Skip secondary menu

Genetic Analysis of Childhood Brain Tumors Improves Diagnosis And Predicts Survival

For release: Friday, March 8, 2002

Doctors who treat brain tumors and other kinds of cancer have long struggled to understand why some patients respond well to therapy while others do not. In recent years, it has become clear that the answer lies at least partially in the genes. Two studies now show that identifying the "genetic fingerprints" of some childhood brain tumors can greatly improve diagnosis and predict patients' long-term survival. The findings help researchers understand how the tumors develop and may lead to improved ways of treating them.

The studies are among the first to tie genetic factors in brain tumors to differences in treatment outcome, the researchers say. The findings may lead to targeted treatments for particular tumor types or, ultimately, for individual patients. The studies were funded in part by the National Institute of Neurological Disorders and Stroke (NINDS).

The researchers focused on two common types of childhood brain tumors: medulloblastomas and malignant gliomas. Diagnosis of tumors is normally based on clinical factors, such as whether the tumor has spread, and on the appearance of tumor cells under a microscope. However, doctors are unable to fully predict patients' response to therapy using these methods.

In the first study,1 Scott L. Pomeroy, M.D., Ph.D., of Children's Hospital in Boston, and colleagues used a DNA microarray, or "gene chip," to identify genes that were active in tumors from patients with medulloblastomas and other childhood brain tumors. They were able to accurately distinguish between medulloblastomas and other types of tumors based on the combination of active genes in each tumor. Genetic analysis also helped to distinguish between two types of medulloblastoma and revealed that several of the genes active in a type called desmoplastic medulloblastoma were related to a protein called sonic hedgehog (SHH) that is important during brain development. Finally, the researchers showed that they could predict which children would be successfully treated for medulloblastoma based on the pattern of gene activity at the time the tumors were diagnosed.

The study shows that medulloblastoma is genetically different from other types of brain tumor and that patterns of gene activity are more accurate than clinical factors for predicting risk, says Dr. Pomeroy. The findings provide "proof of concept" that genetic analysis can distinguish between tumors that otherwise appear to be identical, he adds.

The findings may help to improve treatment of this disease. While most children treated for medulloblastoma survive, many have learning disabilities and other lasting problems that result from treatment. With routine genetic analysis, doctors may be able to identify patients at relatively low risk and reduce the amount of treatment, thereby reducing side effects for these individuals and helping them to live more normal lives, says Dr. Pomeroy. In addition, the finding that desmoplastic medulloblastomas have SHH protein activity suggests that drugs which counteract SHH may be useful in treating this disorder. One such drug, called cyclopamine, has already been identified, although it has not been tested in patients.

In the second study,2 a multicenter team of investigators led by Ian F. Pollack, M.D., of the University of Pittsburgh Medical Center and Children's Hospital of Pittsburgh, studied how mutations in a gene called TP53 and activity of the p53 protein that this gene produces relate to treatment outcome in children with malignant gliomas. Abnormalities, or mutations, in the TP53 gene have been linked to many kinds of tumors.

The researchers found that children with gliomas that had little p53 activity were much more likely than those with high p53 levels to survive for 5 years without tumor progression. Advanced (high-grade) gliomas generally have a poor prognosis. However, using p53 levels as a marker allowed the researchers to distinguish between a subgroup of children who did extremely poorly and others who fared somewhat better, says Dr. Pollack. The researchers are now planning to look at additional factors that may influence prognosis.

These strategies may help doctors distinguish between tumors that will respond successfully to standard treatments, such as chemotherapy, and tumors that may require more aggressive or experimental treatments, the investigators say. They also can help researchers understand what makes these tumors grow and what treatments will be most effective for each patient.

Genetic analysis has exciting implications not only for brain tumor diagnosis and treatment, but also for other diseases, says Thomas P. Jacobs, Ph.D., NINDS executive director of the Brain Tumor Progress Review Group, a joint initiative by the NINDS and the National Cancer Institute that aims to identify and prioritize research needs for brain tumor research. For example, differences in gene activity might help to distinguish between different types of epilepsy or different kinds of stroke, he says.

More research is needed to validate these methods for predicting treatment outcomes with different types of tumors, the researchers say. However, genetic analysis of tumors also may point to therapies that can alter the effects of specific genes in a focused way, without the severe side effects of traditional therapies. Ultimately, it may even be possible to design individualized courses of treatment aimed at the unique characteristics of each person's disease.


1 Pomeroy SL, Tamayo P, Gaasenbeek M, et. al. "Prediction of central nervous system embryonal tumour outcome based on gene expression." Nature, January 24, 2002, Vol. 415, pp. 436-442.

2 Pollack IF, Finkelstein SD, Woods J, et. al. "Expression of p53 and prognosis in children with malignant gliomas." The New England Journal of Medicine, February 7, 2002, Vol. 346, No. 6, pp. 420-427. February 2002, Vol. 8, No. 2, pp. 143-149.


- By Natalie Frazin


Related Items

Fact Sheet

Last Modified April 16, 2014