5. It was shown that high risk of HT were correlated with blood glucose on admission (Odds Ratio (OR) = 5.61, 95%CI:1.86–20.83, P = 0.004) and CRP (OR = 73.52, 95%CI: 25.06-276.77, P < 0.001).

Development of the nomogram

A nomogram was developed based on the independent variables identified by multivariate logistic regression. The predictive nomogram consisting of CRP and blood glucose on admission(Fig. 6).

The application of the nomogram is described as follows. First, assign the individual’s score for CRP and blood glucose on admission by drawing a vertical line from the variable axis to the “Points” axis. Then, cumulate the scores assigned to CRP and blood glucose on admission to obtain the total points. Finally, the probability of post-MT HT is obtained by drawing a vertical line from the “Total points” axis to “Risk” axis.

Validation, performance and clinical usefulness of the nomogram

The discriminative ability of the nomogram was good with an AUC of 0.924 in the training cohort and 0.846 in the validation cohort(Fig. 7).

A calibration plot with bootstraps of 1000 resamples was used to illustrate the degree of fit between actual observation and nomogram prediction. The calibration plot revealed good predictive accuracy of the nomogram (Fig. 8).

Decision curve analysis(DCA) was used to assess whether nomogram-based decision would improve patients’ outcomes. The results of DCA showed that using the nomogram provided a net benefit over “treat all” or “treat none” strategies, indicating the clinical usefulness of the nomogram(Fig. 9).

Mechanical thrombectomy (MT) is currently recommended as the standard treatment for AIS patients caused by intracranial large vessel occlusion[15, 16]. Despite successful reperfusion of intracranial large vessels, hemorrhage transformation is one of the common postoperative complications and is associated with poor prognosis in AIS patients[17, 18]. Therefore, early identification of patients at high-risk of post-MT HT has great clinical significance.

In this study, we used machine learning algorithms to screen risk factors and developed a nomogram based on CRP and blood glucose on admission to predict the risk of post-MT HT in anterior AIS patients. The nomogram showed good predictive ability and might help clinicians identify patients at high risk of post-MT HT.

The pathophysiological mechanism of post-MT HT is the disruption of the blood-brain barrier (BBB) and increased paracellular permeability[19, 20]. Ischemia-reperfusion injury may trigger mitochondrial dysfunction, overproduction of reactive oxygen species (ROS) and neuroinflammatory responses[21], leading to cerebral edema and intracranial hemorrhage. Several studies have shown that old age, preoperative hypertension, worse initial NIHSS score, thrombolytic therapy, long time from puncture to reperfusion, hyperglycemia on admission, and poor collateral circulation are risk factors of post-MT HT[22–27]. Consistent with previous studies, our analysis identified the baseline NIHSS score, ASPECTS, hyperglycemia on admission, CRP and CRP/ALB as independent predictors of post-MT HT.

The thromboinflammatory response may play a role in post-MT HT[28]. Neutrophil infiltration, release of pro-inflammatory cytokines, and degradation of tight junction proteins by matrix metalloproteinases(MMP) contribute to BBB disruption and HT[29–34]. C-reactive protein (CRP) has long been considered as an inflammatory biomarker[35]. Elevated CRP has been shown to be associated with high risk of HT and unfavorable outcome in AIS patients undergoing MT[36–38]. Our results suggest that CRP is an independent predictor of post-MT HT, which is consistent with the inflammatory response in post-MT HT[39, 40].

Hyperglycemia occurrs in approximately 28–40% of AIS patients and may increase the risk of HT, leading to worse clinical outcomes[7, 41–44]. Hyperglycemia is involved in brain injury in AIS patients by promoting inflammatory cascade, reducing blood flow to ischemic penumbra, destroying blood-brain barrier, inducing cerebral microcirculation dysfunction and disequilibrium between fibrinolysis and thrombosis[45, 46]. The result of our study confirmed the importance of blood glucose management in anterior AIS patients. For AIS patients undergoing mechanical thrombectomy, the monitoring of blood sugar and the management of hyperglycemia should be strengthened.

Our study has several advantages. First, all parameters in this study were routinely applied in clinical practice, so this predictive model combines accuracy with simplicity. In addition, we screened as many risk factors as possible for feature selection to identify the optimal predictors for nomogram. Furthermore,the nomogram has good discriminative performance and can provide valuable guidance for clinicians. Inevitably, our study has some limitations. First, because of the relatively small sample size and lack of external validation, the clinical generalizability of this model might be limited. Second, biomarkers such as MMP-9 and/or MMP-2 have been reported to be associated with post-MT HT. Combining biomarkers with traditional parameters may help improve the predictive value of the nomogram.

