Abstract
Background
The causal relationship between interleukin-7 levels and neuroblastoma risk remains unclear, and optimal radiation therapy timing lacks definitive evidence. This study investigated causal associations using Mendelian randomization while examining radiation therapy timing effects.
Methods
We conducted two-sample Mendelian randomization analysis using GWAS summary statistics for interleukin-7 levels and neuroblastoma with three SNPs as instrumental variables. Multiple MRmethods included inverse variance weighted (IVW), MR-Egger, weighted median, and mode approaches. Additionally, 1,007 neuroblastoma patients from SEER database (2000–2018) were analyzed comparing preoperative (n = 416) versus postoperative (n = 591) radiation therapy using propensity score matching and Cox regression models.
Results
Mendelian randomization revealed significant positive causal association between elevated interleukin-7 levels and increased neuroblastoma risk. The IVW method showed higher interleukin-7 levels associated with 3.6-fold increased odds (OR = 3.585, 95% CI: 1.216–10.575, p = 0.021). In clinical analysis, preoperative radiation demonstrated superior survival outcomes with 27% mortality reduction (HR = 0.73, 95% CI: 0.55–0.97, p = 0.031). Subgroup analysis revealed significant racial differences, with White patients deriving greatest benefit from preoperative radiation (HR = 0.57, 95% CI: 0.42–0.78, p < 0.001).
Results
This study provides evidence for causal relationship between interleukin-7 levels and neuroblastoma risk, suggesting inflammatory pathways’ role in pathogenesis.
Keywords: Neuroblastoma, Mendelian randomization, Interleukin-7, Radiation therapy, Causal inference
Introduction
Among pediatric extracranial solid malignancies, neuroblastoma stands out as a particularly prevalent entity, comprising 8–10% of all cancer cases in children. Arising from undifferentiated neural crest cells within the sympathetic nervous system, this tumor demonstrates remarkable diversity in its clinical manifestations, biological characteristics, and prognostic outcomes [1–3]. While significant progress has been achieved through integrated treatment modalities encompassing surgical intervention, systemic chemotherapy, and radiation therapy, considerable uncertainty persists regarding the optimal sequencing of these therapeutic approaches, with particular controversy surrounding the appropriate timing of radiotherapy relative to surgical procedures.
Pediatric oncology research has increasingly focused on the intricate relationships between inflammatory signaling pathways and neoplastic development. Interleukin-7 (IL-7) has emerged as a cytokine of particular interest, given its central importance in lymphoid cell maturation, immune system homeostasis, and surveillance against malignant transformation. This cytokine serves as an essential growth and survival signal for both T and B lymphocyte lineages, orchestrating adaptive immune mechanisms while preserving the integrity of peripheral T cell compartments [4–6]. Multiple cancer types have been associated with increased systemic IL-7 concentrations, implying a possible role in facilitating tumor escape from immune recognition and promoting disease advancement. Nevertheless, establishing definitive causal links between IL-7 levels and neuroblastoma susceptibility has proven challenging, primarily due to methodological constraints of conventional epidemiological investigations, which are susceptible to confounding variables, temporal ambiguity in cause-effect relationships, and patient selection artifacts.
Contemporary management strategies for neuroblastoma employ comprehensive therapeutic regimens that integrate surgical tumor removal, systemic chemotherapy, radiotherapy, and targeted immunological interventions for patients with high-risk disease characteristics. Radiotherapy occupies a central position in these protocols, serving primarily to establishlocal disease control and minimize the likelihood of tumor recurrence [7–9]. Despite its established importance, the question of optimal radiotherapy scheduling in relation to surgical timing continues to generate significant clinical controversy. Conventional therapeutic approaches have traditionally emphasized post-surgical radiation delivery to eliminate residual malignant cells and address microscopic disease extension beyond surgical margins. However, accumulating clinical evidence indicates that pre-surgical radiation administration may confer unique therapeutic benefits, such as reducing tumor burden prior to resection, enhancing the feasibility of complete surgical excision, and potentially improving long-term patient survival rates. The absence of definitive randomized clinical trials specifically examining radiotherapy timing considerations in neuroblastoma management has led to substantial heterogeneity in treatment approaches and clinical recommendations among different medical centers.
