Gina Brown MBBS, Presenter: Nothing to Disclose

Ian Daniels MBBS, Abstract Co-Author: Nothing to Disclose

Andrew R Norman PhD, Abstract Co-Author: Nothing to Disclose

The presence of tumour at the circumferential resection margin (CRM) profoundly affects prognosis in rectal cancer making accurate preoperative prediction of CRM status crucial for selecting therapy.Currently there is no agreement for the cut-off value that should be employed for preoperatively defining patients with tumour free resection margins.We hypothesise that if tumour lies ≥1mm of the mesorectal fascia,this predicts for tumour free CRM in patients undergoing total mesorectal excision surgery.

A multi-centre, multidisciplinary European collaboration (MERCURY) was initiated to prospectively evaluate MRI in rectal cancer staging. Prediction of CRM status was based on the criterion of a 1mm cut-off.Of 715 consecutive patients enrolled between January 2002 and close of study in October 2003, 325 were eligible for this end-point assessing mesorectal fascia involvement on MRI compared with CRM involvement by tumour by histopathology. MRI was undertaken using thin high section high spatial resolution T2 weighted fast-spin echo according to a defined protocol. Standardised reporting proformas were employed for prospective, blinded radiological and histopathological data collection.The minimum distance of tumour to the mesorectal fascia was noted for each patient. The potential CRM was defined as involved if the distance on MRI was <1mm.On axial wholemount histopathological sections of the rectal cancer specimens, the CRM was involved if the distance of tumour to the margin was <1mm.The agreement of CRM status compared with the histopathological gold standard was assessed using the 1mm MRI cut-off.A 2mm MRI cut-off was also evaluated to determine impact on agreement.

Agreement between MRI and histopathology assessment of CRM status when MRI defined CRM involvement as tumour <1mm from the mesorectal fascia was 266/325 (82%, 95% CI 77-85%). Increasing this cut-off to 2mm worsened accuracy 200/325 (61%, 95% CI 56% - 67%).

Our study confirms that MR accurately predicts CRM status if a cut-off of 1mm is used. It is thus an effective tool for the preoperative identification of patients at risk of incomplete resection following total mesorectal excision surgery.

Brown, G, Daniels, I, Norman, A, MRI Predicts Surgical Resection Margin Status in Patients with Rectal Cancer: Results from the MERCURY Study Group.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4417078.html