Does Malignancy Status Effect Outcomes in Patients With Large Vessel Occlusion Stroke and Cancer Who Underwent Endovascular Thrombectomy?

Abstract

Background

Cancer is associated with an increased risk of acute ischemic stroke, including large vessel occlusions. Whether cancer status affects outcomes in patients with large vessel occlusions that undergo endovascular thrombectomy remains unknown.

Methods and Results

All consecutive patients undergoing endovascular thrombectomy for large vessel occlusions were recruited into a prospective ongoing multicenter database, and the data were retrospectively analyzed. Patients with active cancer were compared with patients with cancer in remission. Association of cancer status with 90‐day functional outcome and mortality were calculated in multivariable analyses. We identified 154 patients with cancer and large vessel occlusions that underwent endovascular thrombectomy (mean age, 74±11; 43% men; median National Institutes of Health Stroke Scale 15). Of the included patients, 70 (46%) had a remote history of cancer or cancer in remission, and 84 (54%) had active disease. Outcome data at 90 days poststroke were available for 138 patients (90%) and was classified as favorable in 53 (38%). Patients with active cancer were younger and more often smoked but did not significantly differ from those without malignancy in other risk factors, stroke severity, stroke subtype, or procedural variables. Favorable outcome rates among patients with active cancer did not significantly differ compared with those seen in patients without active cancer, but mortality rates were significantly higher among patients with active cancer on univariate and multivariable analyses.

Conclusions

Our study suggests that endovascular thrombectomy is safe and efficacious in patients with history of malignancy as well as in those with active cancer at the time of stroke onset, although mortality rates are higher among patients with active cancer.

Nonstandard Abbreviations and Acronyms

EVT

endovascular thrombectomy

LVO

large vessel occlusion

mRS

Modified Rankin Scale

sICH

symptomatic intracerebral hemorrhage

TICI

thrombolysis in cerebral infarction

Clinical Perspective.

What Is New?

• The results of the current study suggest that endovascular thrombectomy is safe and efficacious in patients with history of malignancy as well as in those with active cancer at the time of stroke onset, although mortality rates are significantly higher and favorable outcome rates tend to be lower among patients with active cancer.

What Are the Clinical Implications?

• The results imply that all patients with cancer and large vessel occlusion stroke should be considered for endovascular thrombectomy regardless of cancer status.

The prevalence of cerebrovascular disease in the setting of malignancy is ≈15%. ¹ , ² , ³ Cancer associated coagulopathy is the main pathological mechanism underlying cancer‐related stroke. ¹ , ² , ³ Other causes including radiation chemotherapy and biological therapy are all associated with potential vessel wall injury and coagulopathy. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ Additionally, many patients with cancer are prone to have stroke secondary to concomitant atherosclerotic disease secondary to hypertension, diabetes, hyperlipidemia, and smoking as well as to atrial fibrillation. Patients with stroke secondary to large vessel occlusion (LVO) have higher rates of recanalization and favorable outcomes when treated with endovascular thrombectomy (EVT). ¹⁰ Several case series published lately suggest that EVT may be as effective in achieving target vessel recanalization in patients with cancer with stroke and LVO, although the chances for favorable outcome are somewhat reduced in patients with cancer. ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ However, only 1 of these studies differentiated between patients with active cancer to those with cancer in remission. ²⁴ Therefore, our goal was to determine whether cancer state at the time of stroke affects outcomes in patients with cancer with LVO that underwent EVT.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request in compliance with governmental rules.

We conducted a retrospective analysis of our prospective multicenter database for all patients arriving with LVO that underwent EVT at 3 tertiary academic centers. Data from all patients undergoing EVT were accrued continuously without specific exclusion criteria. The study was approved at each individual center by the local institutional review board, and a permission to combine anonymized individual patient data into a unified data set was granted with a waiver to obtain informed consent, given the retrospective nature of the data analysis.

We identified patients with a history of remote cancer in remission or suffering from an active malignancy at the time of presentation. These patients were identified based on personal history obtained using a standardized questionnaire administered to all patients and family members or from medical records when the questionnaire could not be filled (eg, in aphasic patients with no family member present). We compared patients with active cancer to those with history of cancer but no evidence of active disease. For the purpose of the current study, active cancer was defined as a diagnosis of cancer, other than local basal‐cell carcinoma of the skin within 180 days before LVO, any treatment for cancer within 180 days before LVO, new diagnosis of cancer during admission for LVO, or recurrent or metastatic cancer at the time of LVO. ²⁵

Clinically relevant accrued data included demographics, time metrics, baseline stroke severity determined with the National Institutes of Health Stroke Scale, and outcome status determined with the modified Rankin Scale (mRS) scores at discharge and 3 months after discharge. Radiological data acquired before the endovascular treatment included non‐contrast computed tomography (NCCT) or magnetic resonance imaging and computed tomography angiography (CTA).

