Hyperdense sign as a predictor for successful recanalization and clinical outcome in acute ischemic stroke: A systematic review and meta-analysis

Abstract

Objective

To assess the prognostic values of hyperdense sign on pretreatment non-contrast head CT scan for successful recanalization (mTICI ≥2b) and 90-day good functional outcome (mRs 0-2) in patients with acute ischemic stroke undergoing mechanical thrombectomy (MT).

Methods

Literature search on PubMed, EMBASE, and Cochrane databases from inception up to 1 November 2023 was conducted. Twelve studies which reported hyperdense sign, recanalization and clinical outcomes were included in qualitative synthesis and meta-analysis.

Results

Pooled analysis demonstrated a statistically significant association between successful recanalization and hyperdense sign-positive patients who underwent MT (odd ratios (OR) = 1.47, 95% confidence interval (CI) = 1.03–2.10, p = 0.04). No statistically significant association was demonstrated between presence of hyperdense sign and good functional outcome (OR = 1.04, 95% CI: 0.72–1.49, p = 0.85) or symptomatic intracranial hemorrhage sICH (OR: 1.80, 95% CI 0.72–4.47, p = 0.21)

Conclusions

This meta-analysis demonstrated that pre-intervention hyperdense sign on CT imaging might be useful in prediction of successful recanalization after MT.

Introduction

In the evaluation of acute ischemic stroke (AIS), a crucial aspect lies in the utilization of non-contrast computed tomography (NCCT) of the brain, serving a multifaceted purpose beyond the detection of early ischemic changes and the exclusion of hemorrhage. It plays an additional role in identifying other imaging biomarkers, among which the hyperdense vessel sign (HDVS) holds significance. This distinctive sign is characterized by focal hyper-attenuation within the blood vessel on non-contrast brain CT scans. ¹ While the HDVS is deemed a specific indicator, demonstrating a noteworthy ability to discern arterial obstruction, it falls short in sensitivity. ² The relationship between the time elapsed after a stroke and the accuracy of detecting HDVS has been a subject of investigation in the published literature, yielding conflicting findings challenging the notion that early imaging within a specific time frame significantly impacts HDVS detection.^3–5

Prior studies have endeavored to assess recanalization and outcomes following intravenous thrombolysis in patients with HDVS. ⁶ Nevertheless, to our knowledge, this represents the first meta-analysis aimed at discerning whether the presence of HDVS predicts successful recanalization and functional outcomes in AIS patients undergoing mechanical thrombectomy (MT).

Methods

We conducted this systematic review and meta-analysis based on the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA). ⁷ Data is available on request to the corresponding authors.

Data source and search strategy

We searched PubMed, Embase, and Cochrane Library to obtain articles in all languages from inception until 1 November 2023. “Ischemic Stroke,” “Thrombectomy,” and “Hyperdense sign” were the search terms. Synonyms were obtained from PubMed and Embase with the elimination of duplicates. Supplementary Table 1 describes the full search term used in each database searched.

Eligibility criteria

The inclusion criteria for our systematic review and meta-analysis were as follows: (a) Retrospective or prospective observational studies, (b) that performed a direct comparison between patients with or without HDVS on NCCT scan prior to MT, (c) reporting of recanalization grade after MT. Conference abstracts, single-arm studies, and case reports were excluded. Two investigators (KG and MS) independently screened and selected the studies for the final review. Discrepancies were resolved by a third investigator (MJ).

Data extraction

Extracted data included study design, country and year of the study, baseline characteristics, primary, and secondary outcomes. Two investigators (KG and NA) independently extracted the data from the included studies. Any disagreement was resolved by consensus.

Risk of bias assessment and quality of evidence

The methodological quality of the observational studies was evaluated using the Newcastle Ottawa scale (NOS). ⁸ Observational studies with total scores of ≥ 6 on NOS were considered to have a low risk of bias, Supplementary Table 2. Funnel plot analysis was used to evaluate publication bias. The funnel plots for the recanalization grade and good functional outcome appeared asymmetric by visual inspection (Supplementary Figures. 1 and 2). However, Egger's regression analysis did not show evidence of publication bias (p = 0.12, p = 0.46 respectively).

