Abstract Background Intraoperative ultrasonography during open surgery for colorectal cancer may be useful for the detection of unrecognized liver metastases. Laparoscopic ultrasonography (LUS) for the detection of unrecognized liver metastasis has not been studied in a randomized trial. This RCT tested the hypothesis that LUS would change the TNM stage and treatment strategy. Methods Patients with colorectal cancer and no known metastases were randomized (1 : 1) to laparoscopic examination (control or laparoscopy plus LUS) in three Danish centres. Neither participants nor staff were blinded to the group assignment. Results Three hundred patients were randomized, 150 in each group. After randomization, 43 patients were excluded, leaving 128 in the control group and 129 in the LUS group. Intraoperative T and N categories were not altered by LUS, but laparoscopy alone identified previously undetected M1 disease in one patient (0·8 per cent) in the control group and three (2·3 per cent) in the LUS group. In the latter group, LUS suggested that an additional six patients (4·7 per cent) had M1 disease with liver (4) or para-aortal lymph node (2) metastases. The change in treatment strategy was greater in the LUS than in the control group (7·8 (95 per cent c.i. 3·8 to 13·8) and 0·8 (0 to 4·2) per cent respectively; P = 0·010), but the suspected M1 disease was benign in half of the patients. Conclusion Routine LUS during resection of colorectal cancer is not recommended. Registration number: NCT02079389 (http://www.clinicaltrials.gov).