Cervical carcinoma: comparison of standard and pharmacokinetic MR imaging.

Abstract

To stage advanced cervical carcinoma with conventional or pharmacokinetic magnetic resonance (MR) imaging by correlating imaging findings with whole-mount specimens and histopathologic findings. Twenty-six adult patients with primary cervical cancer (stages IIB-IVA) underwent T2-weighted turbo spin-echo (SE) MR imaging; gadolinium-enhanced, T1-weighted SE MR imaging; and gadolinium-enhanced, saturation-recovery, turbo fast low-angle shot MR imaging. All imaging findings were correlated with the whole-mount specimens and histopathologic findings. Signal intensity changes versus time were analyzed by using a pharmacokinetic model and parameter values displayed as a color-coded overlay. Histopathologic stages were IIB (n = 9), IIIB (n = 1), and IVA (n = 16). The overall accuracy for tumor staging was 73% for T2-weighted, 81% for T1-weighted, and 92% for pharmacokinetic MR imaging. Pharmacokinetic MR imaging was accurate (90%) in the diagnosis of tumor extension into the bladder and/or rectal wall but inaccurate (69%) in that of parametrial invasion. T2-weighted images were most accurate (86%) in the assessment of parametrial tumor extension but less accurate (69%) in that of bladder or rectal invasion. T2-weighted turbo SE images are still superior to contract medium-enhanced T1-weighted SE or pharmacokinetic MR images in the diagnosis of parametrial infiltration by uterine cervical carcinoma. However, pharmacokinetic MR imaging is a promising method for demonstrating and staging IVA disease.