Using Gene Expression to Predict Chemotherapy Benefit in NSCLC


The decision to seek additional treatment after surgery (adjuvant chemotherapy) for non-small cell lung cancer (NSCLC) can be difficult, because chemotherapy responses can be highly patient-specific and are often accompanied by serious adverse symptoms. Discovering genetic signatures that are associated with good treatment outcomes could help identify patients who are likely to benefit from adjuvant chemotherapy. A new study in Clinical Cancer Research used this approach to isolate 12 genes that predict chemotherapy success in early-stage NSCLC.

Using tissue samples from adenocarcinomas, researchers at the University of Texas Southwestern Medical Center first identified 18 “hub” genes that were highly predictive of the prognosis in stages I-III NSCLC. Among these genes, RRM2, AURKA, PRC1, and CDKN3 were associated with poor prognosis. Across six different patient cohorts, those who were genetically predicted to be high risk had significantly shorter survival time. Even after adjusting for age, gender, and stage of cancer, the difference in survival time between predicted high-risk and low-risk groups remained. The 18-hub-gene signature appears to be specific for adenocarcinomas, as survival could not be reliably predicted when testing was done on squamous cell carcinoma samples.

Once the investigators had established the prognostic value of the 18-gene signature, they wanted to determine if the gene set could actually predict the survival benefit of chemotherapy in NSCLC. Using a subset of the data from patients who had undergone adjuvant chemotherapy after surgery, they narrowed down 12 survival-related genes that either affected response to chemotherapy drugs or had genetic aberrations in NSCLC: DOCK9, RRM2, AURKA, HOPX, NKX2-1, TTC37, COL4A3, IFT57, C1orf116, HSD17B6, MBIP, and ATP8A1. Nine of these genes are “synthetic lethal,” leading to cell death in NSCLC in the presence of the drug paclitaxel.

Classifying patients as genetically high risk or low risk based on these 12 genes, the researchers found that, with adjuvant chemotherapy, the high-risk group had longer survival while the low-risk group did not receive any significant survival benefit. In fact, the low-risk group experienced worse survival outcomes in the first 21 months after receiving adjuvant chemotherapy. The 12-gene set is predictive for NSCLC chemotherapy outcome with paclitaxel or vinorelbine plus cisplatin treatment and for the combination of carboplatin plus taxanes.

Many of the genes in the set of 18 hub genes are related to cancer metastasis or tumor cell proliferation. This hub gene set was also better at predicting prognosis than a previously used set of top genes associated with survival outcome. This better performance is likely linked to lower information redundancy in the 18-hub-gene set, which captures more variance across the patient population, making it superior for prognosis across different datasets and DNA testing platforms.