Detection and survival of colorectal cancer from a 2 week wait service

To determine detection strategies for colorectal cancer (CRC) and to analyse subsequent survival from a 2 week wait (2WW) service.

Retrospective analysis of 2WW and hospital CRC databases from January 2006 to July 2009. Survival was assessed using Kaplan Meier survival curves and Cox's proportional hazard models. Proximal cancers were those proximal to and including the splenic flexure.

From 1725 patients seen in a 2WW clinic, 108 were identified with CRC. Median follow-up in survivors was 514 days (interquartile range 160-788 days). Of 23 patients investigated for iron deficient anaemia and/or abdominal mass, 78% (18) were found to have proximal cancers. Of 85 patients with symptoms of change in bowel habit, rectal bleeding or abdominal pain, 15% (13) were found to have proximal cancers. Age, haemoglobin and mean corpuscular volumes between these 13 patients and the 72 distal cancers in this group were not significantly different. Multivariable analysis showed that survival was lower for those presenting with proximal cancers (hazard ratio 2.912, 95% confidence interval 1.361-6.227, p=0.006) and for those with increasing Dukes stage (p<0.001).

Flexible sigmoidoscopy would have missed 15% of cancers in those presenting with symptoms alone. Patients with proximal tumours had a worse prognosis. Further research is needed to identify those presenting with symptoms alone who are at high risk of having proximal tumours and thus requiring whole colonic imaging.