Treatment with intravenous tissue plasminogen activator (tPA) was given according to recommendations for thrombolytic treatment in the institutional guidelines. A history of cancer was not considered a contraindication for tPA unless the presence of brain metastasis was known or suspected, or the patient was considered to present high risk of systemic bleeding related to the cancer or systemic metastasis.

Experienced interventional specialists treated all patients undergoing EVT. All procedures were performed using a femoral artery approach under general anesthesia. A balloon guide catheter was routinely used unless technically unfeasible, and a heparinized saline solution was continuously perfused through the catheter during the procedure. Mechanical thrombectomy using stent retrievers or aspiration techniques were attempted at the discretion of the endovascular specialists.

Data on procedural variables including the thrombolysis in cerebral infarction (TICI) score ²⁶ at the end of the procedure and the number of passes needed to achieve the best possible recanalization status were also studied. TICI2b‐3 was considered as successful target vessel recanalization.

Patients were seen after discharge at the outpatient stroke clinics. We also obtained the information on mortality by reviewing medical records.

Primary outcomes included functional outcome determined with the modified Rankin Scale (mRS) at 3‐month post‐EVT. Favorable outcome was defined as either an mRS ≤2 at 90 days for patients that were independent before stroke and no change in mRS for those with pre‐existing disability (mRS ≥3) before the stroke.

Secondary safety outcomes were symptomatic hemorrhagic transformation (symptomatic intracranial hemorrhage [sICH]) according to the European Cooperative Acute Stroke Study criteria ²⁷ and death.

Statistical Analysis

Statistical analysis was performed using the SPSS 25 (IBM USA). P<0.05 was considered significant. The ᵡ² test was used to explore the link between qualitative variables. The Student t‐test was used to compare quantitative variables. We next performed multivariable logistic regression modeling to test the impact on favorable outcome and survival. Only variables with P<0.1 were entered into our regression analysis.

Results

Among 1263 patients with acute LVO that underwent EVT and were included in the registry, 154 patients (43% men, median age, 74 years [interquartile range [IQR], 67–82]) had a concomitant history of cancer (Figure). The most common cancer types included breast (21%), lung (14%), colon (11%), and lymphoma (6%). The median National Institutes of Health Stroke Scale was 15 (IQR, 15–20) on admission and 5 (IQR, 5–12) at discharge. Favorable target vessel recanalization (TICI ≥2b) was achieved in 132 (85%) patients, and the median number of passes needed to recanalize the vessel was 2 (IQR, 1–3). sICH was observed in 10 patients (6.5%). Outcome data at 90 days poststroke was available for 139 (90%) patients, with 15 patients lost to follow‐up. Among patients with available follow‐up data, 54 (39%) achieved a favorable functional outcome, and 44 (34%) were dead before 90 days from onset.

Figure 1. Study flowchart and outcomes.

Cancer status and EVT results in patients with cancer and LVO stroke. EVT indicates endovascular thrombectomy; ICH, intracerebral hemorrhage; and LVO, large vessel occlusion.

Of the included patients with cancer, 84 (54%) had an active malignancy at the time of presentation (mean age±SD, 71.6±12.2, 51% men), while 70 (46%) had a history of cancer with no known activity at the time of EVT (mean age±SD, 75.5±11.1, 39% men). Comparison of patients with active cancer to those with a history of cancer (Table 1) showed that patients with active cancer were significantly younger and were more often smokers. However, other baseline risk factors as well as stroke severity, suspected stroke pathogenesis, and site of vessel occlusion did not differ between the groups. Similarly, the frequency of use of bridging with systemic thrombolysis did not differ between the groups. Procedural variables including time metrics and number of passes needed to achieve target vessel recanalization also did not significantly differ between the groups. Favorable target vessel recanalization defined as TICI2b‐3 was achieved in 86% of patients in both groups. Only 10 patients were found to have symptomatic hemorrhagic transformations, and there was no significant difference in the frequency of ICH between the groups.

Table 1.

