Objectives: To determine whether magnetic resonance imaging (MRI)-detected extramural vascular invasion (EMVI) could predict synchronous distant metastases in rectal cancer. Methods: Patients who underwent rectal MRI between July 2011 and December 2012 were screened. This study included 447 patients with pathologically confirmed rectal adenocarcinoma who had undergone MRI without previous treatment. Distant metastases were recorded at the initial work-up and over a 6-month follow-up. Univariate/multivariate logistic regression models were used to determine the risk of metastasis. The diagnostic performance was calculated using pathologic lymphovascular invasion (LVI) as a gold standard. Results: Among 447 patients, 79 patients (17.7 %) were confirmed to have distant metastases. Three MRI features are significantly associated with a high risk of distant metastasis: positive EMVI (odds ratio 3.02), high T stage (odds ratio 2.10) and positive regional lymph node metastasis (odds ratio 6.01). EMVI in a large vessel (≥3 mm) had a higher risk for metastasis than EMVI in a small vessel (<3 mm). Sensitivity, specificity and accuracy of MRI-detected EMVI were 28.2 %, 94.0 % and 80.3 %, respectively. Conclusions: MRI-detected EMVI is an independent risk factor for synchronous metastasis in rectal cancer. EMVI in large vessels is a stronger risk factor for distant metastasis than EMVI in small vessels. Key points: • EMVI, LN metastasis and T staging on MRI are risk factors for metastasis. • EMVI in large vessels has greater risk for metastasis than in small vessels. • Regional LN metastasis on MRI has highest risk for predicting metastasis. • MR findings could be helpful for selecting patients at high risk for metastasis.
Bibliographical noteFunding Information:
The scientific guarantor of this publication is Joon-seok Lim. The authors of this manuscript declare no relationships with any companies whose products or services may be related to the subject matter of the article. This study was supported by faculty research grants of Yonsei University College of Medicine (6-2012-0093 and 6-2011-0139). One of the authors has significant statistical expertise (Sungmin Myoung). Institutional review board approval was obtained. Written informed consent was waived by the institutional review board. Methodology: retrospective, cross-sectional study, performed at one institution.
© 2014, European Society of Radiology.
All Science Journal Classification (ASJC) codes
- Radiology Nuclear Medicine and imaging