Systematic Review—Role of MRI in Cervical Cancer Staging

Abstract

A systematic review of the diagnostic accuracy of MRI in the staging of cervical cancer was conducted based on the literature from the last 5 years. A literature search was performed in the Cochrane Library, EMBASE, MEDLINE and PubMed databases using the MeSH terms “cervical cancer”, “MRI” and “neoplasm staging”. A total of 110 studies were identified, of which 8 fit the inclusion criteria. MRI showed adequate accuracy (74–95%) and high sensitivity (92–100%) in assessing stromal invasion. The data for MRI in terms of assessing vaginal and pelvic side wall involvement were wide ranging and inconclusive. In assessing lymph node metastasis, MRI showed an adequate accuracy (73–90%), specificity (75–91%) and NPV (71–96%) but poor sensitivity (52–75%) and PPV (52–75%). MRI showed high accuracy (95%), sensitivity (78–96%), specificity (87–94%), and NPV (98–100%) but poor PPV (27–42%) in detecting bladder involvement. There was a paucity of data on the use of MRI in assessing rectal involvement in cervical cancer. Overall, the literature was heterogenous in the definitions and language used, which reduced the comparability between articles. More research is required into the diagnostic accuracy of MRI in the staging of cervical cancer and there must be increased consistency in the definitions and language used in the literature.