Magnetic resonance imaging of rectal tumours

Abstract

Cross sectional imaging techniques introduced during the last two decades have been increasingly used in the evaluation of patients with pelvic tumours. The extent of tumour - the tumour stage - at the time of diagnosis constitutes a guideline for both the immediate treatment and the follow-up of the patient. Continuous development of imaging modalities, such as magnetic resonance imaging (MRI), has implications which motivates a change of the examination routines.

State-of-the-art high resolution MRI was evaluated in patients with rectal tumours considered as resectable by the surgeon. Tumours considered to be primarily not resectable were also evaluated with both computed tomography (CT) and MRI. After surgery, resected rectal specimens were examined using similar MR-techniques. In patients who after surgery presented with a clinical suspicion of local recurrence, dynamic gadolinium contrast-enhanced MRI was evaluated in order to differentiate local tumour recurrence from changes in the pelvis related to the treatment. MRl was also compared to CT and monoclonal antibody (CEA)-scintigraphy for the diagnosis of local recurrence.

The results demonstrated that tumour penetration through the rectal wall to the perirectal tissues and the presence of Iymph local Iymph node metastases could be predicted in 75 % of patients with resectable rectal tumours. In patients with unresectable rectal cancer, MRI better predicted involvement of the uterus and the urinary bladder than CT. However, sensitivity of CT and specificity of MRl were both low in terms of diagnosis of organ involvement. After surgery, no parameters were found which helped to distinguish local recurrent tumours from benign changes in contrast-enhanced dynamic MRI. When CT, MRI and CEA-scintigraphy were compared for the diagnosis of locally recurrent rectal tumours, the diagnosis was most effectively established by MRl.

It is concluded that local excision of rectal tumours can presently not be performed based on results of MRI. Sensitivity of CT and specificity of MRI is not sufficient to allow for general recommendations in the evaluation unresectable rectal tumours. However, If the cross-sectional investigation begins with CT, MRl will contribute to a more complete evaluation, especially if involvement of the bladder and the internal genitalia cannot be completely ruled out. After surgery, a tumour-free lateral resection margin can be verfied by MRI of rectal specimen if the measured distance is more than one mm. Local tumour recurrence and benign changes in the pelvis related to the treatment of the patient can presently not be differentiated on account of dynamic contrast-enhanced MRI. In the diagnosis of locally recurrent rectal cancer with cross-sectional evaluation, MRl should be considered as the first choice rather than CT or CEA-scintigraphy.