Last week’s BMJ cast yet further doubt on prostate specific antigen as a screening tool for prostate cancer (2009;339:b3572, b3537, b3601), and this week we warm to the theme of tumour markers and their limitations. The Clinical Review (doi:10.1136/bmj.b3527) and the Practice article on rational testing (doi:10.1136/bmj.b3111) make clear that tumour markers are a mixed bag, not to be used in screening, unhelpful for diagnosis in patients with non-specific symptoms, and best reserved for monitoring once a diagnosis has been made. Unfocused requests and inappropriate use of tumour markers cost health services around the world large sums of money. They provide …