Magnetic resonance imaging for detection of parametrial invasion in cervical cancer: An updated systematic review and meta-analysis of the literature between 2012 and 2016.

Abstract

To review the diagnostic performanceof MRI for detection of parametrial invasion (PMI) in cervical cancer patients. MEDLINE and EMBASE databases were searched for studies providing diagnostic performance of MRI for detecting PMI in patients with cervical cancer. Studies published between 2012 and 2016 using surgico-pathological results as reference standard were included. Study quality was evaluated using QUADAS-2. Sensitivity and specificity of all studies were calculated. Results were pooled and plotted in a hierarchical summary receiver operating characteristic plot. Meta-regression and subgroup analyses were performed. Fourteen studies (1,028 patients) were included. Study quality was generally moderate. Pooled sensitivity was 0.76 (95% CI 0.67-0.84) and specificity was0.94 (95% CI 0.91-0.95). The possibility of heterogeneity was considered low: Cochran's Q-test (p = 0.471), Tau2 (0.240), Higgins I2 (0%). With meta-regression analysis, magnet strength, use of DWI, and antispasmodic drugs were significant factors affecting heterogeneity (p < 0.01). Subgroup analysis for studies solely using radical hysterectomy as reference standard yielded pooled sensitivity and specificity of 0.73 (95% CI 0.60-0.83) and 0.93 (95% CI 0.90-0.95), respectively. MRI shows good performance for detection of PMI in cervical cancer. Using 3-T scanners and DWI may improve diagnostic performance. • MRI shows good performance for detection of parametrial invasion in cervical cancer. • Subgroup of studies using only radical hysterectomy showed consistent results. • Using 3-Tesla scanners and diffusion-weighted imaging may improve diagnostic performance.