Author response to: Comment on: Transanal total mesorectal excision for rectal cancer has been abandoned in Norway

Editor We thank Warrier and colleagues for their interesting comments on the Norwegian negative results study. Indications for chemoradiotherapy (CRT) are strict in Norway, but were the same for both transanal total mesorectal excision (TaTME) patients and those in the national cohort. In the Norwegian study, the stages are given after pathological examination, not before, as Warrier et al. may have thought. We do agree that the patient with a pT4a N2 tumour might have benefitted from CRT. However, the effect of CRT on local recurrence (LR) did not affect the hazard ratio for LR between the TaTME group and the national cohort. Circumferential resection margin (CRM) less than 1 mm is a risk factor for LR in non-radiated tumours1. A short CRM is believed to act as a surrogate for the risk of LR. In this study, 12 per cent had a CRM of 1 mm or less, but only 5⋅1 per cent had a CRM less than 1 mm. Few of the LRs had a short CRM. The high LR rate does not seem to be a consequence of a short CRM. We believe studies that rely on CRM or other surrogate endpoints are of less value in establishing TaTME as a safe oncological method. As the results revealed, tumours were less advanced than in the national cohort and can be characterized as ‘easy’. Lessadvanced tumours could give excellent results with TaTME, as well as other methods. Even before total mesorectal excision, 80 per cent of patients had no LR2. This presents a very high risk of selection bias in observational studies. The results adjusted for case mix and selection bias cannot be explained by a failure to complete the learning curve. To advocate a long learning curve in modern cancer surgery when excellent simple methods are available is unacceptable. The method itself must be critically evaluated3–5.