Mendelian randomization has revolutionized causal inference in medical research by leveraging genetic variants as instrumental variables to overcome the limitations of conventional observational studies. This approach exploits the random allocation of genetic variants at conception, effectively mimicking the randomization process inherent in controlled experiments while avoiding ethical constraints associated with interventional studies in pediatric populations. By utilizing genome-wide association study (GWAS) data, Mendelian randomization enables robust assessment of causal relationships between biomarkers and disease outcomes while minimizing confounding effects and reverse causation bias [10, 11].
The integration of genetic epidemiological approaches with comprehensive clinical outcome analyses represents a powerful strategy for advancing our understanding of complex diseases like neuroblastoma. By combining Mendelian randomization methodology to establish causal relationships between inflammatory mediators and disease risk with detailed survival analyses of treatment timing effects, this study aims to provide novel insights into both the pathobiological mechanisms underlying neuroblastoma development and optimal therapeutic strategies for patient management. The findings from this dual-approach study may have significant implications for both risk stratification strategies and clinical decision-making in neuroblastoma management, potentially informing the development of personalized treatment protocols and improved patient outcomes.
Methods
Study design and conceptual framework
Our study utilized a robust dual-sample Mendelian randomization (2SMR) methodological approach to rigorously assess the potential causal association between systemic interleukin-7 concentrations and neuroblastoma susceptibility. The Mendelian randomization technique constitutes an influential epidemiological method that exploits spontaneously occurring genetic polymorphisms as proxy variables, successfully replicating the random allocation mechanisms characteristic of randomized controlled investigations. This analytical approach takes advantage of the fundamental concept that genetic variations are distributed randomly during fertilization, thus dramatically minimizing the impact of environmental confounders, temporal causality issues, and recruitment biases that commonly undermine conventional observational research designs. The dual-sample methodology employs distinct, non-overlapping datasets for exposure and outcome assessments, thereby amplifying analytical power while preserving the validity of causal inference principles.
Data sources and population characteristics
Summary statistics from genome-wide association studies (GWAS) examining circulating interleukin-7 concentrations were carefully obtained from extensive, thoroughly documented population studies incorporating detailed proteomic analyses of individuals with European genetic background. Quality assurance protocols were rigorously applied to these datasets, involving comprehensive individual-level data validation, systematic genetic variant screening, and advanced statistical corrections for population structure [12, 13]. To prevent potential interference with cytokine quantification, individuals presenting with active inflammatory diseases, autoimmune conditions, or receiving immunomodulatory treatments were systematically removed from analyses.
GWAS summary data for neuroblastoma were obtained through extensive meta-analytical approaches combining multiple large-scale case-control investigations and community-based cohort studies. These analyses incorporated clearly defined patient cohorts with neuroblastoma diagnoses confirmed through histopathological examination following internationally recognized classification protocols. Patient data spanned the complete spectrum of ages and disease severities, accompanied by comprehensive clinical documentation encompassing disease staging information, pathological classifications, and available molecular profiling data.
Control populations underwent careful demographic matching procedures considering age distribution, gender representation, and geographic location, with European ancestry verification performed using principal component analytical methods to prevent population structure confounding. Stringent case identification protocols maintained diagnostic precision through independent expert pathological evaluation and strict adherence to established diagnostic standards.
Instrumental variable strength assessment and validation
Rigorous evaluation of instrumental variable strength involved calculating multiple statistical measures including the proportion of phenotypic variance explained (R²) by each genetic variant and corresponding F-statistics to assess instrument strength. F-statistics exceeding 10 were required for inclusion to mitigate weak instrument bias, a critical assumption violation that can lead to biased causal estimates. Additionally, we computed conditional F-statistics for scenarios involving multiple correlated instruments and assessed the cumulative explanatory power of the entire instrumental variable set. Instrumental variables were further evaluated for potential pleiotropic effects through systematic database searches and literature reviews to identify known associations with potential confounding factors.
Comprehensive statistical analysis strategy
The inverse variance weighted (IVW) method served as the primary analytical approach for causal effect estimation, representing the gold standard for Mendelian randomization analysis under the assumption of valid instrumental variables. This method employs a weighted regression framework where individual SNP-outcome associations are regressed against corresponding SNP-exposure associations, with weights determined by the inverse variance of SNP-outcome estimates. The IVW approach provides optimal statistical efficiency when all instrumental variables satisfy the exclusion restriction criterion, yielding unbiased causal effect estimates. Results were expressed as odds ratios (OR) with corresponding 95% confidence intervals (CI), facilitating direct interpretation of effect magnitudes and clinical significance assessment.
