MRI can non-invasively characterise brain tumours
24 March 2008
Magnetic resonance imaging (MRI) technology has the potential to
non-invasively characterise tumours and determine which of them may be
responsive to specific forms of treatment, based on their specific
molecular properties, according to researchers at the University of
California, San Diego (UCSD) School of Medicine. The study will be
published online by the Proceedings of the National Academy of
Science (PNAS) this week.
Normally tumour tissue and cell samples are obtained through invasive
biopsy or surgery and studied under a microscope to diagnose and
“This approach reveals that, using existing imaging techniques, we
can identify the molecular properties of tumours,” said Dr Michael Kuo,
assistant professor of interventional radiology at UCSD School of
Medicine. Kuo and colleagues analyzed more than 2,000 genes that had
previously been shown to have altered expression in Glioblastoma
multiforme (GBM) tumours. They then mapped the correlations between gene
expression and MRI features.
The researchers also identified characteristic imaging features
associated with overall survival of patients with GBM, the most common
and lethal type of primary brain tumour.
The researchers discovered five distinct MRI features that were
significantly linked with particular gene expression patterns. For
example, one specific characteristic seen in some images is associated
with proliferation of the tumour, and another with growth and formation
of new blood vessels within the tumour–both of which are susceptible to
treatment with specific drugs.
These physiological changes seen in the images are caused by genetic
programs, or patterns of gene activation within the tumour cells. Some
of these programs are tightly associated with drug targets, so when they
are detected, they could indicate which patients would respond to a
particular anti-cancer therapy, according to the researchers.
“For the first time, we have shown that the activity of specific
molecular programs in these tumours can be determined based on MRI scans
alone,” said Kuo. “We were also able to link the MRI with a group of
genes that appear to be involved in tumour cell invasion — a phenotype
associated with a reduced rate of patient survival.”
Laboratory work that relies on tissue samples is routinely used to
diagnose and guide treatment for GBM. However, the biological activity
shown may depend on the portion of the tumour from which the tissue
sample is obtained. The researchers have shown that MRI could be used to
identify differences in gene expression programs within the same tumour.
“Gene expression results in the production of proteins, which largely
determine a tumour's characteristics and behaviour. This non-invasive
MRI method could, for example, detect which part of a tumour expresses
genes related to blood vessel formation and growth or tumour cell
invasion,” said Kuo. “Understanding the genetic activity could prove to
be a very strong predictor of survival in patients, and help explain why
some patients have better outcomes than others.”
Kuo also led a study, published in Nature Biotechnology in May 2007,
correlating CT images of cancerous tissue with gene expression patterns
in liver tumors. “In the new study, we were able to take a different
imaging technology, MRI, and apply it to a totally different tumour
type,” he said, noting that the studies open up promising new avenues
for non-invasive diagnoses and classification of cancer.