Fig 1.
The flow diagram and the quality assessment.
A) The flow diagram. This diagram shows the process of study selection. Finally, 13 eligible studies were included in our meta-analysis. B) The quality assessment of the included studies by QUADAS-2. This assessment summarizes “risk of bias” and “applicability concerns” through judging each domain in each included study and shows the major biases concentrated upon the “index text” and lesser biases that were related to the “flow and timing”.
Table 1.
Main characteristics of the studies included in this meta-analysis.
Table 2.
The details of NETs patients included in the meta-analysis.
Fig 2.
The forest plots show the pooled diagnostic indices of CgA for NETs.
The heterogeneity caused by the non-threshold effect is quantified by inconsistency (I2). Because this heterogeneity exists, a random effects model was used to pool these data. The point efficiencies from each study are shown as squares, and the pooled efficiencies are shown as diamonds. The degree of freedom is abbreviated as df. As shown, A) the pooled sensitivity and specificity are 0.73 (95% CI: 0.71 to 0.76) and 0.95 (95% CI: 0.93 to 0.96), respectively. B) the pooled PLR and NLR are 14.56 (95% CI: 6.62 to 32.02) and 0.26 (95% CI: 0.18 to 0.38), respectively. C) the pooled DOR is 56.29 (95% CI: 25.27 to 125.38).
Fig 3.
The summary receiver operating characteristic curves (SROC).
Every square represents a study. The SROC curve is symmetric with the 0.8962 AUC, which intimates a higher diagnostic accuracy for the diagnosis of NETs.
Fig 4.
The funnel plots of publication bias.
Every point represents one study, and the line is the regression line. The funnel shape is symmetrical, which indicates no publication bias.