Data source
This study utilized data from the Surveillance, Epidemiology, and End Results (SEER) database, which is a population-based cancer registry sponsored by the National Cancer Institute. The SEER database collects cancer incidence, treatment, and survival data from approximately 28% of the U.S. population across multiple geographic regions [14–16]. We extracted data on patients diagnosed with neuroblastoma between 2000 and 2018 using the International Classification of Diseases for Oncology, Third Edition (ICD-O-3) histology codes.
Study population
Patients with histologically confirmed neuroblastoma who received both surgical resection and radiation therapy were included in this analysis. We excluded patients with unknown treatment sequence, those who received neither surgery nor radiation, patients with incomplete follow-up information, and those with missing data on critical variables including age, sex, race, tumor grade, or TNM staging. The final cohort consisted of 1,007 patients, of whom 591 received radiation after surgery and 416 received radiation before surgery.
Variables and outcomes
The primary exposure variable was the sequence of radiation therapy relative to surgery (preoperative versus postoperative). Demographic variables included age at diagnosis (< 65 years, ≥ 65 years), sex, race (White, Black, Other, Unknown), and marital status (Single, Married, Divorced, Separated, Widowed, Unknown). Clinical variables included tumor size (continuous), tumor grade (I-IV, Unknown), TNM staging (T1-T4, TX; N0-N1; M0-M1), lymph node positivity (LNP), and receipt of chemotherapy (Yes/No).The primary outcome was overall survival (OS), defined as the time from diagnosis to death from any cause or last follow-up. Survival time was measured in months, with a maximum follow-up period of 556 time units.
Statistical analysis
Baseline characteristics between the preoperative and postoperative radiation groups were compared using Student’s t-test for continuous variables and chi-square or Fisher’s exact tests for categorical variables. Standardized mean differences (SMD) were calculated to assess the balance of covariates between the two groups.To address potential selection bias inherent in observational studies, we employed propensity score matching (PSM) techniques. Propensity scores were calculated using logistic regression with treatment sequence as the dependent variable and all measured baseline characteristics as covariates. Patients were matched using a 1:1 nearest-neighbor matching algorithm with a caliper width of 0.2 times the standard deviation of the logit of the propensity score.The association between treatment sequence and overall survival was assessed using Cox proportional hazards regression models [17, 18]. Three models with increasing levels of adjustment were constructed: Model 1 (unadjusted), Model 2 (adjusted for age, race, marital status, sex, grade, and TNM staging), and Model 3 (further adjusted for lymph node positivity and tumor size).Subgroup analyses were performed to identify potential effect modifiers by stratifying according to age, race, sex, tumor grade, TNM staging, and marital status. Interaction terms were tested to assess potential heterogeneity in treatment effects across subgroups.Kaplan-Meier survival curves were generated to visualize the differences in survival probability between treatment groups, and the log-rank test was used to determine statistical significance. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated.All statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY) and R version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria). A two-sided p-value < 0.05 was considered statistically significant.
Results
Patient characteristics before PSM: comparison between pre- and post-surgery radiation groups
This data presents a comparison of baseline characteristics between patients who received radiation after surgery (n = 591) and those who received radiation before surgery (n = 416) for neuroblastoma treatment. Several significant differences exist between the two groups. Patients in the pre-surgery radiation group had significantly larger mean tumor sizes (3.19 vs. 2.93 cm, p = 0.035) and a higher proportion of younger patients (< 65 years: 93.99% vs. 88.32%, p = 0.002). Males were more represented in the pre-surgery radiation group (57.21% vs. 51.10%, p = 0.05). There were notable differences in grade distribution (p = 0.002), with the pre-surgery group having a higher proportion of Grade IV tumors (18.99% vs. 11.68%), as well as significant differences in T staging (p = 0.005), with the pre-surgery group having fewer T2 tumors. More patients in the pre-surgery radiation group received chemotherapy (36.30% vs. 29.10%, p = 0.016). Major differences were observed in marital status distribution (p < 0.001), with the pre-surgery group having more married patients (45.43% vs. 28.60%) and fewer single patients (44.23% vs. 57.02%). No significant differences were found in lymph node positivity (LNP), race, N stage, or M stage between the two groups. These baseline characteristic differences highlight the need for propensity score matching (PSM) to minimize selection bias when comparing outcomes between these treatment approaches (Table 1).
Table 1.
