Pulmonary vascular enlargement on thoracic CT for diagnosis and differential diagnosis of COVID-19: a systematic review and meta-analysis.

ABSTRACT

BACKGROUND: The 2019 coronavirus disease (COVID-19) has become a global pandemic. To date, although many studies have reported on the computed tomography (CT) manifestations of COVID-19, the vascular enlargement sign (VES) of COVID-19 has not been deeply examined, with the few available studies reporting an inconsistent prevalence. We thus performed a systematic review and meta-analysis based on the best available studies to estimate the prevalence and identify the underlying differential diagnostic value of VES. METHODS: We searched nine English and Chinese language databases up to April 23, 2020. Studies that evaluated CT features of COVID-19 patients and reported VES, with or without comparison with other pneumonia were included. The methodologic quality was assessed using Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2). Meta-analyses with random effects models were performed to calculate the aggregate prevalence and pooled odds ratios (ORs) of VES. We also conducted meta-regression and subgroup analyses to analyze heterogeneity. RESULTS: VES findings from a total of 1969 patients were summarized and pooled across 22 studies. Our analysis demonstrated that the prevalence of VES among COVID-19 patients was 69.37% [95% confidence interval (CI) 57.40-79.20%]. Compared with non-COVID-19 patients, VES manifestation was more frequently observed in confirmed COVID-19 patients (OR =6.43, 95% CI 3.39-12.22). Studies that explicitly defined distribution of VES in the lesion area demonstrated a significantly higher prevalence (P=0.03). Subgroup analyses also revealed a relatively higher VES rate in studies with a sample size larger than 50, but the difference was not statistically significant. No significant difference in VES rates was found between different countries (China/Italy), regions (Hubei/outside Hubei), average age groups (over/less than 50-year-old), or slice thicknesses of CT scan. Extensive heterogeneity was identified across most estimates (I2>80%). Some of the variations (R2=19.73%) could be explained by VES distribution, and sample size. No significant publication bias was seen (P=0.29). CONCLUSIONS: VES on thoracic CT was found in almost two-thirds of COVID-19 patients, and was more prevalent compared with that of the non-COVID-19 patients, supporting a promising role for VES in identifying pneumonia caused by coronavirus.