BACKGROUND DATA: MRI assessment of rectal cancer not only assesses tumour depth and surgical resectability but also extramural disease which affects prognosis. We have observed that non-nodal tumour nodules (tumour deposits; mrTDs) have a distinct MRI appearance compared to lymph node metastases (mrLNMs). OBJECTIVE: We aimed to assess whether mrTDs and mrLNMs have different prognostic implications and compare these to other known prognostic markers. METHODS: This was a retrospective cohort study of 233 patients undergoing resection for rectal cancer from January 2007-October 2015. Data were obtained from electronic records and MRIs blindly re-reported. Survival was determined using Kaplan-Meier method. Prognostic markers were evaluated using Cox regression and competing risks analysis. Inter-observer agreement for mrTD was measured using Cohen's Kappa. RESULTS: On multivariable analysis, baseline mrTD/mrEMVI (extramural venous invasion) status was the only significant MRI factor for adverse survival (HR 2.36 (1.54-3.61) for OS, 2.37 (1.47-3.80) for DFS (both p < 0.001), superseding T and N categories. mrLNMs were associated with good prognosis (HR 0.50 (0.31-0.80)p= 0.004 for OS, 0.60 (0.40-0.90)p = 0.014 for DFS). On multivariable analysis, mrTDs/mrEMVI were strongly associated with distant recurrence (HR 6.53 (2.52-16.91) p = < 0.001) whereas T and N category were not. In a subgroup analysis of post-treatment MRIs in post-chemoradiotherapy (CRT) patients, mrTD/mrEMVI status was again the only significant prognostic factor; furthermore those who showed a good treatment response had a prognosis similar to patients who were negative at baseline. Inter-observer agreement for detection of mrTDs was κ0.77 and κ0.83. CONCLUSION: Current MRI staging predicting T and N status does not adequately predict prognosis. Positive mrTD/mrEMVI status has greater prognostic accuracy and would be superior in determining treatment and follow-up protocols. CRT may be a highly effective treatment strategy in mrTD/mrEMVI positive patients.