Patient characteristics before PSM: comparison between pre- and post-surgery radiation groups
| Variable | Total (n = 1007) | Radiation after surgery (n = 591) | Radiation before surgery (n = 416) | Statistic | P | SMD |
|---|---|---|---|---|---|---|
| LNP, Mean ± SD | 0.17 ± 2.48 | 0.11 ± 0.32 | 0.26 ± 3.83 | t=– 0.963 | 0.336 | 0.040 |
| Tumor size, Mean ± SD | 3.06 ± 1.03 | 2.93 ± 1.05 | 3.19 ± 1.02 | t=– 1.964 | 0.035 | 0.631 |
| Age, n (%) | χ²=9.260 | 0.002 | ||||
| <65 | 913 (90.67) | 522 (88.32) | 391 (93.99) | 0.238 | ||
| ≥65 | 94 (9.33) | 69 (11.68) | 25 (6.01) | -0.238 | ||
| Race, n (%) | – | 0.140 | ||||
| Black | 104 (10.33) | 62 (10.49) | 42 (10.10) | -0.013 | ||
| Other | 136 (13.51) | 72 (12.18) | 64 (15.38) | 0.089 | ||
| Unknown | 5 (0.5) | 5 (0.85) | 0 (0.00) | -0.121 | ||
| White | 762 (75.67) | 452 (76.48) | 310 (74.52) | -0.045 | ||
| Sex, n (%) | χ²=3.667 | 0.05 | ||||
| Female | 467 (46.38) | 289 (48.90) | 178 (42.79) | -0.124 | ||
| Male | 540 (53.62) | 302 (51.10) | 238 (57.21) | 0.124 | ||
| Grade, n (%) | χ²=16.622 | 0.002 | ||||
| Grade I | 36 (3.57) | 28 (4.74) | 8 (1.92) | – 0.205 | ||
| Grade II | 88 (8.74) | 48 (8.12) | 40 (9.62) | 0.051 | ||
| Grade III | 163 (16.19) | 95 (16.07) | 68 (16.35) | 0.007 | ||
| Grade IV | 148 (14.7) | 69 (11.68) | 79 (18.99) | 0.187 | ||
| Unknown | 572 (56.8) | 351 (59.39) | 221 (53.12) | – 0.126 | ||
| T, n (%) | χ²=8.196 | 0.005 | ||||
| T1 | 21 (2.09) | 10 (1.69) | 11 (2.64) | 0.059 | ||
| T2 | 948 (94.14) | 566 (95.77) | 382 (91.83) | – 0.144 | ||
| T3 | 16 (1.59) | 6 (1.02) | 10 (2.40) | 0.091 | ||
| T4 | 10 (0.99) | 3 (0.51) | 7 (1.68) | 0.091 | ||
| TX | 12 (1.19) | 6 (1.02) | 6 (1.44) | 0.036 | ||
| N, n (%) | χ²=0.094 | 0.760 | ||||
| N0 | 1004 (99.7) | 590 (99.83) | 414 (99.52) | – 0.045 | ||
| N1 | 3 (0.3) | 1 (0.17) | 2 (0.48) | 0.045 | ||
| M, n (%) | χ²=0.102 | 0.749 | ||||
| M0 | 995 (98.81) | 585 (98.98) | 410 (98.56) | – 0.036 | ||
| M1 | 12 (1.19) | 6 (1.02) | 6 (1.44) | 0.036 | ||
| Chemotherapy, n (%) | χ²=5.801 | 0.016 | ||||
| No | 684 (67.92) | 419 (70.90) | 265 (63.70) | – 0.150 | ||
| Yes | 323 (32.08) | 172 (29.10) | 151(36.30) | 0.150 | ||
| Marital, n (%) | χ²=35.859 | < 0.001 | ||||
| Divorced | 43 (4.27) | 24 (4.06) | 19 (4.57) | 0.024 | ||
| Married | 358 (35.55) | 169 (28.60) | 189 (45.43) | 0.338 | ||
| Separated | 2 (0.2) | 1 (0.17) | 1 (0.24) | 0.015 | ||
| Single | 521 (51.74) | 337 (57.02) | 184 (44.23) | – 0.258 | ||
| Unknown | 37 (3.67) | 30 (5.08) | 7 (1.68) | – 0.264 | ||
| Widowed | 46 (4.57) | 30 (5.08) | 16 (3.85) | – 0.064 |
Impact of Pre-Surgery versus Post-Surgery radiation on neuroblastoma outcomes
The data presents three statistical models comparing survival outcomes between neuroblastoma patients who received radiation before surgery versus those who received radiation after surgery. Across all three models, patients who received radiation before surgery consistently demonstrated better survival outcomes. In the unadjusted analysis (Model 1), pre-surgery radiation was associated with a 29% reduction in mortality risk compared to post-surgery radiation (HR = 0.71, 95% CI: 0.55–0.92, p = 0.010). After adjusting for demographic and clinical characteristics including age, race, marital status, sex, tumor grade, and TNM staging (Model 2), the survival benefit remained significant with a 26% reduction in mortality risk (HR = 0.74, 95% CI: 0.57–0.98, p = 0.032). In the fully adjusted model (Model 3), which additionally controlled for lymph node positivity and tumor size, pre-surgery radiation maintained a significant 27% reduction in mortality risk (HR = 0.73, 95% CI: 0.55–0.97, p = 0.031). These findings consistently suggest that administering radiation before surgical intervention may offer a survival advantage for neuroblastoma patients compared to the traditional approach of post-surgical radiation, with the benefit persisting even after accounting for multiple potential confounding variables(Table 2).
