Objective: The goal of this study was to determine the accuracy of thin-section magnetic resonance imaging (MRI) using a pelvic phased-array coil for preoperative local staging of rectal carcinoma.
Methods: Ninety-three patients with primary rectal cancer were prospectively evaluated by thin-section MRI using a pelvic phased-array coil for T and N staging. After radical surgery, MRI results were compared with the histological findings of the resected specimens, and accuracy, sensitivity, specificity, and positive and negative predictive values were assessed.
Results: According to pathological staging, 93 tumors were classified based on the TNM classification as follows:
Four tumors were pT1, 26 tumors were pT2, 51 tumors were pT3, and 12 tumors were pT4. The MRI had an overall accuracy of 82.8% (77 of 93) for discriminating between early (ie, ≤pT2) and advanced (ie, pT3-pT4) tumors, with a sensitivity of 67% (20 of 30) for early tumors and a sensitivity of 90% (57 of 63) for advanced tumors. It yielded a diagnostic accuracy of 75% (70 of 93) in predicting T stage.
Regarding lymph node metastases, the MRI prediction was correct in 58 of 93 patients, which yielded an overall accuracy of 62.4%.
Conclusions: Despite some limitations, thin-section MRI with pelvic phased-array coil can be reliably used in the preoperative evaluation of rectal cancer patients for local tumor staging. It seems highly effective in predicting T stages and can provide useful information regarding lymph node metastases of rectal cancer.