A group of researchers from several US institutions has identified a gene expression signature that can predict which patients with idiopathic pulmonary fibrosis are likely to have a worse outcome, based on how long they lived with the disease before death or before requiring a lung transplant.
The researchers published the results of their initial discovery and validation in Science Translational Medicine earlier this month, and hope the signature could offer clinicians a tool to personalize care for IPF patients by better timing lung transplantation and more effectively stratifying patients in clinical trials for experimental therapy.
Naftali Kaminsky, the study's senior author and chief of pulmonary, critical care, and sleep medicine at the Yale School of Medicine, said that the group also hopes that further research into the genes implicated in the study may eventually inform development of additional therapies for the disease, or expanded indications of current drugs.
The lowest hanging fruit is transplantation, he told PGx Reporter this week.
"When you have a patient with IPF you don’t want to refer them too early, because evaluation is costly and has risks in itself," he said. Additionally, in areas where transplantation is less available, knowing which patients are likely to "crash and burn" would be important to make sure that the most needy receive transplants first.
On the clinical trial side, being able to predict patients likely to progress to a worse disease state could help make sure clinical trials are conducted on those for whom a drug is more likely to show a real effect, Kaminsky added.
IPF is a chronic and progressive disease with an overall median survival of about three years. However, individual patients can survive significantly shorter or longer than that average, Kaminsky said. And current clinical tools that help doctors predict how severe or mild a patient's disease will be leave a lot to be desired.
According to the study authors, the fact that outcomes in IPF are so variable and unpredictable has generated great interest in pinpointing molecular markers that may predict more reliably how patients will progress so doctors can plan and prescribe lung transplantation in a more personalized manner.
In the study, Kaminsky and his colleagues used microarrays to profile samples for genes differentially expressed in IPF patients with varying outcomes, measured in terms of transplant-free survival — the length of time they survived before either dying or having a lung transplant.
Based on microarray results from a discovery cohort of 45 patients recruited at the University of Chicago, the team identified an initial 52 genes whose expression differed between the groups with the better outcome (longer TFS) and the worse outcome (shorter TFS).
Looking at these 52 genes in a replication cohort of another 75 patients from the University of Pittsburgh, the researchers clustered patients based on their gene expression. Using the differences in expression again, they were able to distinguish two main clusters with significantly different outcomes.
The researchers also looked at specific pathways associated with different TFS, and found that one, including the genes CD28, ICOS, LCK, and ITK, had the strongest association.
The group looked at these four genes alone in both cohorts, and found that this smaller signature was also predictive of shorter or longer TFS. When the group compared the genomic predictor alone against a set of clinical factors, and a combined clinical and molecular model, the combined set was the most powerful in predicting which outcome group patients would fall into, with a statistically significant higher area under the receiver operating characteristic curve — 78.5 percent versus 76.6 percent for the genomic-only model, and 70.9 percent for clinical variables alone.
Following up on the current study, the group is now attempting to replicate, validate, and standardize the findings in two larger cohorts, Kaminsky said, as well as test whether the 52-gene set they identified might also predict other aspects of disease progression beyond overall survival.
The team also plans to look at whether patients' gene expression-based risk profiles change over time, with the hope that the signature could also be used to monitor disease.
According to Kaminsky, other follow-up studies on genes that had differential expression in this initial effort could also inform the development of targeted interventions, or provide evidence for expanding the indications for marketed drugs.
For example, "patients with IPF used to be treated with immunosupression for years with no results,” Kaminsky said. “Now we know that things like steroids, if anything, are bad for [these patients] as a whole, but that doesn’t mean its bad for everybody.”
However, since in this study researchers found that the two outcome groups differed in gene expression that could be associated with T-cell activation, “we could potentially identify a group that could still benefit from immunomodulation,” Kaminsky added.