News Analysis
The Year Ahead
and Finance
Science, Technology,
and Medicine
The Biology of Hope
AIDS Research
Old drugs, new indications
Immune Modulators
Plastic Surgery
Drug delivery: the Biopump
DNA replication
Junk DNA
The polypill
Molecular pathology
Tom Cech
Stanley Falkow
Heino von Prondzynski
The Bibliome
Future Trend
Policy, Society,
and Ethics
Common Sense
Editor's Letter


Forensic Evidence

Genetic clues may help solve the case of cancer.

The first thing a detective wants to do is visit the scene of the crime. For sleuths investigating cancer, pinpointing where malignancy first took hold in a patientís body is by no means clear-cut. A CAT scan may reveal the presence of a lump, and a biopsy can determine whether or not it is malignant, but finding something in the liver does not exclude the possibility that the original lesion formed in the breast, colon, ovary, or lung.

Cancer is today defined by the organ in which it arises and by its microscopic appearance, so for the 60,000 patients in the United States whose primary cancer site is unknown, treatment is difficult. Hope may be at hand, however, from DNA microarray technology, which can distinguish between types of tissue based upon gene expression. Since cancer is the result of aberrant gene expression, it is an obvious first application of this tool. A study of DNA expression profiling that will get under way within a year will be one of the first initiatives in what promises to be a revolution leading diagnosis into the era of molecular pathology.

Recent research has shown that cancers have distinct patterns of gene expression that reflect their tissue of origin, their likelihood of developing metastases, and their response to therapy. For example, in 2002, investigators at the Netherlands Cancer Institute and at Rosetta Inpharmatics of Kirkland, Washington, confirmed that a previously identified set of 70 genes could predict the natural history of breast cancer, allowing prospective classification of patients into those with good or poor prognoses.1 The teamís DNA microarray data were better at predicting tumor behavior than were traditional histologic and clinical parameters.


1  Van de Vijver, M.J. et al. A gene-expression signature as a predictor of survival in breast cancer. New England Journal of Medicine 347:1999Ė2009.