Our study developed a nomogram consisting of CRP and blood glucose on admission to predict the risk of post-MT HT in anterior AIS patients. The nomogram helps clinicians identify the high-risk population and optimize monitoring and management strategies.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of the First Affiliated Hospital of Dalian Medical University. The patients provided their written informed consent to participate in this study.

Consent for publication

Not applicable

Competing interests

All authors declare no conflict of interest.

Funding

This research was supported by the Natural science foundation of liaoning province (LZ2020068)

Author Contribution

Yuanhao Li, Chao Wen and Susuo Yang recorded the information of enrolled patients. Jianhua Li and Rui Xia preprocessed the data used in this study. Dongliang Yang analysed the data. Jian Kang critically revised the manuscript and offered constructive suggestions for this study. Yuling Li completed the design and wrote the manuscript for this study. The final manuscript version was approved by all authors.

Acknowledgments

We thank all the participants of this study for their participation.

Availability of data and materials

The datasets used in the current study are available from the corresponding author if required.

Hollist M, Morgan L, Cabatbat R, Au K, Kirmani MF, Kirmani BF. Acute Stroke Management: Overview and Recent Updates. Aging and disease. 2021;12(4):1000–9.
Silverman A, Kodali S, Sheth KN, Petersen NH. Hemodynamics and Hemorrhagic Transformation After Endovascular Therapy for Ischemic Stroke. Frontiers in neurology. 2020;11:728.
Ni H, Lu GD. Association between Infarct Location and Hemorrhagic Transformation of Acute Ischemic Stroke following Successful Recanalization after Mechanical Thrombectomy. 2022.
Merlino G, Smeralda C. Dynamic Hyperglycemic Patterns Predict Adverse Outcomes in Patients with Acute Ischemic Stroke Undergoing Mechanical Thrombectomy. 2020;9(6).
Semerano A, Laredo C, Zhao Y, Rudilosso S, Renú A, Llull L, et al. Leukocytes, Collateral Circulation, and Reperfusion in Ischemic Stroke Patients Treated With Mechanical Thrombectomy. Stroke. 2019;50(12):3456–64.
Zhang X, Xie Y, Wang H, Yang D, Jiang T, Yuan K, et al. Symptomatic Intracranial Hemorrhage After Mechanical Thrombectomy in Chinese Ischemic Stroke Patients: The ASIAN Score. Stroke. 2020;51(9):2690–6.
Janvier P, Kerleroux B. TAGE Score for Symptomatic Intracranial Hemorrhage Prediction After Successful Endovascular Treatment in Acute Ischemic Stroke. 2022;53(9):2809–17.
Tang G, Cao Z, Luo Y, Wu S, Sun X. Prognosis associated with asymptomatic intracranial hemorrhage after acute ischemic stroke: a systematic review and meta-analysis. 2022;269(7):3470–81.
Choi RY, Coyner AS, Kalpathy-Cramer J, Chiang MF, Campbell JP. Introduction to Machine Learning, Neural Networks, and Deep Learning. Translational vision science & technology. 2020;9(2):14.
Dong YM, Sun J, Li YX, Chen Q, Liu QQ, Sun Z, et al. Development and Validation of a Nomogram for Assessing Survival in Patients With COVID-19 Pneumonia. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2021;72(4):652–60.
Barber PA, Demchuk AM, Zhang J, Buchan AM. Validity and reliability of a quantitative computed tomography score in predicting outcome of hyperacute stroke before thrombolytic therapy. ASPECTS Study Group. Alberta Stroke Programme Early CT Score. Lancet. 2000;355(9216):1670–4.
Larrue V, von Kummer RR, Müller A, Bluhmki E. Risk factors for severe hemorrhagic transformation in ischemic stroke patients treated with recombinant tissue plasminogen activator: a secondary analysis of the European-Australasian Acute Stroke Study (ECASS II). Stroke. 2001;32(2):438–41.
Dargazanli C, Fahed R, Blanc R, Gory B, Labreuche J, Duhamel A, et al. Modified Thrombolysis in Cerebral Infarction 2C/Thrombolysis in Cerebral Infarction 3 Reperfusion Should Be the Aim of Mechanical Thrombectomy: Insights From the ASTER Trial (Contact Aspiration Versus Stent Retriever for Successful Revascularization). Stroke. 2018;49(5):1189–96.