Effect measures

The primary outcome was recanalization grade (mTICI 2b-3). The secondary outcomes were functional independence (mRS 0-2), sICH defined according to study criteria, and cardioembolisim as stroke mechanism.

Statistical analysis

Statistical analysis was performed using IBM SPSS Software Statistics (Version 29.0). Absolute counts are provided in addition to effect estimates, which are expressed as odd ratios (OR) with corresponding 95% confidence intervals (CIs). The heterogeneity of the results was evaluated using the I2 statistic as defined by the Cochrane handbook for systematic reviews. It was considered that an I2 < 50% and p > 0.1 indicated that the combined results were homogeneous.

Sensitivity analysis

To confirm the robustness of our results, sensitivity analysis for the primary outcomes using a leave-one-out meta-analysis was performed to see if it had a significant influence on the meta-analysis result (Supplementary Figures 3).

Standard protocol approvals and registrations

This systematic review and meta-analysis was registered prospectively PROSPERO. No individual level data was used, so informed consent or IRB approval was not required.

Results

Study selection

A total of 52 studies were retrieved by our search strategy. Among these, 14 were eligible for the systematic review. Twelve studies^9–21 met our inclusion criteria and were included in the meta-analysis. Figure 1 shows the PRISMA flow chart that illustrates how the final studies were selected. Basic characteristics are shown in Table 1. There was a low risk of bias for all the included studies.

Figure 1.

PRISMA flow diagram for the selection of studies. PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-analysis.

Table 1.

Demographics and basic characteristics of the observational studies.

Study	Region	Design	Site	Sample size	Gender		CT slice thickness (mm)	Site of occlusion	Reperfusion therapy	Thrombectomy technique (%)
					Male	Female
Mowla 2021 9	USA	Retrospective	Single center	93	47 (50.5%)	46 (49.5%)	-	Proximal M1 93 (100%)	MT + IVT 93 (100%)	-
Ramachandran 2022 10	India	Prospective	Single center	102	60 (57.9)	42 (42.1%)	5	ICA 18 (17.6) Proximal MCA 26 (25.5%) Mid MCA 28 (27.5%) Distal MCA 28 (27.5%) MCA M2 20 (1.9%)	MT only 49 (65.7%) MT + IVT 35 (34.3%)	SR (75) AT & SR (25)
Kovilapu 2021 11	India	Retrospective	Single center	52	NA	NA	-	ICA terminus 6 (11.5%) MCA 46 (88.5%)	MT only 27 (51.9%) MT + IVT 25 (48.1%)	SR (64) AT & SR (36)
Kang 2022 12	China	Retrospective	Multi-center	318	173 (54.4%)	145 (45.9%)	1.25, 2.5	Proximal MCA 249 (78.3%) Distal MCA 69 (21.7%)	MT only 169 (53.1%) MT + IVT 148 (46.9%)	SR (44) AT (27) AT & SR (29)
Baek 2016 13	South Korea	Retrospective	Single center	224	132 (58.9%)	92 (41.1%)	3	Anterior circulation 224 Posterior circulation	MT only 133 (59.4%) MT + IVT 91 (40.6%)	SR (100)
Chen 2023 14	China	Retrospective	Multi-center	143	95 (66.4%)	48 (33.6%)	-	Tandem ICA + MCA 41 (28.7%) MCA M1 102 (71.3%)	MT only 133 (58.1%) MT + IVT 60 (41.9%)	SR (80) Others (20)
Aykac 2022 15	Turkey	Retrospective	Single center	136	67 (49.3%)	69 (50.7%)	2.5	ICA 53 (38.9%) MCA 83 (61.1%)	MT only 133 (52.2%) MT + IVT 65 (47.8%)	AT (12) SR (55) AT & SR (33)
Man 2014 16	USA	Prospective	Single center	126	65 (51.6%)	61 (48.4%)	4.8	ICA MCA Proximal	MT only 94 (74.6%) MT + IVT 23 (25.4%)	AT (70) SR (4) Others (26)
Kim 2017 17	Germany	Retrospective	Single-center	212	107 (50.5%)	105 (49.5%)	4.8	MCA M1 72 (61.1%) MCA M2 46 (38.9%) *	MT only 78 (32.5%) MT + IVT 143 (67.5%)	SR (89) AT & SR (11)
Froehler 2013 18	USA	Retrospective	Single-center	67	26	41	4.8	ICA 20 (29.9%) MCA 43 (64.2%) Basilar 4 (5.9%)	MT only 44 (65.6%) MT + IVT 23 (34.4%)	Merci Device (100)
Merlino 2022 19	Italy	Retrospective	Single center	191	93	98	5	MCA M1 131 (68.6%) MCA M2 40 (20.9%) Tandem 20 (10.5%)	MT only 79 (41.4%) MT + IVT 112 (58.6%)	AT (48) SR (3) AT & SR (39) Others (10)
Ume 2020 20	USA	Retrospective	Single center	102	39	63	3	ICA terminus 23 (23%) Proximal M1 52 (51%) Distal M1 27 (26%)	MT only 46 (45%) MT + IVT 56 (55%)	AT, SR, Both

