Gene Signature Predicts Response to Immune-Checkpoint Blockade in Lung Cancer

By Will Boggs MD

December 10, 2019

NEW YORK (Reuters Health) - A combined gene-signature score accurately predicts responses to immune-checkpoint blockade (ICB) in patients with non-small-cell lung cancer (NSCLC), researchers report.

"If validated, these genomic scores could be used to help physicians and patients understand the probability that they would respond to checkpoint blockade," Dr. Jeffrey C. Thompson of the University of Pennsylvania Perelman School of Medicine, in Philadelphia, told Reuters Health by email.

While ICB using anti-PD1 or anti-PD-L1 antibodies has been shown to enhance response rates and improve survival in patients with NSCLC, only a subset of patients achieve durable clinical benefit.

Dr. Thompson's team used data from pretreatment tumor specimens from 52 patients with metastatic NSCLC treated with ICB monotherapy to assess the predictive power of genomic signatures reflecting inflammation, chronic interferon stimulation, the IPRES genes (reflecting innate anti-PD1 resistance), TGF-beta-related genes, and the level of epithelial-to-mesenchymal transition (EMT) of tumor cells.

Nearly half (48%) of the patients had a clinical response (partial response or stable disease) to ICB, while the rest had progressive disease.

PD-L1-gene-expression levels did not differ significantly between responders and nonresponders, but inflammation-signature scores were significantly higher in responders than in nonresponders.

EMT-signature scores were significantly lower (i.e., more epithelial) in the responder group than in the nonresponders group, the researchers report in Lung Cancer, online October 18.

IPRES scores, chronic interferon scores, and TGF-beta signature scores did not differ significantly between responders and nonresponders.

An unweighted score combining EMT and inflammation signature scores was significantly higher in responders than in nonresponders and differentiated these groups with 81% accuracy (using AUROC). A weighted score combining these scores differentiated responders and nonresponders with 92% accuracy.

A weighted combined score above the median was associated with 68% lower risk of progression and 60% lower overall mortality, compared with a score below the median.

"This indicates that both the immune status of the patient (reflected in the inflammation score) and, independently, properties of the tumor cells (the EMT status) play a role in response to checkpoint blockade," Dr. Thompson said.

"If the scores predict a very low chance of response, a better risk/benefit ratio could be estimated as checkpoint inhibitors have significant side effects and are very expensive," he said. "A score predicting lack of response could also trigger the initiation of alternative therapies or enrollment in clinical trials. It may also be possible to alter the immune status of the patient or the tumor's characteristics to make it more likely to respond."

"While the development of immune-checkpoint inhibitors has changed the treatment paradigm for patients with NSCLC, only a subset of patients respond," Dr. Thompson said. "Improving our understanding of the tumor-immune interactions is critical to optimizing a 'precision medicine' approach tailored to the specifics of a given patient's immune system and tumor characteristics. These types of data are also important in identifying resistance mechanisms to immunotherapeutic approaches and developing novel therapeutics."

He added, "Gene signatures, along with tissue tumor mutation burden (TMB) and tissue PD-L1 expression, will likely all play a role in patient selection for immunotherapy approaches."

Janssen Research and Development support in this study and employed several of the authors.


Lung Cancer 2019.