Wang Y, Xia ST, Tang Q, Wu J, Zhu X. A Novel Consistent Random Forest Framework: Bernoulli Random Forests. IEEE transactions on neural networks and learning systems. 2018;29(8):3510–23.
Warner JJ, Harrington RA, Sacco RL, Elkind MSV. Guidelines for the Early Management of Patients With Acute Ischemic Stroke: 2019 Update to the 2018 Guidelines for the Early Management of Acute Ischemic Stroke. Stroke. 2019;50(12):3331–2.
Rodgers ML, Fox E, Abdelhak T, Franker LM, Johnson BJ, Kirchner-Sullivan C, et al. Care of the Patient With Acute Ischemic Stroke (Endovascular/Intensive Care Unit-Postinterventional Therapy): Update to 2009 Comprehensive Nursing Care Scientific Statement: A Scientific Statement From the American Heart Association. Stroke. 2021;52(5):e198-e210.
Huang K, Zha M. Increased intracranial hemorrhage of mechanical thrombectomy in acute ischemic stroke patients with atrial fibrillation. 2021;51(2):536–44.
Balami JS, White PM, McMeekin PJ, Ford GA, Buchan AM. Complications of endovascular treatment for acute ischemic stroke: Prevention and management. International journal of stroke: official journal of the International Stroke Society. 2018;13(4):348–61.
Bernardo-Castro S, Sousa JA, Brás A, Cecília C, Rodrigues B, Almendra L, et al. Pathophysiology of Blood-Brain Barrier Permeability Throughout the Different Stages of Ischemic Stroke and Its Implication on Hemorrhagic Transformation and Recovery. Frontiers in neurology. 2020;11:594672.
Krishnan R, Mays W, Elijovich L. Complications of Mechanical Thrombectomy in Acute Ischemic Stroke. Neurology. 2021;97(20 Suppl 2):S115-s25.
Khatri R, McKinney AM, Swenson B, Janardhan V. Blood-brain barrier, reperfusion injury, and hemorrhagic transformation in acute ischemic stroke. Neurology. 2012;79(13 Suppl 1):S52-7.
Baracchini C, Farina F, Palmieri A, Kulyk C, Pieroni A, Viaro F, et al. Early hemodynamic predictors of good outcome and reperfusion injury after endovascular treatment. Neurology. 2019;92(24):e2774-e83.
Nawabi J, Kniep H, Schon G, Flottmann F, Leischner H, Kabiri R, et al. Hemorrhage After Endovascular Recanalization in Acute Stroke: Lesion Extent, Collaterals and Degree of Ischemic Water Uptake Mediate Tissue Vulnerability. Frontiers in neurology. 2019;10:569.
Hao Y, Yang D, Wang H, Zi W, Zhang M, Geng Y, et al. Predictors for Symptomatic Intracranial Hemorrhage After Endovascular Treatment of Acute Ischemic Stroke. Stroke. 2017;48(5):1203–9.
Lee JS, Bang OY. Collateral Status and Outcomes after Thrombectomy. Translational stroke research. 2022.
Uniken Venema SM, Dankbaar JW, Wolff L, van Es A, Sprengers M, van der Lugt A, et al. Collateral status and recanalization after endovascular treatment for acute ischemic stroke. Journal of neurointerventional surgery. 2022.
Tian B, Tian X, Shi Z, Peng W, Zhang X, Yang P, et al. Clinical and Imaging Indicators of Hemorrhagic Transformation in Acute Ischemic Stroke After Endovascular Thrombectomy. Stroke. 2022;53(5):1674–81.
Spronk E, Sykes G, Falcione S, Munsterman D, Joy T, Kamtchum-Tatuene J, et al. Hemorrhagic Transformation in Ischemic Stroke and the Role of Inflammation. Frontiers in neurology. 2021;12:661955.
Winneberger J, Schöls S, Lessmann K, Rández-Garbayo J, Bauer AT, Mohamud Yusuf A, et al. Platelet endothelial cell adhesion molecule-1 is a gatekeeper of neutrophil transendothelial migration in ischemic stroke. Brain, behavior, and immunity. 2021;93:277–87.
Franke M, Bieber M, Kraft P, Weber ANR, Stoll G, Schuhmann MK. The NLRP3 inflammasome drives inflammation in ischemia/reperfusion injury after transient middle cerebral artery occlusion in mice. Brain, behavior, and immunity. 2021;92:223–33.
Chen H, Luo Y, Tsoi B, Gu B, Qi S, Shen J. Angong Niuhuang Wan reduces hemorrhagic transformation and mortality in ischemic stroke rats with delayed thrombolysis: involvement of peroxynitrite-mediated MMP-9 activation. 2022;17(1):51.