ICA: Internal Cerebral Artery, MCA: Middle Cerebral Artery, AT: Aspiration Thrombectomy, SR: Stent Retrieval

Primary outcome: successful recanalization (mTICI 2b-3)

In this study, a random effect model was employed to combine the findings from 12 observation studies, as depicted in Figure 2. The primary outcome of interest was successful recanalization. Among the participants, 816 out of 1012 in the HDVS positive group achieved successful recanalization, while 556 out of 733 in the HDVS negative group achieved the same outcome. These results indicate that patients in the HDVS positive group exhibited a statistically significant higher likelihood of achieving successful recanalization when compared to those with negative HDVS group (OR = 1.47, 95% CI = 1.03–2.10, p = 0.04) (Figure 2). It is worth noting that there was moderate heterogeneity among the study results, with an I² statistic of 47%.

Figure 2.

Forest plot showing the association of the hyperdense sign with the likehood of successful recanalization in AIS patients receiving mechanical thrombectomy. AIS: acute ischemic stroke.

Secondary outcomes

Cardioembolism as stroke etiology

This result was reported in six observational studies. 355 out of 616 patients HDVS positive group did have cardioembolism as stroke etiology, while 205 out of 452 in the HDVS negative group had the same outcome. Random effect model was used. HDVS positive group had higher numerical likelihood of having cardioembolism as underlying (OR: 1.43, 95% CI: 0.98–2.07; p = 0.06) (Figure 3a). There was moderate heterogeneity detected (I²=48%)

Figure 3.

(a) Forest plot showing the association of the hyperdense sign with the likehood of cardioembolisim as stoke etiology in AIS patients receiving mechanical thrombectomy. (b) Forest plot showing the association of the hyperdense sign with the likehood of achieving 90-day good functional outcome in AIS patients receiving mechanical thrombectomy. (c) Forest plot showing the association of the hyperdense sign with the risk of sICH in AIS patients receiving mechanical thrombectomy. AIS: acute ischemic stroke.

Symptomatic intracranial hemorrhage

Nine observational studies were included. Analysis using random effect model also showed that in the HDVS positive group, 107 out of 771 patients experienced sICH, while in the HDVS negative group, 44 out of 580 patients encountered this complication. There is no statically significant difference in the risk of sICH between these two groups (OR: 1.80, 95% CI 0.72–4.47, p = 0.21) (Figure 3b). The heterogeneity was observed to be high (I²=73%).

Good functional outcome (mrs 0–2)

Seven studies were combined and analyzed using the random effect model. Good functional outcome did not differ between the HDVS positive group (268/646) and the HDVS negative group (222/539), with a pooled effect estimate of 1.04 (95% CI: 0.72-1.49, p = 0.85) (Figure 3c), showing a moderate heterogeneity (I² = 49%).

