Can preoperative MRI accurately evaluate nodal and parametrial invasion in early stage cervical cancer?

Objective: To evaluate the diagnostic performance of magnetic resonance imaging (MRI) in the pretreatment evaluation of invasive cervical cancer especially for the parametrial invasion and lymph node (LN) involvement. Methods: We retrospectively recruited consecutive patients with biopsy-confirmed cervical cancer who had undergone preoperative MRI and were scheduled for surgery based on clinical assessment between January 2004 and May 2006. We evaluated the diagnostic performance of MRI for the parametrial invasion and LN involvement using surgicopathologic findings as the reference standard. Results: A total of 119 eligible patients completed preoperative and intra-operative survey, of whom 34 (28.6%) had pelvic LN metastasis and four (3.4%) had para-aortic LN metastasis histologically. The sensitivity, specificity and accuracy of MRI in detecting LN involvement by region-specific analysis were 40.5, 91.3 and 86.8% respectively. The sensitivity, specificity and accuracy of MRI in detecting parametrial invasion were 44.4, 89.1 and 88.3% respectively. The positive predictive value (PPV) of preoperative MRI for detecting region-specific LN involvement and parametrial invasion was 31.3 and 61.2%, respectively. Imaging findings of suspected parametrial invasion were not to influence the treatment decision in the study. Conclusion: Preoperative MRI showed low PPV for detecting LN involvement and parametrial invasion in cervical cancer. Further studies are necessary to determine the cost-effectiveness of using MRI in place of conventional clinical staging tests according to clinical indication and also its use in comparison with that of integrated positron emission tomography/computed tomography.