[Nuclear matrix protein 22 and urinary cytology test in the diagnosis of bladder cancer: a meta-analysis].

Abstract

Systematic reviews of diagnostic value of the nuclear matrix protein 22 (NMP22) and urine cytology for bladder cancer. Development of inclusion criteria, exclusion criteria and search strategy to retrieve relevant literature. Screening the literature according to inclusion criteria and exclusion criteria. Quality evaluation of the screening and data extraction, using MetaDiSc 1.4 software for Meta analysis. In total, 266 relevant studies were searched, excluded 256 studies, and then 10 studies were included, with 4895 patients involved. The pooled sensitivity and specificity of NMP22 to detect bladder cancer were 0.76 (95%CI: 0.74 - 0.77), 0.80 (95%CI: 0.79 - 0.82), respectively. The pooled sensitivity and specificity of urine cytology were 0.36 (95%CI: 0.34 - 0.38), 0.94 (95%CI: 0.93 - 0.95), respectively. The area under curve (AUC) for NMP22 and urine cytology were 0.8533 and 0.8628, and Q(*) index were 0.7863 and 0.7934, respectively. For the diagnosis of bladder cancer, the sensitivity of NMP22 was higher than urine cytology, but the specificity was lower than urine cytology. Overall diagnostic performance of NMP22 was medium, it was no significant difference with urine cytology. It can't replace urine cytology now.