Abstract Aim The sigmoid take‐off (STO), the point on imaging where the sigmoid sweeps ventral from the sacrum, was chosen as the definition of the rectum during an international Delphi consensus meeting and has been incorporated into the Dutch guidelines since October 2019. The aim of this study was to evaluate the implementation of this landmark 1 year after the guideline implementation and to perform a quality assessment of the STO training. Method Dutch radiologists, surgeons, surgical residents, interns, PhD students and physician assistants were asked to complete a survey and classify 20 tumours on MRI as ‘below’, ‘on’ or ‘above’ the STO. Outcomes were agreement with the expert reference, inter‐rater variability and accuracy before and after the training. Results Eighty‐six collaborators participated. Six radiologists (32%) and 11 surgeons (73%) used the STO as the standard landmark to distinguish between rectal and sigmoidal tumours during multidisciplinary meetings. Overall agreement with the expert reference improved from 53% to 70% ( p < 0.001) after the training. The positive predictive value for diagnosing rectal tumours was high before and after the training (92% vs. 90%); the negative predictive value for diagnosing sigmoidal tumours improved from 39% to 63%. Conclusion Approximately half of the represented hospitals have implemented the new definition of rectal cancer 1 year after the implementation of the Dutch national guidelines. Overall baseline agreement with the expert reference and accuracy for the tumours around the STO was low, but improved significantly after training. These results highlight the added value of training in implementation of radiological landmarks to ensure unambiguous assessment.