Magnetic resonance imaging of rectal tumours

Abstract

Magnetic Resonallce Imaging Or Rectal Tumours

by Lennart Blomqvist, M.D. Department of Surgical Sciences,Division for Diagnostic Radiology, Karolinska Hospital and Institute, S-171 76, Stockholm,Sweden

Cross sectional imaging techniques introduced during the last two decades havebeen increasingly used in the evaluation of patients with pelvic tumours.

The extent of tumour - the tumour stage - at the time of diagnosis constitutesa guide line for both the immediate treatment and the follow-up of the patient. Continuousdevelopment of imaging modalities, such as magnetic resonance imaging (MRI), hasimplications which motivates a change of the examination routines.

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

The results demonstrated that tumour penetration through the rectal wall to theperirectal tissues and the presence of Iymph local Iymph node metastases could bepredicted in 75 % of patients with resectable rectal tumours. In patients with unresectablerectal cancer, MRI better predicted involvement of the uterus and the urinary bladderthan CT. However, sensitivity of CT and specificity of MRl were both low in termsof diagnosis of organ involvement. After surgery, no parameters were found whichhelped to distinguish local recurrent tumours from benign changes in contrast-enhanceddynamic MRI. When CT, MRI and CEA-scintigraphy were compared for the diagnosis oflocally recurrent rectal tumours, the diagnosis was most effectively establishedby MRl.

It is concluded that local excision of rectal tumours can presently not be performedbased on results of MRI. Sensitivity of CT and specificity of MRI is not sufficientto allow for general recommendations in the evaluation unresectable rectal tumours.However, If the cross-sectional investigation begins with CT, MRl will contributeto a more complete evaluation, especially if involvement of the bladder and the internalgenitalia cannot be completely ruled out.

Afler surgery, a tumour-free lateral resection margin can be verfied by MRI ofrectal specimen if the measured distance is more than one mm. Local tumour recurrenceand benign changes in the pelvis related to the treatment of the patient can presentlynot be differentiated on account of dynamic contrast-enhanced MRI. In the diagnosisof locally recurrent rectal cancer with cross-sectional evaluation, MRl should beconsidered as the first choice rather than CT or CEA-scintigraphy.

Key words: Rectum, Neoplasms, Magnetic Resonance Imaging

Stockholm 1997 ISBN-91-628-2797-9