As the most common extracranial solid tumor in children, there are many studies related to the prognosis, diagnosis, and treatment of NB. Although there have been many studies on the prognosis of NB, most of them only involved an individual prognostic factor or only investigated the factors influencing the overall survival (OS) of NB patients. Fewer authors have studied the prognostic factors for the tumor-specific survival (CCS) and event-free survival (EFS) of NB. Most children with NB were already in the INSS stage 4 and COG high-risk group at the time of diagnosis, so their prognosis was generally poor, and they were prone to tumor recurrence or metastasis or even death after systematic treatment with surgery combined with chemotherapy (16). The recurrence and metastasis of tumors, even if they did not cause immediate death, could bring considerable trauma and burden to the physical and mental health of the children and their families (17).
Nomogram is a multiple-indicator combination model that predicts disease occurrence or progression in tumor survival prediction. The advantage of nomograms is that the total score is calculated based on the values of the patient’s predictive variables, simplifying a complex statistical prediction model involving a considerable number of variables into a simple numerical prediction model to predict the occurrence risk of an event or the probability of survival (18). In a pre-study of ours involving the NB patient data from the TARGET database, we found that 341 of the 448 (76.1%) dead patients experienced critical events such as tumor recurrence or progression before death, while 120 of the 667 (18.0%) currently alive patients had experienced such events. Therefore, to identify separate predictive factors associated with EFS outcomes, we conducted a large population-based data analysis of NB patients based on the TARGET database and developed a nomogram to predict EFS in NB patients. The constructed nomogram in this study provided a quantifiable prediction of EFS for each NB patient because it could easily incorporate the key prognostic predictors and balance the effects among them. Moreover, a new risk stratification system based on the nomogram was constructed to allow clinicians to make better choices about patient treatment.
Therefore, we investigated the influencing factors of EFS in NB patients by performing a retrospective study involving 763 patients from the TARGET database and concluded that age at diagnosis > 1425 days, INSS stage 4, and DNA diploid were independent predictive risk factors. Then, a nomogram was developed to predict the 3-, 5-, and 10-year EFS of NB. There was no significant deviation between the event-free survival rates of the training and the validation sets, suggesting that the nomogram has good discriminatory capability and predictive accuracy. A risk stratification system for the EFS of NB patients based on the abovementioned three risk factors was constructed subsequently. As determined by the nomogram overall point, NB patients were categorized into low- (< 125), middle- (125–133), and high- (> 133) risk subgroups, and the EFS of the three subgroups differed significantly (p < 0.001).
Age at diagnosis is considered an important factor affecting the prognosis of patients with various tumors, and NB is no exception (19). As early as 2005, London et al. found that NB patients with a diagnosis age < 18 months had a greater chance of experiencing spontaneous tumor regression and were more likely to be cured by surgery alone (20). In addition, age at diagnosis is used as an important basis for the INGRSS and COG risk subgroups. Older children with NB usually have tumors that tend to earlier recurrence; correspondingly, the prognosis of these patients is poor, which might be related to the fact that older children are more susceptible to invasive tumors that are insensitive to multimodal and cytotoxic therapy (19, 21, 22). In this study, age also served as one of the independent prognostic factors, and patients aged > 1425 days had a significantly worse prognosis. As for gender and race, both uni- and multivariate Cox analyses demonstrated that neither of them were independent predictive factors for the EFS of NB patients (p > 0.05).
According to previous studies, tumor size and stage may affect OS in patients with NB (23). Wang et al. reported that the primary tumor size was considered a key prognostic factor for NB, with tumors > 4 cm suggesting a poor prognosis (24). Previously, the OS of NB patients with distant metastasis and INSS stage 4 was reported to be significantly lower than that of patients with low INSS stage and regional tumors (25). Our findings supported previous reports that higher INSS stage and COG risk groups were associated with poorer outcomes of EFS in NB patients. Both the uni- and multivariate Cox analysis identified INSS stage as an important predictor for the EFS of NB patients. These trends further demonstrated the importance of the early diagnosis and treatment of NB to improve the patient’s survival and reduce the rate of tumor recurrence and metastasis.
MYCN gene amplification status and DNA ploidy also had important prognostic influences on NB patients (26). According to previous reports, MYCN gene amplification is associated with primary giant abdominal tumors, chromosomal aberrations, and poor prognostic histological type (27). It has also been reported that patients with DNA hyperdiploid had a better prognosis and were characterized by chromosomal instability and that the aggressiveness of tumor cells in NB might be related to the degree of chromosomal instability (28). The univariate Cox analysis in this study depicted that MYCN status and DNA ploidy were associated with the EFS of NB patients. However, the multivariate Cox analysis excluded MYCN status and considered DNA ploidy to be an independent prognostic factor.
In conclusion, the nomogram and the risk stratification system established could better predict and classify the prognosis of NB patients. However, there were some shortcomings of this study: (I) as a retrospective study, the selection bias was inevitable; (II) the TARGET database does not contain certain detailed diagnostic and treatment data, such as whether chemotherapy, radiotherapy, and immunotherapy were administered or other specific treatment information of patients; (III) the predictive accuracy of our nomogram has not been validated with patient data from other centers or databases yet.