Answers are grounded in this paper and cite their sources.
Objective This study aimed to synthesize the prevalence of baseline cerebral microbleeds (CMBs) and the incidence of hemorrhagic transformation (HT) and symptomatic intracranial hemorrhage reported in magnetic resonance imaging (MRI)-based studies of acute ischemic stroke. Methods From the earliest records up to December 2025, the electronic databases (including PubMed, Embase, Cochrane Library, and Web of Science) were searched to gather diagnostic reports or cohort investigations addressing the utilization of MRI in forecasting HT after acute ischemic stroke (AIS). Two reviewers independently carried out the processes of literature selection, data extraction, and methodological quality evaluation of the enrolled articles. A single-arm meta-analytical approach was undertaken using Stata and RevMan software. Results This study included 14 studies on hemorrhage-related outcomes in AIS, with a total of 5736 patients. The meta-analysis results showed that the pooled prevalence of baseline cerebral microbleeds was 24.2% (95% CI: 18.3%–30.0%), whereas the pooled incidence proportions of hemorrhagic transformation and symptomatic intracranial hemorrhage were 22.1% (95% CI:9.80%–34.4%) and 3% (95% CI, 1%–6%), respectively. All pooled estimates showed very high between-study heterogeneity (I 2 >90%) and therefore represent broad summaries of heterogeneous cohorts rather than fixed population benchmarks. Publication bias assessment for the CMBs outcome revealed an asymmetric funnel plot, and both Egger's test ( p = 0.286) and Begg's test ( p = 0.211) indicated no significant small-study effects or publication bias. Conclusion Through this meta-analysis, based on the pooled prevalence estimates of MRI-derived features, the proportions of patients with AIS presenting with CMBs and HT were relatively high, whereas symptomatic intracranial hemorrhage was comparatively rare. These descriptive estimates demonstrate substantial between-study variability and should not be interpreted as validated individual bleeding-risk probabilities; prospective studies with standardized MRI and clinical variables are needed before multimodal prediction models can be reliably developed.
No takes yet. Share an insight, caveat, or question.
Liangjuan Yang (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: