Objective:To explore the diagnostic value of four types of humoral microRNAs（plasma

saliva

serum

and whole blood）for oral squamous cell carcinoma（OSCC）. Methods:Databases were searched for diagnostic trials in which OSCC was diagnosed using humoral microRNAs; these trials were collected based on strict inclusion and exclusion criteria. The QUADAS-2 tool was used to assess the risk of bias in all the collected literature.The included literature was divided into four groups according to the types of humoral samples. Meta-Disc1.4 was used to determine the presence of threshold effect and non-threshold effect based on Spearman correlation coefficient and Cochran-Q test. Pooled statistics in the four groups were analyzed with Reveman5.3. Stata13.0 was used to draw a Deek’s funnel plot to determine the presence of publication bias. Results:Eleven pieces of literature（including 19 humoral microRNAs）which had a relatively low risk of bias were included. The pooled sensitivity

pooled specificity

and area under the summary receiver operating characteristic curve were 0.87（95% confidence interval [CI]:0.83 ～ 0.90）

0.85（95% CI:0.81 ～ 0.89）

and 0.93

respectively

for the plasma group

0.7（95%CI:10.72 ～ 0.85）

0.78（95% CI:0.70 ～ 0.85）

and 0.87

respectively

for the serum group

0.77（95% CI:0.71 ～ 0.83）

0.78（95% CI:0.70 ～ 0.85）

and 0.84

respectively

for the saliva group

and 0.73（95% CI:0.66～ 0.80）

0.81（95% CI:0.72 ～ 0.88）

0.82

respectively

for the whole blood group. Conclusion:Among the four types of humoral microRNA samples

the plasma group ranks first in pooled diagnostic parameters such as pooled sensitivity and pooled specificity

showing higher overall diagnostic performance for OSCC. The diagnostic performance of the serum group and saliva group are similar

ranking second and third

respectively; the whole blood group has the poorest diagnostic performance for OSCC.