Provided sufficient numbers of cases are studied, analysis of studies comparing antigen expression for diagnostic purposes in tumours from two body sites is likely to be supported in the wider literature. The design of such comparative studies is informed by aggregating data from single tumour studies.