Discussion

In this meta-analysis, our findings suggest that among patients undergoing MT for LVO, the presence of HDVS on non-contrast CT scan upon admission is linked to a higher likelihood of achieving successful recanalization. HDVS was not found to be associated with 90-day good functional outcome or sICH aligning with previous studies.²¹ In contrast to a previous meta-analysis that predominantly focused on intravenous thrombolytics,²² our study exclusively included patients undergoing MT, either alone or in conjunction with intravenous thrombolytics. Notably, this meta-analysis incorporated only four studies involving MT. A key difference between the two studies lies in the association between HDVS and functional outcomes. The earlier meta-analysis reported 1.43-fold higher odds of poor functional outcome across all treatment modalities in the presence of HDVS, a finding incongruent with our results. This discrepancy prompts a critical examination of the diverse patient populations and treatment strategies considered in both studies.

In our meta-analysis, we observed a trend toward higher rate of cardio embolism as a stroke etiology in patients with HDVS. This is significant as cardioembolic strokes have different microstructural features including a higher proportion of red blood cells (RBCs than those from large artery atherosclerosis as suggested by previous studies.²³ This aligns with Liebeskind et al. who conducted a histologic evaluation of clots retrieved from patients with AIS and identified a significantly greater RBC content in clots exhibiting the HDVS (47%) compared to iso-dense clots (22%) (p = 0.016). ²⁴ Importantly, our synthesis of previous studies resonates with the notion that clots enriched with RBCs are more amenable to successful recanalization, which was previously noted in the literature.²⁵

Another important imaging biomarker includes the location of the HDVS.²⁶ Previous studies indicate that thrombolytic therapy demonstrated reduced effectiveness in recanalizing occlusions within the internal carotid artery (ICA) and proximal middle cerebral artery (MCA). ²⁷ Froehler et al. added depth to this discourse by comparing the locations of HDVS in the ICA and MCA, albeit without stratification into proximal MCA, distal MCA, or M2 segments. ¹⁸ Their intriguing conclusion suggested that recanalization was not significantly associated with clot location, sparking a need for further investigation. Among the studies we included, Man et al. delved into a more refined analysis to address this gap, stratifying HDVS location into proximal MCA, distal MCA, and full-length MCA. Their findings revealed a noteworthy distinction, as the distal M1 hyperdense group exhibited a higher recanalization rate compared to the hyperdense proximal MCA or full-length group, ¹⁶ offering valuable insights for a more tailored and effective approach to stroke intervention.

Kovilapu et al. explored the correlation between HDVS and first pass recanalization. Their study revealed that the presence of HDVS independently predicted the success of both first pass and early recanalization (OR: 3.8, 95% CI 1.11–13.03, p = 0.03). ¹¹ This observation carries significant clinical implications, as various studies underscore the critical role of the first pass effect as an independent prognostic factor for attaining improved functional outcomes in LVO.²⁸ The emphasis on achieving first pass success stems from the recognition that multiple passes can contribute to heightened clot friction and compaction. Additionally, it can lead to thrombus fragmentation and distal embolization and re-occlusion which may adversely affect the functional outcome.²⁹

Our study indicates that patients with HDVS showed a numerically higher incidence of sICH following MT, even though the statistical significance was not achieved, aligning with the findings of Kang et al. ¹² who investigated specifically the relation between HDVS and hemorrhagic transformation (HT) following MT. This observed trend may be attributed to the previously established correlation between the presence of HDVS on initial NCCT scans and a larger cerebral infarction volume as estimated by Alberta Stroke Program Early CT Score (ASPECTS) and subsequently the risk of HT after MT.³⁰ The link between HDVS and increased susceptibility to sICH underscores the importance of considering pre-existing indicators when evaluating potential outcomes and complications in patients undergoing endovascular interventions.

Our study has limitations: first, the inclusion of a relatively small number of studies, and second, the reliance on observational studies introduces inherent biases and confounding. Third, the diverse array of endovascular strategies employed across the included studies poses a challenge in establishing consistent comparisons. Fourth, the lack of other crucial biomarkers such as length, volume, and permeability HDVS, represents a notable gap in the analysis. Finally, this meta-analysis relied on the aggregation of published data rather than individual data, which unfortunately precluded us from performing subgroup analysis and a further detailed examination of the diverse locations of HDVS implementation.

Conclusion

In conclusion, patients with HDVS who underwent MT may achieve a higher rate of successful recanalization. However, larger confirmatory prospective studies including other pre-interventional imaging biomarkers are warranted.