The influence of endorectal filling on rectal cancer staging with MRI

Objective: To assess the influence of endorectal filling (EF) on rectal cancer staging. Methods: 47 patients who underwent a staging MRI of rectal cancer in the period from 2011 to 2014 were included. The MRI protocol included T-2 weighted fast spin echo sequences without and with EF at 3 T (EF-MRI). Images were scored by two readers for T-stage, distance of the lower pole of the tumour to the anorectal junction, distance to the mesorectal fascia (MRF), and number of (suspicious) lymph nodes. Agreement in T-staging was calculated using the Cohen's kappa value. Comparison of continuous variables was performed using Wilcoxon matched pairs signed-rank test. Results: The interobserver agreement for T-staging with and without EF-MRI showed a poor agreement between both readers (weighted kappa = 0.156, weighted kappa = 0.037, respectively). Tumours tended to be overstaged more prominently with EF-MRI. The accuracy of predicting the pathological T-stage slightly improved from 55% with EF to 64% without EF for Reader 1 and from 59 to 68% for Reader 2, respectively. The distance of the tumour to the anorectal junction increased from 33.9 to 49.3 mm (p < 0.001) after EF for Reader 2. EF-MRI did not significantly influence the number of (suspicious) lymph nodes and distance to the mesorectal fascia. Conclusion: EF-MRI did not lead to an improved tumour staging and it has the potential to influence the distance to a key anatomical landmark. EF-MRI is therefore not recommended in primary staging rectal cancer. Advances in knowledge: EF-MRI may not be used as an additional tool to stage rectal cancer patients, as it does not seem to facilitate in locoregionally staging the disease.