Background:
Survival of high-risk neuroblastoma (NB) patients remains poor. Advances in our understanding of neuroblastoma biology are critical for improving outcome. Hypoxia is a condition of low-oxygen tension occurring in poorly or abnormally vascularized tissues related to tumor aggressiveness. We report on the clinical significance of new 7 genes hypoxia signature.
Methods:
Gene expression of 498 samples from NB patients of all stages was measured by RNA-seq technology. Dataset was split into two groups of 49 and 449 patients to build a classifier and test its prognostic power. Classifier was built utilizing Logic Learning Machine algorithm (LLM).
Results:
We defined a new seven-gene signature (NB-hop) to assess hypoxia in primary NB tumors (PGK1, PDK1, EGLN1, ALDOC, FAM162a, AK4, and MTFP1 genes). A multi-gene classifier was built based on the expression of the aforementioned genes and patients’ outcome. In the test cohort, the clinical significance of NB-hop predictions was assessed in relevant subgroups of patients defined by combinations of established clinical risk factors. NB-hop stratified patients into groups with significantly divergent OS (P =.006) and EFS (P =.03) in the population of high-risk patients with metastatic neuroblastoma (INSS 4) diagnosed after 18 months of age. Furthermore, NB-hop classifier distinguished patients with different probability of relapse in the population of localized (INSS 1, 2) or metastatic (INSS 4s) MYCN not amplified tumors, which are notoriously at low-risk of relapse (P=.004). The NB-hop classifier emerged as an independent risk factor in multivariate cox regression models on both OS and EFS (OS HR 2.4; 95% CI 1.4-4.2 and EFS HR 2.5; 95% CI 1.7-3.8, both P<.001).
Conclusions:
NB-hop is a novel independent prognostic factor capable of improving stratification of clinically relevant neuroblastoma populations while measuring tumor hypoxia. NB-hop may guide the use of hypoxia-activated prodrugs.