|
Objective To investigate the serum CEA and CYFRA21-1 non-small cell lung cancer diagnosis . Methods We searched PubMed, Embase, Cochrane library, Ovid, CBM and other documents in the database , in accordance with pre-established inclusion and exclusion criteria screening literature , extracted characteristic information included in the study . Statistical analysis using the Meta-DiSc 1.4, SPSS 17.0 software . QUADAS scale of included studies using rigorous quality evaluation, then heterogeneity , according to the characteristics of heterogeneity select the appropriate model to calculate the combined effect of the sensitivity , specificity, positive likelihood ratio and negative likelihood ratio , drawing integrated receiver operating curve and calculate the area under the curve (AUC), Finally, according to the possible sources of heterogeneity in subgroup analyzes . Results A total of 1679 to the relevant literature , and ultimately into the 38 documents . After the gold standard pathological diagnosis of NSCLC patients 5140 cases , 3834 cases of the control group . Meta-analysis showed , CEA, CYFRA21-1 and CEA CYFRA21-1 in each group were included in the study between the presence of heterogeneity , SROC area under the curve were 0.7230,0.8804,0.9134 . Conclusion CEA, CYFRA21-1 non-small cell lung cancer diagnosis has some value , contribute to disease diagnosis and disease monitoring, combined CEA and CYFRA21-1 can increase its diagnostic sensitivity , but CEA auxiliary CYFRA21-1 combined detection of CYFRA21-1 independently tested to improve the overall diagnostic performance no obvious significance.
|