Predictive value of age and serum neuron enolase analysis in the pathological prognosis of histopathology subgroup in intermediate and high-risk neuroblastoma

Objective To improve substantially the prediction accuracy of histopathology subgroup through combined analysis of clinical and biological features, and accordingly build a decision tree to predict the histopathology subgroup in intermediate and high-risk neuroblastoma. Methods A total of 62 intermediate and high-risk neuroblastoma patients were included retrospectively in this study, who received chemotherapy and surgery at Beijing Children′s Hospital (BCH), Capital Medical University between January 2015 and January 2017. The fin-dings of urinary vanillylmandelic acid (VMA), and homovanillic acid (HVA), serum neuron-specific enolase (NSE), lactate dehydrogenase (LDH) and ferritin, ultrasound, CT, MRI, positron emission tomography-computed tomography (PET-CT), bone marrow aspiration and biopsy, MYCN gene, and histopathology were collected and analyzed retrospectively. Statistical analysis was performed by using SAS 9.4. Univariate and multivariate Logistic regression analysis were conducted to select potentially useful characteristics for prediction. Based on the results of Logistic regression analysis, a classification tree was developed to predict histopathology subgroup. Results To identify the characteristics related to histopathology, tumor markers and six clinico-pathologic factors were evaluated by univariate analysis.The results showed that unfavorable histopathology(UH) was more frequently associated with bone marrow metastasis, older age, as well as higher serum NSE, ferritin and LDH levels.The result of multivariate analysis showed that age and NSE were significant independent predictors of histopathology.The adjusted odds ratio(OR) of NSE and age was 33.2 and 13.0, respectively.The area under the receiver-operating-characteristic (AUC) of the prediction mo-del was 0.889.The sensitivity and specificity were 91.90% and 76.00%, respectively.Furthermore, to provide a visua-lization of the significant predictors found by Logistic regression analysis, a decision tree was developed for predicating of histopathology. Conclusions Age and NSE are significant independent predictors of histopathology subgroup.The decision tree based on age and NSE can help to predict the histopathology subgroup in intermediate and high-risk neuroblastoma effectively. Key words: Neuroblastoma; Histopathology subgroup; Age; Neuron specific enolase; Prediction