The present landscape of clinical trials for metastatic colorectal cancer is dominated by visceral metastases, as was highlighted at the recent American Society of Clinical Oncology Gastrointestinal Cancers conference held in San Francisco, CA, USA, in 2017. Peritoneal metastases are difficult to image by cross-sectional imaging and this leads to a disproportionate underrepresentation of this site of metastases in clinical trials. Peritoneal metastases differ in their presentation from visceral metastases, which are often incidentally detected. Peritoneal metastases tend to be more symptomatic, leading to bowel obstructions, hydroureter, and ascites, which rapidly lead to inanition and death. Additionally, peritoneal metastases tend to have a higher percentage of the worse prognosis BRAF-mutated tumours compared with other sites. This is seen in pooled analysis of NCCTG trials, which showed a median survival of 12· 7 months compared with 17· 6 months for other disease sites (hazard ratio 1· 32, 95% CI 1· 15–1· 50, p= 0· 001). 1 The true incidence of isolated peritoneal-only metastases is difficult to ascertain. The National Comprehensive Cancer Network guidelines quote an incidence of 2%, extrapolated from the pooled analysis of clinical trials. This might be disproportionately lower than the true incidence of isolated peritoneal disease, because of the systematic exclusion of such patients from the same clinical trials that were used to calculate incidence. In fact, autopsy series of 5817 autopsies revealed an incidence of 6% isolated peritoneal metastases in adenocarcinomas and 15% isolated peritoneal metastases in mucinous adenocarcinomas and signet-ring cell carcinomas. 2 The percentage of patients with any peritoneal metastases was 20% in adenocarcinomas, 48% in mucinous adenocarcinomas, and 51% in signet-ring cell carcinoma. Although it is possible that patients dying of peritoneal disease are over-represented in autopsy series, it is intriguing to