Article Text

Download PDFPDF
PWE-328 Accuracy of pre-operative mri assessment of tumour staging in rectal cancer
  1. HG Jones1,
  2. J Williamson1,
  3. T Wells2,
  4. P Chowdhury2,
  5. J Beynon1,
  6. D Harris1
  1. 1Department of Colorectal Surgery, Singleton Hospital, Swansea, UK
  2. 2Department of Radiology, Singleton Hospital, Swansea, UK


Introduction Magnetic resonance imaging (MRI) is routinely used for pre-operative rectal cancer (RC) staging. Several meta-analyses have demonstrated that MRI has good sensitivity and specificity when assessing the T stage and the circumferential resection margin (CRM), but was consistently poor for detection of lymph node metastases.1,2The current study aims to compare the pre-operative staging on the MRI scan with the post-operative pathological stage in patients patients not undergoing neoadjuvant therapy in a single tertiary colorectal centre.

Method 20 blinded MRI scans of the pelvis in patients with RC were given to two consultant gastrointestinal radiologists and their interpretation was compared to each other, the original MRI report, and to the final pathological report. These scans were reported by the use of a standardised proforma. A Cohen–s kappa coefficient analysis was performed on 145 historic RC cases to test the correlation between T and N stage on pre-operative MRI scans compared to the definitive histopathological result.

Results There was inter-reporter agreement on T and N stage in 11/20 (55%) and 10/20 (50%) of scans respectively. Agreement on EMVI and CRM positivity was 60% and 75% respectively. When these results were compared to the pathology results, at least one radiologist had the correct T and N stage in 60% and 75% of cases using the proforma, compared to 35% for T and N stage on the original MRI report. When 145 MRI reports were retrospectively reviewed (all reported without a proforma), the kappa statistic (κ) of 0.131 indicated a poor strength of agreement between the MRI stage and the pathological stage. Similarly T stage was predicted correctly in only 42.4% (76/145) of cases, and a kappa statistic (κ) of 0.135.

Conclusion There is a growing consensus is the medical literature that we should not be relying on MRI scans to accurately stage RC. The discrepancy between MRI staging and histology may be explained by variable quality of the scans between different machines and institutions, and standardising the scanning protocols may improve accuracy. The findings fit with the growing consensus in the medical literature that we should not be relying on MRI scans to accurately stage RC. Dual consultant reporting of pelvic MRI scans, as well as the utilisation of the reporting proforma is a potential way of increasing the accuracy of this staging tool.

Disclosure of interest None Declared.


  1. Kwok H, Bissett IP, Hill GL. Preoperative staging of rectal cancer. Int J Colorectal Dis. 2000;15

  2. Shandra B, et al. Rectal cancer: local staging and assessment of lymph node involvement with endoluminal US, CT, and MR imaging—a meta-analysis 1. Radiology 2004;232

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.