Epigenome-wide association study reveals CpG sites related to COG of neuroblastoma

Hao Zhao,Hu Sun,Dongyun Zhao,Xiaojun Zhou,Xingang Li,Hongfei Liu,Bin Huang,Yinghao Gu

doi:10.1042/bsr20200826

Hao Zhao, Hu Sun + Show 6 more

Open Access

https://doi.org/10.1042/bsr20200826

Copy DOI

Abstract

Background. Neuroblastoma (NB) is the most common extracranial solid tumor in infants and children. Its variable location and complex pathogenesis make NB hard for early diagnosis and risk classification. Methodology. We analyzed the methylation data of 236 samples from patients with NB in Therapeutically Applicable Research to Generate Effective Treatments (TARGET) database. Kaplan–Meier survival analysis was used for comparing overall survival of NB patients in different groups. Epigenome-wide association study (EWAS) was conducted to screen CpGs significantly associated with NB patients’ Children’s Oncology Group (COG). Logistic regression method was used for constructing a model to predict NB patients’ COG. Results. NB patients in low COG showed significantly superior prognosis than those in high COG. A total of seven CpG sites were found closely related to COG. Logistic regression model based on those CpGs showed superior performance in separating NB patients in different COGs. Conclusions. The present study highlights the important role of DNA methylation in NB development, which might provide evidence for treatment decisions for children NB.

Highlights

Neuroblastoma (NB) is one of the most common malignant tumors in infants and children
hazard rate (HR) > 1 indicated that samples with the factor had a higher risk of death than reference, and HR < 1 indicated samples with the factor had a lower risk of death than reference
Using information-gain-based approach to sort the reliability of the seven CpG island (CpG) sites in predicting Children’s Oncology Group (COG) risk group from high to low, we found that the prediction accuracy of cg25241559 + cg20989926 reached 0.84, while the accuracy of the points did not significantly increase, indicating that the model constructed by these two CpG sites was accurate enough for COG group prediction (Figure 3C)

Summary

Introduction

Neuroblastoma (NB) is one of the most common malignant tumors in infants and children. Appropriate treatments based on risk classification could reduce the treatment-related toxicities and improve the prognosis. The occult primary site and diverse phenotypes make NB hard for early diagnosis and most of the patients were at high risk when diagnosed. Neuroblastoma (NB) is the most common extracranial solid tumor in infants and children. Its variable location and complex pathogenesis make NB hard for early diagnosis and risk classification. Epigenome-wide association study (EWAS) was conducted to screen CpGs significantly associated with NB patients’ Children’s Oncology Group (COG). NB patients in low COG showed significantly superior prognosis than those in high COG. Logistic regression model based on those CpGs showed superior performance in separating NB patients in different COGs. Conclusions. The present study highlights the important role of DNA methylation in NB development, which might provide evidence for treatment decisions for children NB

Methods

Results

Discussion

Conclusion