Objective: To compare the clinical ability of MRl taken before and after neo-adjuvant treatment in locally advanced rectal cancer (LARC) to predict the necessary extension of TME (ETME) and the possibility to achieve a R0 resection.
Method: Prospective registration of 92 MRI evaluated T4a cancers undergoing elective surgery between 2002 and 2007 in a tertiary referral centre for multimodal treatment of rectal cancer.
Results: MRI identified patients in need of neo-adjuvant treatment and predicted T-downstaging in 10% and N-downstaging in 59%. Seventy-nine percent R0 resections, 18% R1 and 3% R2 were obtained after ETME in 95% of the patients and TME in the rest. Higher tumour regression grade (TRG) was achieved in higher ypT-stage (P < 0.01). Preoperative chemo radiotherapy resulted in that more patients obtained TRG1-3 compared to those receiving radiotherapy (79% vs. 57%, P = 0.02). The pelvic wall was the area of failure in 70% of the R1 resections. Tumour cells outside the mesorectal fascia scattered within fibrosis was found in 18 TRG2-3 among 33 ypT4 tumours (55%).
Conclusion: MRl cannot discriminate tumour within fibrosis. Therefore, if a R0 resection is the goal, we advocate optimal surgery in accordance with the pre-treatment MRI. Post treatment MRI is a poor predictor of final histology and should not be relied upon to guide the extent of surgical resection. The study has initiated a new approach to histopathological classification of the removed specimen where we introduce a MRI assisted technique for investigating the areas at risk outside the mesorectal fascia in the specimen.