Extramural depth of tumor invasion at thin-section MR in patients with rectal cancer: results of the MERCURY study

• Published in: Radiology Vol. 243; no. 1; p. 132
• Format: Journal Article
• Language: English
• Published: United States 01.04.2007
• Subjects: Adenocarcinoma - pathology; Adenocarcinoma - surgery; Adenocarcinoma - therapy; Adult; Aged; Aged, 80 and over; Combined Modality Therapy; Female; Humans; Magnetic Resonance Imaging - methods; Male; Middle Aged; Preoperative Care; Prognosis; Prospective Studies; Rectal Neoplasms - pathology; Rectal Neoplasms - surgery; Rectal Neoplasms - therapy; Reference Standards; Sensitivity and Specificity
• ISSN: 0033-8419; 1527-1315
• DOI: 10.1148/radiol.2431051825

To prospectively evaluate the accuracy of magnetic resonance (MR) imaging in depicting the extramural depth of tumor invasion in patients who have rectal cancer, with histopathologic results as the reference standard. The Magnetic Resonance Imaging and Rectal Cancer European Equivalence (MERCURY) Study received ethics approval from all participating centers, and all patients gave informed consent. Consecutive patients (n = 679) with adenocarcinoma of the rectum consented to participate. Imaging workshops for participating specialist gastrointestinal radiologists were held to ensure standardization of image acquisition techniques. Standardized MR image interpretation and data reporting were performed by using previously validated criteria. MR images were prospectively singly read by the specialist gastrointestinal radiologists. The maximal extramural depth (EMD) of tumor spread, defined at histopathologic analysis as the distance from the outer edge of the longitudinal muscularis propria to the outer edge of the tumor, was measured and recorded. The maximal EMD was the reference standard. The MR and histopathologic results were considered to be equivalent when the 95% confidence interval of the difference between them was within +/-0.5 mm. Tumor EMD measurements obtained at both MR imaging and histopathologic analysis were available for 295 (95%) of 311 patients after primary surgery. Mean EMDs were 2.80 mm +/- 4.60 (standard deviation) and 2.81 mm +/- 4.28 at MR imaging and histopathologic analysis, respectively. The mean difference between the MR-derived and histopathologically derived EMDs was -0.05 mm +/- 3.85 (95% confidence interval: -0.49 mm, 0.40 mm). Therefore, MR and histopathologic assessments of tumor spread were considered equivalent to within 0.5 mm. Demonstration of accurate measurement of the depth of extramural tumor spread in the MERCURY Study enabled accurate preoperative prognostication.