Early-stage cervical cancer: tumor delineation by magnetic resonance imaging and ultrasound - a European multicenter trial

Objective. To compare the diagnostic accuracy of ultrasound (US) and magnetic resonance imaging (MRI) in the preoperative assessment of early-stage cervical cancer using pathologic findings as the reference standard. Patients and methods. Prospectivemulti-center trial enrolling 209 consecutivewomen with early-stage cervical cancer (FIGO IA2–IIA) scheduled for surgery. The following parameterswere assessed on US andMRI and compared to pathology: remaining tumor, size, tumor stromal invasionb2/3 (superficial) or ≥2/3 (deep), and parametrial invasion. Results. Complete data were available for 182 patients. The agreement between US and pathologywas excellent for detecting tumors, correctly classifying bulky tumors (>4 cm), and detecting deep stromal invasion (kappa values 0.84, 0.82, and 0.81 respectively); and good for classifying small tumors (b2 cm) and detecting parametrial invasion (kappa values 0.78 and 0.75, respectively). The agreement between MRI and histology was good for classifying tumors as b2 cm, or >4 cm, and detecting deep stromal invasion (kappa values 0.71, 0.76, and 0.77, respectively). It was moderately accurate in tumor detection, and in assessing parametrial invasion (kappa values 0.52 and 0.45, respectively). The agreement between histology and US was significantly better in assessing residual tumor (pb0.001) and parametrial invasion (pb0.001) than the results obtained by MRI. Imaging methods were not significantly influenced by previous cone biopsy.