Table 2.
Impact of Pre-Surgery versus Post-Surgery radiation on neuroblastoma outcomes
| Variables | Model1 | Model2 | Model3 | |||
|---|---|---|---|---|---|---|
| HR (95%CI) | P | HR (95%CI) | P | HR (95%CI) | P | |
| Therapy | ||||||
| Radiation after surgery | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Radiation before surgery | 0.71 (0.55–0.92) | 0.010 | 0.74 (0.57–0.98) | 0.032 | 0.73 (0.55–0.97) | 0.031 |
HR Hazard Ratio, CI Confidence Interval
Model1: Crude
Model2: Adjust: Age, Race, Marital, Sex, Grade, T,M, N
Model3: Adjust: Age, Race, Marital, Sex, Grade, T,M, N, LNP, Tumor Size
Subgroup analysis of pre-surgery radiation effects in neuroblastoma patients
This data presents subgroup analyses exploring how the benefit of pre-surgery radiation versus post-surgery radiation varies across different patient populations with neuroblastoma. Overall, the protective effect of pre-surgery radiation was consistent across most subgroups, but with notable variations. Among patients younger than 65, pre-surgery radiation showed a significant 28% reduction in mortality risk (HR = 0.72, 95% CI: 0.55–0.93, p = 0.012), while the effect could not be reliably estimated in the older age group. Race demonstrated a significant interaction effect (p for interaction = 0.013), with White patients deriving the greatest benefit from pre-surgery radiation (HR = 0.57, 95% CI: 0.42–0.78, p < 0.001), while Black patients showed a non-significant trend toward harm (HR = 1.95, 95% CI: 0.85–4.48, p = 0.115). By sex, males showed a significant benefit from pre-surgery radiation (HR = 0.65, 95% CI: 0.46–0.93, p = 0.019), while females showed a non-significant trend in the same direction. Grade III tumors demonstrated the most pronounced benefit from pre-surgery radiation (HR = 0.34, 95% CI: 0.20–0.58, p < 0.001). T2 stage tumors (HR = 0.66, 95% CI: 0.51–0.86, p = 0.002), N0 stage (HR = 0.71, 95% CI: 0.55–0.92, p = 0.009), and M0 stage (HR = 0.70, 95% CI: 0.54–0.91, p = 0.008) all showed significant benefits from pre-surgery radiation. Regarding marital status, single patients derived significant benefit from pre-surgery radiation (HR = 0.69, 95% CI: 0.51–0.94, p = 0.018), while other marital categories showed varying non-significanttrends. These findings suggest that while pre-surgery radiation generally offers survival benefits across neuroblastoma patients, the magnitude of benefit varies significantly by patient characteristics, particularly race and tumor grade(Table 3).
Table 3.
