Abstract
OBJECTIVE: To evaluate the diagnostic value of magnetic resonance (MR) imaging in the diagnosis of mandibular invasion caused by oral cancers. METHODS: Medline, EMBASE, SIGLE, and Chinese biomedical literature database were searched electronically. Manual searching for 19 relevant Chinese journals was also performed. Two reviewers evaluated the literature and extracted the data. Meta-Disc 1.4 was chosen to conduct the sensitivity (SEN), specificity (SPE), and 95% confidence interval (95%CI). RESULTS: Twelve studies with a total of 476 patients, namely, 5 prospective studies and 7 retrospective studies, were included. All the studies had unclear risk of bias. Meta-analysis result showed that the combination of SEN of MR in diagnosing mandibular invasion was 0.779 (95%CI: 0.719-0.831), combination of SPE was 0.823 (95%CI: 0.767-0.870), positive likelihood ratio was 3.442 (95%CI: 2.181-5.431), negative likelihood ratio was 0.286 (95%CI: 0.181- 0.451), and diagnostic odds ratio was 25.702 (95%CI: 13.406-49.273). The area under curve was 0.903 9 and Q* was 0.835 4. Meta-analysis was not processed when diagnosing mandibular medullary invasion because only two studies with 55 patients had been reported. The SEN was 0.838, and the SPE was 0.722. CONCLUSION: MR is efficient and has clinical value in the diagnosis of mandibular invasion caused by oral cancer.