Early invasive cervical cancer: tumor delineation by magnetic resonance imaging, computed tomography, and clinical examination, verified by pathologic results, in the ACRIN 6651/GOG 183 Intergroup Study.

PURPOSE To compare magnetic resonance imaging (MRI), computed tomography (CT), and clinical examination for delineating early cervical cancer and for measuring tumor size. PATIENTS AND METHODS A 25-center study enrolled 208 patients with biopsy-proven invasive cervical cancer for MRI and CT before attempted curative radical hysterectomy. Each imaging study was interpreted prospectively by one onsite radiologist and retrospectively by four independent offsite radiologists, who were all blinded to surgical, histopathologic, and other imaging findings. Likelihood of cervical stromal and uterine body involvement was rated on a 5-point scale. Tumor size measurements were attempted in three axes. Surgical pathology was the standard of reference. RESULTS Neither MRI nor CT was accurate for evaluating cervical stroma. For uterine body involvement, the area under the receiver operating characteristic curve was higher for MRI than for CT for both prospective (0.80 v 0.66, respectively; P = .01) and retrospective (0.68 v 0.57, respectively; P = .02) readings. Retrospective readers could measure diameter by CT in 35% to 73% of patients and by MRI in 79% to 94% of patients. Prospective readers had the highest Spearman correlation coefficient with pathologic measurement for MRI (r(s) = 0.54), followed by CT (r(s) = 0.45) and clinical examination (r(s) = 0.37; P < .0001 for all). Spearman correlation of multiobserver diameter measurements for MRI (r(s) = 0.58; P < .0001) was double that for CT (r(s) = 0.27; P = .03). CONCLUSION In patients with cervical cancer, MRI is superior to CT and clinical examination for evaluating uterine body involvement and measuring tumor size, but no method was accurate for evaluating cervical stroma.