Baseline Characteristics of the Patients

	All (N=154)	Active cancer n=84	Cancer history n=70	P value
Age, (±SD)	73.4±11.7	71.6±12.2	75.5±10.8	0.04
Sex, men (%)	71 (46%)	43 (51%)	28 (40%)	0.16
Hypertension (%)	102 (66%)	51 (61%)	51 (73%)	0.11
Diabetes (%)	47 (31%)	24 (29%)	23 (33%)	0.56
Dyslipidemia (%)	70 (46%)	41 (49%)	29 (41%)	0.08
AF (%)	62 (40%)	32 (38%)	30 (43%)	0.08
Smoking (%)	39 (25%)	31 (37%)	8 (11%)	0.001
IHD (%)	62 (40%)	32 (38%)	30 (43%)	0.08
Prior stroke (%)	27 (17.5%)	13 (15.5%)	14 (20%)	0.46
Antiplatelets	31 (21%)	19 (24%)	12 (18%)	0.38
Anticoagulants	31 (20%)	18 (21%)	13 (19%)	0.66
Vessel lesion (%)				0.41
ICA	23 (15%)	13 (15%)	10 (14%)
MCA	93 (60%)	47 (56%)	46 (66%)
Tandem	8 (5%)	5 (6%)	3 (4%)
Other	30 (20%)	19 (23%)	11 (16%)
TOAST (%)				0.07
Cardioembolic	67 (44%)	36 (43%)	31 (44%)
Large‐vessel atherosclerosis	57 (37%)	26 (31%)	31 (44%)
Other known	6 (4%)	5 (6%)	1 (1%)
Unknown	24 (16%)	17 (20%)	7 (10%)
Initial NIHSS score (±SD)	15.1±6.0	14.6±6.3	15.6±5.5	0.30
IV tPA (%)	38 (25%)	16 (19%)	22 (31%)	0.08
Baseline mRS (%)				0.32
0–2	125 (81%)	66 (78%)	59 (84%)
3	23 (15%)	14 (17%)	9 (13%)
4	6 (4%)	4 (5%)	2 (3%)

AF indicates atrial fibrillation; ICA, internal carotid artery; IHD, ischemic heart disease; IV tPA, intravenous tissue plasminogen activator; MCA, middle cerebral artery; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; sICH, symptomatic intracerebral hemorrhage; TICI, thrombolysis in cerebral infarction; and TOAST, Trial of Org 10172 in Acute Stroke Treatment.

There were no significant differences in the rates of favorable outcome at discharge as defined by an mRS ≤2 or no change from baseline for those arriving with an mRS ≥3 (Table 2). However, mortality rates during the acute admission were doubled in patients with active cancer (28% versus 14%, P=0.04).

Table 2.

Outcomes

	All (n=154)	Active cancer (n=84)	Cancer history (n=70)	P value
Favorable recanalization (TICI 2b‐3, %)	132 (85%)	72 (86%)	60 (86%)	1
Discharge favorable outcome (n=149, %)	44 (29%)	25 (32%)	19 (27%)	0.55
Death during admission (n=149, %)	32 (21%)	22 (28%)	10 (14%)	0.04
90‐d favorable outcome (n=139, %)	55 (39%)	24 (32%)	30 (47%)	0.08
sICH (n=149, %)	10 (7%)	5 (6%)	5 (8%)	0.08
Death at 90 d (n=141, %)	44 (34%)	30 (38%)	14 (21%)	0.04

sICH indicates symptomatic intracerebral hemorrhage; and TICI, thrombolysis in cerebral infarction.

Similarly, favorable outcome at 90 days was seen in 32% of the patients with active cancer and in 47% of patients in the group of patients with a previous history of malignancy but no evidence of active disease (P=0.08). The 90‐day mortality rates were significantly higher in the patients with active cancer (38% versus 21%, P=0.04), and patients that survived were more often independent before stroke and more often reached target vessel recanalization (Table S1).

We next compared patients with favorable and unfavorable outcomes on day 90 after stroke. The results (Table 3) show that patients with favorable outcomes less often had diabetes, were more often independent at the time of stroke onset, and more often reached favorable target vessel recanalization. On multivariable analysis (Table 4) that included variables yielding a significance value of <0.1 on the univariate analyses, the presence of active cancer was not found as an independent predictor of favorable functional independence at 90 days poststroke. In contrast, being independent before stroke and achieving a favorable target vessel recanalization during EVT were positively associated with increased chances of achieving an independent functional state at 90 days poststroke.

Table 3.

