Clinical examination versus magnetic resonance imaging in the pretreatment staging of cervical carcinoma: systematic review and meta-analysis

Abstract

To review the literature on the diagnostic performance of clinical examination and magnetic resonance imaging (MRI) in detecting parametrial invasion and advanced stage disease (FIGO stage ≥ IIB) in patients with cervical carcinoma. Reports of studies were searched using the MEDLINE, EMBASE and Cochrane databases. Two observers reported on data relevant for analysis and methodological quality using the QUADAS scoring system. Publication bias was analysed using Deeks funnel plots. Covariates were added to the model to study the influence on the summary results of the technical and methodological aspects of the clinical examination and MRI. In total, 3,254 patients were included. Partial verification bias was often encountered. Pooled sensitivity was 40 % (95 % CI 25-58) for the evaluation of parametrial invasion with clinical examination and 84 % (95 % CI 76-90) with MRI, 53 % (95 % CI 41-66) for the evaluation of advanced disease with clinical examination, and 79 % (95 % CI 64-89) with MRI. Pooled specificities were comparable between clinical examination and MRI. Different technical aspects of MRI influenced the summary results. MRI is significantly better than clinical examination in ruling out parametrial invasion and advanced disease in patients with cervical carcinoma. • MRI has a higher sensitivity than clinical examination for staging cervical carcinoma. • Clinical examination and MRI have comparably high specificity for staging cervical carcinoma. • Quality of clinical examination studies was lower than that of MRI studies. • The use of newer MRI techniques positively influences the summary results. • Anaesthesia during clinical examination positively influences the summary results.