In some cases of cancer, the major challenge is to identify
the site at which the cancer initially arose. A cancer of unknown primary
origin (CUP) refers to a disorder in which the location of the primary tumor
remains a mystery, even after routine diagnostic tests and biopsy. Not knowing
the primary source of a cancer can setback the development of an effective
treatment plan for the patient and reduce their overall chance of survival.
This week in Genome Medicine, Olli
Kallioniemi and colleagues at the University of Helsinki present an improved
method for pinpointing the tissue origin of a primary tumor, using gene expression data. Their method could
improve the accuracy of diagnosis and pave the way for tailored anti-cancer
therapy in CUP patients.
A number of methods for predicting the site of tumor origin
have been described, and most of these compare the gene expression profile of
the CUP sample with a “reference set” of tumor-specific hallmarks. These
conventional approaches rely on an a
priori defined set of genes, limiting the adaptability of the method to
emerging information about specific cancers. The method described by
Kallioniemi and colleagues, wAGEP (weighted Alignment of Gene Expression
Profiles) can be adapted to any reference dataset, allowing it to be
continually optimized as new tumor expression data becomes available. As well as being flexible, wAGEP proved to be highly accurate in classifying
tumor samples according to tissue origin.
Another key advantage of wAGEP is that it can be used to investigate a CUP case on a
gene-by-gene level, so provides information about the individual genes involved
in initiating the cancer and also driving its spread to different tissue types
(metastasis). Every cancer is unique and
its evolution can be multifaceted, but by defining some of the systematic
changes involved, novel molecular predictors could be revealed.
The method described is accurate, scalable and enables gene-by-gene
analysis of cancers of unknown primary origin. Application of wAGEP in a
clinical setting could allow the rapid diagnosis of unclassified cancers and
implementation of personalized treatment regimes.