Factors Associated With Favorable Outcome

	Favorable outcomes n=55	Unfavorable outcomes n=84	P value
Age, (±SD)	72.5±12.2	73.4±12.1	0.68
Sex, men (%)	21 (38%)	43 (51%)	0.13
Hypertension (%)	30 (63%)	58 (69%)	0.08
Diabetes (%)	10 (18%)	29 (34%)	0.04
Dyslipidemia (%)	20 (36%)	36 (43%)	0.44
Atrial fibrillation (%)	27 (49%)	30 (36%)	0.12
Smoking (%)	11 (20%)	26 (31%)	0.15
Ischemic heart disease (%)	20 (36%)	34 (41%)	0.63
Prior stroke (%)	10 (18%)	14 (17%)	0.82
Antiplatelets	11 (20%)	16 (19%)	0.89
Anticoagulants	16 (29%)	14 (17%)	0.08
Active cancer (%)	24 (44%)	50 (59%)	0.08
Vessel lesion (%)			0.45
Internal carotid	6 (11%)	16 (19%)
Middle cerebral	36 (66%)	49 (58%)
Tandem	2 (4%)	6 (7%)
Other	11 (19%)	13 (16%)
TOAST (%)			0.10
Cardioembolic	31 (56%)	33 (39%)
Large vessel atherosclerosis	19 (35%)	31 (37%)
Other known	1 (2%)	5 (6%)
Unknown	4 (7%)	15 (11%)
Baseline mRS (%)			0.003
0–2	51 (93%)	59 (71%)
3	3 (5%)	19 (23%)
4	1 (2%)	6 (6%)
Initial NIHSS score (±SD)	14.6±5.7	15.6±5.9	0.34
IV tPA (%)	16 (29%)	17 (20%)	0.23
TICI2b‐3 (%)	54 (98%)	67 (80%)	0.002
No. of passes (mean±SD)	1.8±1.2	3.1±4.4	0.06
sICH (%)	1 (2%)	9 (10%)	0.05

IV tPA indicates intravenous tissue plasminogen activator; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; sICH, symptomatic intracerebral hemorrhage; TICI, thrombolysis in cerebral infarction; and TOAST, Trial of Org 10172 in Acute Stroke Treatment.

Table 4.

Multivariable Regression for Favorable Outcomes at 90 Days Poststroke

Variable	OR
95% CI		P value
Diabetes	0.80	0.30	2.13	0.65
Hypertension	0.45	0.19	1.02	0.07
On oral anticoagulant	2.30	0.89	5.94	0.09
Active cancer	0.49	0.22	1.10	0.08
Independent prior	3.96	1.19	13.21	0.02
TICI 2b‐3	10.01	1.20	83.55	0.03
sICH	0.18	0.02	1.66	0.13

sICH indicates symptomatic intracerebral hemorrhage; and TICI, thrombolysis in cerebral infarction.

In contrast, on multivariable analysis, the presence of active cancer remained an independent negative predictor of survival at 90 days poststroke (odds ratio [OR], 0.41 [95% CI, 0.17–0.97]), while favorable recanalization to TICI2b‐3 (OR, 3.66 [95% CI, 1.14–2.26]) and being independent before stroke (OR, 6.24 [95% CI, 2.26–17.22]) were positively associated with survival (Table 5).

Table 5.

Multivariable Regression for Survival at 90 Days

	OR	95% CI		P value
TICI 2b‐3	3.66	1.15	2.26	0.03
Active cancer	0.41	0.17	0.97	0.04
Independent status before stroke	6.24	2.26	17.22	<0.001
tPA	0.46	0.67	1.23	0.462
Internal carotid occlusion	2.56	0.80	13.37	0.47

OR indicates odds ratio; TICI, thrombolysis in cerebral infarction; and tPA, tissue plasminogen activator.

Discussion

The main findings of the current multicenter analysis are that outcomes tend to be less often favorable, and mortality is significantly higher in patients with active cancer when compared with patients with a remote history of cancer. Importantly, stroke severity, occluded vessel location, favorable target vessel recanalization, and sICH rates were similar among the groups and could not account for these findings. Patients with active cancer were younger and also showed trends toward being more likely to be treated with bridging tPA as well as trends toward lower rates of having atrial fibrillation and large vessel atherothrombosis, all of which were expected to tilt mortality and favorable outcome rates toward better outcomes. However, mortality rates were significantly higher in patients with active cancer, and an active cancer status was associated with numerically lower rates of favorable outcome, although this difference did not reach statistical significance.

