Send to

Choose Destination
Br J Cancer. 2017 Nov 7;117(10):1478-1485. doi: 10.1038/bjc.2017.320. Epub 2017 Sep 21.

Comparison between MRI and pathology in the assessment of tumour regression grade in rectal cancer.

Author information

Department of Medicine, The Royal Marsden NHS Foundation Trust, London and Surrey, UK.
Department of Radiology, The Royal Marsden NHS Foundation Trust, London and Surrey, UK.
Department of Histopathology, The Royal Marsden NHS Foundation Trust, London and Surrey, UK.
Clinical Research & Development, The Royal Marsden NHS Foundation Trust, London and Surrey, UK.
Department of Medical Oncology, Vall d'Hebron University Hospital, Universitat Autònoma de Barcelona, Barcelona, Spain.
Department of Immunology, Genetics and Pathology, Section of Experimental and Clinical Oncology, University of Uppsala, Uppsala, Sweden.
Department of Haematology and Medical Oncology, Biomedical Research Institute INCLIVA, University of Valencia, Valencia, Spain.



Limited data exist regarding the correlation between MRI tumour regression grade (mrTRG) and pathological TRG (pTRG) in rectal cancer.


mrTRG and pTRG were compared in rectal cancer patients from two phase II trials (EXPERT and EXPERT-C). The agreement between radiologist and pathologist was assessed with the weighted κ test while the Kaplan-Meier method was used to estimate survival outcomes.


One hundred ninety-one patients were included. Median time from completion of neoadjuvant treatment to pre-operative MRI and surgery was 4.1 weeks (interquartile range (IQR): 3.7-4.7) and 6.6 weeks (IQR: 5.9-7.6), respectively. Fair agreement was found between mrTRG and pTRG when regression was classified according to standard five-tier systems (κ=0.24) or modified three-tier systems (κ=0.25). Sensitivity and specificity of mrTRG 1-2 (complete/good radiological regression) for the prediction of pathological complete response was 74.4% (95% CI: 58.8-86.5) and 62.8% (95% CI: 54.5-70.6), respectively. Survival outcomes of patients with intermediate pathological regression (pTRG 2) were numerically better if complete/good regression was also observed on imaging (mrTRG 1-2) compared to poor regression (mrTRG 3-5) (5-year recurrence-free survival 76.9% vs 65.9%, P=0.18; 5-year overall survival 80.6% vs 68.8%, P=0.22).


The agreement between mrTRG and pTRG is low and mrTRG cannot be used as a surrogate of pTRG. Further studies are warranted to assess the ability of mrTRG to identify pathological complete responders for the adoption of non-operative management strategies and to provide complementary prognostic information to pTRG for better risk-stratification after surgery.

[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Nature Publishing Group Icon for PubMed Central
Loading ...
Support Center