Background: Tumour extension beyond the mesorectal plane (ymrT4) occurs in 5-10 per cent of patients with rectal cancer and 10 per cent of patients develop locally recurrent rectal cancer (LRRC) after primary surgery. There is global variation in healthcare delivery for these conditions. Methods: An international benchmark trial of the management of ymrT4 tumours and LRRC was undertaken in France and Australia between 2015 and 2017. Heterogeneity in management and operative decision-making were analysed by comparison of surgical resection rates, blinded intercountry reading of pelvic MRI, quality-of-life assessment and qualitative evaluations. Results: Among 154 patients (97 in France and 57 in Australia), 31⋅8 per cent had ymrT4 disease and 68⋅2 per cent LRRC. The surgical resection rates were 88 and 79 per cent in France and Australia respectively (P = 0⋅112). The concordance in operative planning was low (= 0⋅314); the rate of pelvic exenteration was lower in France than Australia both in clinical practice (36 of 78 versus 34 of 40; P < 0⋅001) and in theoretical conditions (10 of 25 versus 50 of 57; P = 0⋅002). The R0 resection rate was lower in France than Australia for LRRC (25 of 49 versus 18 of 21; P = 0⋅007) but not for ymrT4 tumours (21 of 26 versus 15 of 15; P = 0⋅139). Morbidity rates were similar. Patients who underwent non-exenterative procedures had higher scores on the mental functioning subscale at 12 months (P = 0⋅047), and a lower level of distress at 6 months (P = 0⋅049). Qualitative analysis highlighted five categories of psychosocial factors influencing treatment decisions: patient, strategy, specialist, organization and culture. Conclusion: This international benchmark trial has highlighted the differences in worldwide treatment of locally advanced and LRRC. Standardized care should improve outcomes for these patients.