Patients with cancer are at increased risk for stroke, which may be related to a cancer‐associated hypercoagulable state as well as to vessel wall damage secondary to radiation, biological, and chemical treatments. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ Although the stroke risk decreases over time, it remains elevated even a decade following the cancer diagnosis. In addition, the prognosis is poorer in patients with stroke with systemic malignancy than in those without. ²¹ , ²⁸

Patients with acute ischemic stroke and cancer can be treated with intravenous thrombolysis with recombinant tissue plasminogen activator. ²⁴ Certain limitations, such as bleeding tendencies, time from symptom onset, and presence of LVO may limit the use and effectiveness of intravenous thrombolysis. EVT is the treatment of choice for LVO stroke. ¹⁰ Several studies have explored the safety and effectiveness of EVT in patients with malignancy. ¹² , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²² , ²³ , ²⁴ These studies were based on retrospective single‐center data with a low number of included patients, ¹¹ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ on nationwide registry data that did not offer specifics about cancer type, ²² , ²⁴ and only 1 prospective randomized trial. ²³ Data from the nationwide registries and the randomized controlled study showed that patients with cancer comprise between 4.5% and 6.7% of all LVO patients treated with EVT, which is slightly lower than the 12% seen in the current study. These differences may stem from different definitions of cancer and cancer in remission, which were used in different studies. Thus, we included patients with a history of remote cancer even years ago who were probably not included in patients with cancer cohorts in some of the previous studies. Taken together, these studies recorded similar rates of favorable target vessel recanalization and sICH ¹¹ , ¹² , ¹⁴ , ¹⁶ , ¹⁷ , ¹⁸ , ²⁰ , ²² , ²³ , ²⁴ but lower rates of favorable outcome and increased mortality rates in patients with cancer when compared with patients without cancer. ¹¹ , ¹² , ¹⁶ , ¹⁷ , ¹⁸ , ²⁰ , ²² , ²³ , ²⁴ Only 1 prior study compared patients with active cancer to those with nonactive cancer in remission. ²⁴ This is surprising in light of data suggesting that these populations may differ in stroke cause and distribution. ²⁹ The study, ²⁴ which was based on national registry data in Korea, showed that the outcomes of patients with active cancer are less favorable compared with those with cancer in remission, which in turn resembles those outcomes of patients without cancer. The current data confirm these results and also shows higher mortality rates in patients with active cancer compared with those with remote cancer. Of note, mortality rates were higher in patients with active cancer both during the admission for the acute stroke and at 90 days after stroke. Most of the deaths that occurred between discharge and 90 days occurred in the active cancer group, suggesting that at least some of the deaths observed in patients with cancer at 90 days were related to the cancer itself and not to the stroke. However, because we did not have complete access to the causes of death, we could not ascertain this hypothesis.

In these regards, the current study presents novel data suggesting that patients with active cancer represent a group of high‐risk patients with especially poor outcome, despite achieving similarly high recanalization rates and low sICH rates compared with patients with a remote history of cancer. These data further suggest that at least some patients with cancer, including one‐third of those with active cancer and nearly half of the patients with remote cancer, can benefit from EVT and have favorable outcomes.

The current study has several strengths. First, we are reporting on multicenter prospectively accrued data that reflect current everyday practice in centers that are EVT capable. As such, it provides much needed information for clinicians dealing with stroke in patients with cancer at the emergency department level. Second, the included number of patients is relatively large, and the frequency with which cancer associated LVO was observed is larger than that seen in larger nationwide registry data in the United States or Korea.

Our study has limitations. First, although data collection was prospective, the data analysis was retrospective and therefore prone to bias. Second, we do not have information about patients with cancer who were excluded from EVT because of reduced life expectancy or other family‐ or physician‐based decisions. Third, we did not have access to causes of mortality in the included patients, limiting our efforts to determine if mortality was secondary to cancer or cancer therapy or stroke‐related causes.

Conclusions

Our study suggests that EVT is safe and efficacious in patients with history of malignancy at the time of stroke as well as in those with active cancer. While outcomes appear to be better in patients with a history of cancer, these results may suggest that all patients with cancer with LVO may benefit from EVT, despite higher mortality rates and less favorable rates seen in patients with active cancer. Future studies identifying subpopulations of patients with active cancer who may benefit more from EVT in contrast to those in which futile recanalization occurs are needed.

Sources of Funding

This study was supported in part by an unrestricted grant from the Peritz and Chantal Scheinberg Cerebrovascular Research Fund.

Disclosures

Dr. Leker received speaker honoraria from iSchemaView, Boehringer Ingelheim, Pfizer, Jansen, Biogen, Medtronic, and Abbott and advisory board honoraria from Jansen. The remaining authors have no disclosures to report.

Supporting information

Table S1.