To evaluate and compare the value of gray scale ultrasound based ultrasound microangiography (SMI) and color Doppler flow imaging (CDFI) in the differential diagnosis of thyroid nodule.

CNKI, VIP, Sinomed, Wanfang, Cochrane Library, PubMed, Medline, Embase, Web of Science, and Scopus databases were searched to screen all the literature published before June 1, 2020. According to the inclusion and exclusion criteria, the qualified literature was screened, the diagnostic accuracy research quality assessment tool (QUADAS-2) was used to evaluate the quality of included literature, and the Review Manager 5.3 software was used to make the quality evaluation chart. Stata V15.0 software was used to carry out heterogeneity test and pool effect sizes such as sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio and summary receiver operating characteristic curve (SROC curve) for comprehensive evaluation. Meta-regression analysis was used to explore the source of heterogeneity and conduct subgroup analysis. Funnel diagram and Fagan diagram were plotted for publication bias evaluation and clinical applicability evaluation.

Finally, 12 articles were included, with a total of 1359 lesions. The combined sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, odds ratio, and area under the SROC curve of SMI were 0.83 (95%CI: 0.79-0.85), 0.86 (95%CI: 0.78-0.91), 5.7 (95%CI: 3.8-8.7), 0.20 (95%CI: 0.17-0.24), 28 (95%CI: 17-46), and 0.85 (95%CI: 0.82-0.88), respectively; the corresponding values of CDFI were 0.65 (95%CI: 0.58-0.70), 0.78 (95%CI: 0.72-0.83), 2.9 (95%CI: 2.4-3.6), 0.46 (95%CI: 0.39-0.53), 6 (95%CI: 5-9), and 0.77 (95%CI: 0.73-0.81). Meta-regression analysis showed that the source of CDFI heterogeneity was nodular type (P＜0.05). Subgroup analysis showed that there was no significant heterogeneity when only unclassified nodules were included in the study (P=0.12); the combined diagnostic sensitivity and specificity were 0.64 (0.57-0.71) and 0.81 (0.76-0.85), respectively. There was also no significant heterogeneity when only the literature that studied TI-RADS4 type nodules was included (P=0.62); the combined diagnostic sensitivity and specificity were 0.68 (0.54-0.81) and 0.64 (0.52-0.76), respectively. Deek funnel pattern showed no significant publication bias. Fagan chart results showed that the post-test probability of SMI and CDFI was 60% and 40% , respectively, which was higher than the pre-test probability of 20%.

The diagnostic efficiency of SMI for benign and malignant thyroid nodules is better than CDFI, and SMI technology can reflect the distribution of blood flow, make up for the deficiency of CDFI, and has better clinical application value.