Editors' ChoiceCancer

Divide and Conquer

See allHide authors and affiliations

Science Translational Medicine  07 Jul 2010:
Vol. 2, Issue 39, pp. 39ec109
DOI: 10.1126/scitranslmed.3001432

To the patient, calling a tumor benign implies that it will not do much harm, whereas a malignant tumor is likely to metastasize and invade into neighboring tissues. To the pathologist, this distinction has special meaning, referring to how abnormal tumor cells look under the microscope and subsequently quantified as tumor grade. In most tumors, especially in sarcoma, tumor grade is one of the main predictors of tumor behavior, presumably because both cell behavior and nuclear morphology are both dependent on whether the cell has a normal set of chromosomes. However, even among patients with tumors of the same type and of the same grade, there is significant heterogeneity. Multiple studies over the last decade have exploited technological advances in genomic analysis of both DNA and RNA in order to increase our ability to predict tumor behavior. In sarcoma, some idea about the underlying genetic derangements can be gleaned from histology: Ewing sarcomas are generally considered normal at the genome level, with the exception of a single chromosomal translocation. Similarly, dedifferentiated liposarcomas carry a single abnormality: amplification of a portion of chromosome 12. However, the vast majority of other sarcomas have complex, highly abnormal sets of chromosomes.

Chibon and colleagues focused on the sarcomas that do not carry translocations and performed array comparative genomic hybridization so as to quantify the degree of aneusomy, or abnormal chromosome content. Surprisingly, the degree of aneusomy did not correlate with prognosis. Instead, the authors developed a very powerful predictor of prognosis on the basis of gene expression of a collection of 67 genes thought to be involved in mitosis and chromosomal stability. This index, which they termed CINSARC, predicted prognosis in sarcomas independently of any other factor, including tumor grade. The authors validated the index in an independent set of sarcoma cases and also in breast cancer and lymphoma. The prognostic power of the index is striking: For example, it can divide a seemingly identical population of high-grade leiomyosarcoma patients into those with a 75% and a 19% chance of developing metastases in 5 years.

The authors are currently testing CINSARC in a large prospective clinical trial. As an index applicable to multiple tumor types, it has the potential to become a routine test on par with tumor grade. But it remains a mystery why these 67 genes predict prognosis much more powerfully than the direct measurement of chromosome content itself.

F. Chibon et al., Validated prediction of clinical outcome in sarcomas and multiple types of cancer on the basis of a gene expression signature related to genome complexity. Nat. Med. 16, 781–787 (2010).[Abstract]

Navigate This Article