Chen H, He Y, Chen S, Qi S, Shen J. Therapeutic targets of oxidative/nitrosative stress and neuroinflammation in ischemic stroke: Applications for natural product efficacy with omics and systemic biology. Pharmacological research. 2020;158:104877.
Shi K, Zou M, Jia DM, Shi S, Yang X, Liu Q, et al. tPA Mobilizes Immune Cells That Exacerbate Hemorrhagic Transformation in Stroke. Circulation research. 2021;128(1):62–75.
Przykaza Ł. Understanding the Connection Between Common Stroke Comorbidities, Their Associated Inflammation, and the Course of the Cerebral Ischemia/Reperfusion Cascade. Frontiers in immunology. 2021;12:782569.
Cai Y, Zhang H, Li Q, Zhang P. Correlation Between Blood Glucose Variability and Early Therapeutic Effects After Intravenous Thrombolysis With Alteplase and Levels of Serum Inflammatory Factors in Patients With Acute Ischemic Stroke. Frontiers in neurology. 2022;13:806013.
Zang N, Lin Z, Huang K, Pan Y, Wu Y, Wu Y, et al. Biomarkers of Unfavorable Outcome in Acute Ischemic Stroke Patients with Successful Recanalization by Endovascular Thrombectomy. Cerebrovascular diseases (Basel, Switzerland). 2020;49(6):583–92.
Kim S, Yi HJ, Lee DH, Sung JH. Association of High-sensitivity C-reactive Protein with Patient Prognosis Following Mechanical Thrombectomy for Acute Ischemic Stroke. Current neurovascular research. 2020;17(4):402–10.
Chen Z, He Y, Su Y, Sun Y, Zhang Y, Chen H. Association of inflammatory and platelet volume markers with clinical outcome in patients with anterior circulation ischaemic stroke after endovascular thrombectomy. Neurological research. 2021;43(6):503–10.
Mizuma A, Yenari MA. Anti-Inflammatory Targets for the Treatment of Reperfusion Injury in Stroke. Frontiers in neurology. 2017;8:467.
Haupt M, Gerner ST, Bähr M, Doeppner TR. Neuroprotective Strategies for Ischemic Stroke-Future Perspectives. International journal of molecular sciences. 2023;24(5).
Johnston KC, Bruno A, Pauls Q, Hall CE, Barrett KM, Barsan W, et al. Intensive vs Standard Treatment of Hyperglycemia and Functional Outcome in Patients With Acute Ischemic Stroke: The SHINE Randomized Clinical Trial. Jama. 2019;322(4):326–35.
Perez-Vega C, Domingo RA. Influence of glucose levels on clinical outcome after mechanical thrombectomy for large-vessel occlusion: a systematic review and meta-analysis. 2022;14(1).
Liu C, Zhang Y, Li X, Liu Y, Jiang T, Wang M, et al. A Glycemia-Based Nomogram for Predicting Outcome in Stroke Patients after Endovascular Treatment. 2022;12(11).
Gelhard S, Kestner RI, Armbrust M, Steinmetz H, Foerch C, Bohmann FO. Exploring Contraindications for Thrombolysis: Risk of Hemorrhagic Transformation and Neurological Deterioration after Thrombolysis in Mice with Recent Ischemic Stroke and Hyperglycemia. 2022;11(12).
Lu GD, Ren ZQ, Zhang JX, Zu QQ, Shi HB. Effects of Diabetes Mellitus and Admission Glucose in Patients Receiving Mechanical Thrombectomy: A Systematic Review and Meta-analysis. 2018;29(3):426–34.
Rom S, Heldt NA, Gajghate S, Seliga A, Reichenbach NL, Persidsky Y. Hyperglycemia and advanced glycation end products disrupt BBB and promote occludin and claudin-5 protein secretion on extracellular microvesicles. Scientific reports. 2020;10(1):7274.

No competing interests reported.

Development of a nomogram to predict hemorrhage transformation after mechanical thrombectomy in patients with acute ischemic stroke caused by large vessel occlusion in the anterior circulation

Status:

Version 1

Abstract

Figures

Introduction

Methods

Study design

Data Collection

Definition

Statistical analysis

Results

Baseline characteristics

Feature Selection

Multivariate logistic regression

Development of the nomogram

Validation, performance and clinical usefulness of the nomogram

Discussion

Conclusion

Declarations

Declarations

Funding

Author Contribution

Acknowledgments

Availability of data and materials

References

Additional Declarations

Status:

Version 1