Factors associated with clinical outcomes in colorectal cancer: a regression analysis
| Variables | No | Yes | HR (95%CI) | P | P for interaction |
|---|---|---|---|---|---|
| Age | 0.989 | ||||
| <65 | 67/185 | 409/728 | 0.72 (0.55–0.93) | 0.012 | |
| ≥65 | 0/1 | 17/93 | 9158538.89 (0.00–Inf) | 0.998 | |
| Race | 0.013 | ||||
| Black | 8/26 | 38/78 | 1.95 (0.85–4.48) | 0.115 | |
| Other | 7/29 | 50/107 | 0.81 (0.36–1.80) | 0.605 | |
| Unknown | 3/3 | 2/2 | 3.31 (0.30–36.71) | 0.329 | |
| White | 49/128 | 336/634 | 0.57 (0.42–0.78) | < 0.001 | |
| Sex | 0.580 | ||||
| Female | 31/74 | 215/393 | 0.78 (0.53–1.14) | 0.193 | |
| Male | 36/112 | 211/428 | 0.65 (0.46–0.93) | 0.019 | |
| Grade | 0.234 | ||||
| Grade I | 0/1 | 22/35 | |||
| Grade II | 0/5 | 41/83 | 3370905.82 (0.00–Inf) | 0.997 | |
| Grade III | 19/23 | 64/140 | 0.34 (0.20–0.58) | < 0.001 | |
| Grade IV | 2/21 | 46/127 | 0.85 (0.20–3.56) | 0.829 | |
| Unknown | 46/136 | 253/436 | 0.89 (0.65–1.22) | 0.458 | |
| T | 0.313 | ||||
| T1 | 1/3 | 14/18 | 1.07 (0.13–8.71) | 0.953 | |
| T2 | 65/171 | 394/777 | 0.66 (0.51–0.86) | 0.002 | |
| T3 | 1/3 | 9/13 | 0.49 (0.05–4.41) | 0.522 | |
| T4 | 0/2 | 4/8 | |||
| TX | 0/7 | 5/5 | 1618926998.27 (0.00–Inf) | 0.999 | |
| N | 0.988 | ||||
| N0 | 67/184 | 425/820 | 0.71 (0.55–0.92) | 0.009 | |
| N1 | 0/2 | 1/1 | 1615474875.20 (0.00–Inf) | 1.000 | |
| M | 0.991 | ||||
| M0 | 67/178 | 423/817 | 0.70 (0.54–0.91) | 0.008 | |
| M1 | 0/8 | 3/4 | |||
| Marital | 0.644 | ||||
| Divorced | 0/5 | 17/38 | 3443512.91 (0.00–Inf) | 0.999 | |
| Married | 9/67 | 122/291 | 0.68 (0.34–1.37) | 0.281 | |
| Separated | 0/0 | 1/2 | |||
| Single | 49/90 | 266/431 | 0.69 (0.51–0.94) | 0.018 | |
| Unknown | 8/13 | 16/24 | 1.14 (0.48–2.73) | 0.765 | |
| Widowed | 1/11 | 4/35 | 0.21 (0.01–3.39) | 0.273 |
HR Hazard Ratio, CI Confidence Interval
Survival analysis of radiation timing in neuroblastoma treatment
The Kaplan-Meier survival curve illustrates the significant difference in survival outcomes between neuroblastoma patients receiving radiation before versus after surgery. Patients who received radiation after surgery (red line) demonstrated consistently higher survival probability compared to those who received radiation before surgery (blue line) throughout the follow-up period of 556 time units. The statistical analysis confirms this difference is highly significant (p = 6.4e-3). Specifically, patients receiving radiation after surgery had a 43% lower risk of mortality compared to those receiving radiation before surgery (HR = 1.43, 95% CI: 1.10–1.85). The survival probability curves begin to diverge noticeably around the 100-time unit mark and continue to separate through the remainder of the observation period. By the end of follow-up, the radiation-after-surgery group maintained a small but persistent survival advantage. The confidence intervals (shaded areas) indicate greater uncertainty in the radiation-before-surgery group, possibly due to smaller sample size or higher variability in outcomes. These findings contrast with the previous Cox regression analyses, suggesting that when visualized as time-to-event data without adjustment for confounding variables, post-surgical radiation appears to offer superior survival outcomes for neuroblastoma patients(Fig. 1).
Fig. 1.
