Classification: the great deception.

Abstract

The linchpin of oncologic statistics might well be thought to be classification based upon the hope that comparisons of natural history and treatment regimes could be compared and controlled worldwide. Many such systems exist for sites within the head and neck; all possess varying degrees of intrinsic error--usually unquantified! A review of three sites by an experienced clinician highlights some of these deficiencies together with their likely significance. The conclusion is reached that although encouraging better recording of clinical and radiologic assessment, cancer statistics based upon existing classification systems can best be compared with a pyramid perched precariously upon its apex!