MRI in Evaluation of Rectal Cancer pre- and post-Chemo-Radiation Treatment.

Rectal cancer is associated with a high risk of metastases and local recurrence; local recurrence rates after surgical treatment being up to 32%. An accurate local staging at the time of initial diagnosis is therefore very important. Magnetic Resonance Imaging (MRI) is already established as an accurate tool for the preoperative staging of rectal cancer and has resulted in marked improvements in staging accuracy. Material and Methods: This study used MRI in comparing the morphologic features of rectal cancer before and after 8 weeks of chemo-radiation treatment (CRT) and to correlate the post treatment MRI appearances with the histological findings in resected tumors. 45 patients with histo-pathologically proven rectal adenocarcinoma received standardized 8 weeks chemo-radiation therapy and subjected to MRI before and after treatment for clinical staging. A correlation between pathological response and MRI findings was done. Results: The MRI diagnostic accuracy to diagnose T2 is 74.2% with relatively low specificity (64.7%). The diagnostic accuracy of MRI in evaluation of stage T3 and T4, the MRI sensitivity was 96.2% however of low specificity 26.3%. The diagnostic accuracy was 66.7%. Additionally, in evaluation of T2 stage, the sensitivity of MRI was very low 27.3% and specificity relatively high 94.7%. Diagnostic accuracy was 70%. Post RCT, based on downstaging after CRT, the sensitivity of MRI to show no tumor was very low 0% with diagnostic accuracy 88.9%. However, to evaluate stage T2, the sensitivity was 84.6% with low specificity 66.7% and the diagnostic accuracy was 74.2%. Conclusion: MRI had an accuracy average of 81.6% in T stage and 68.9% in N stage in re-staging rectal tumors after CRT. Over-staging results of majority of the inaccuracy. The statistical agreement between post-CRT MRI and the pathologic staging involving T and N stages was not satisfactory. In view of the above, Post CRT, restaging rectal cancer remains a challenge.