Survival Analysis of Radiation Timing in Neuroblastoma Treatment. The Kaplan-Meier survival curve shows that neuroblastoma patients who received radiation after surgery (red line) had significantly better survival outcomes than those who received radiation before surgery (blue line) throughout the 556-time unit follow-up period. The difference is statistically significant (p = 6.4e–3), with patients receiving radiation after surgery having a 43% lower mortality risk (HR = 1.43, 95% CI: 1.10–1.85)
Mendelian randomization analysis results of Interleukin-7 levels and neuroblastoma association
This study employed multiple Mendelian randomization methods to analyze the association between interleukin-7 levels and neuroblastoma, based on 3 single nucleotide polymorphisms. The inverse variance weighted method revealed a significant positive association between interleukin-7 levels and neuroblastoma, with an odds ratio of 3.585 (95% confidence interval: 1.216–10.575, p = 0.021), suggesting that elevated interleukin-7 levels may increase the risk of neuroblastoma. However, the MR-Egger method showed extremely wide confidence intervals and non-significant p-values (p = 0.647), which may indicate potential horizontal pleiotropy issues requiring cautious interpretation of causal relationships. Other sensitivity analysis methods including simple mode, weighted median, and weighted mode did not reach statistical significance, with p-values of 0.433, 0.217, and 0.447, respectively. Nevertheless, except for the MR-Egger method, other approaches demonstrated relatively consistent positive association trends, which to some extent supports the reliability of the main analysis results (Fig. 2).
Fig. 2.
Mendelian Randomization Analysis of Interleukin-7 Levels and Neuroblastoma Association. Forest plot displaying odds ratios and 95% confidence intervals from multiple Mendelian randomization methods examining the causal relationship between interleukin-7 levels and neuroblastoma risk. Analysis based on 3 single nucleotide polymorphisms (SNPs). The inverse variance weighted (IVW) method showed a significant positive association (OR = 3.585, 95% CI: 1.216–10.575, p = 0.021), while sensitivity analyses including MR-Egger, simple mode, weighted median, and weighted mode did not reach statistical significance. OR, odds ratio; CI, confidence interval
Multi-Method Mendelian randomization sensitivity analysis for neuroblastoma risk
This four-panel figure presents comprehensive Mendelian randomization analysis results examining the association between interleukin-7 levels and neuroblastoma risk. Figure 3A shows a scatter plot comparing inverse variance weighted and MR-Egger methods, with SNP effect estimates clustering relatively consistently, suggesting result coherence. Figure 3B displays trend line comparisons across multiple MR methods including inverse variance weighted, MR-Egger, weighted median, and simple mode approaches, where the inverse variance weighted method (blue line) demonstrates a clear positive trend while other method trend lines remain relatively flat. Figure 3C and D present forest plots showing individual SNP effect estimates and 95% confidence intervals respectively. Panel C shows relatively consistent effect estimates across individual SNPs with moderately narrow confidence intervals.
Fig. 3.
Multi-Method Mendelian Randomization Sensitivity Analysis of Interleukin-7 Levels and Neuroblastoma Risk. Four-panel analysis showing: A Scatter plot comparing inverse variance weighted and MR-Egger methods; B Trend line comparison across multiple MR approaches including inverse variance weighted, MR-Egger, weighted median, and simple mode; C, D Forest plots displaying individual SNP effects and overall effect estimates with 95% confidence intervals
Discussion
Pediatric oncology continues to face substantial obstacles in neuroblastoma management, even with progressive developments in comprehensive therapeutic strategies. This investigation targets a crucial evidence gap concerning the ideal coordination between radiotherapy and surgical procedures in neuroblastoma care, carrying significant ramifications for therapeutic practice and subsequent research endeavors [19–21].
The longstanding controversy surrounding the timing of radiation delivery whether preceding or following surgical resection has persisted across various solid malignancies, with each strategy presenting distinct theoretical merits. Pre-surgical radiation administration potentially confers advantages through tumor volume reduction, which may enhance the achievability of comprehensive resection while minimizing the likelihood of malignant cell dispersal during operative procedures. Furthermore, the improved tissue oxygenation present in treatment-naive tumors could augment radiosensitive responses. In contrast, post-surgical radiation permits targeted treatment of residual malignancy detected during resection, facilitates comprehensive histopathological evaluation for radiation treatment planning, and prevents delays in primary surgical management.
The SEER registry serves as an invaluable platform for investigating this therapeutic dilemma, providing access to extensive population-derived data that facilitates examination of actual clinical treatment patterns and patient outcomes [22–24]. Nevertheless, observational investigations inherently face selection bias challenges, where therapeutic choices may be influenced by unaccounted clinical variables. While our implementation of propensity score matching methodology aimed to reduce such biases, complete elimination of residual confounding remains unattainable.
The contrast between our crude Kaplan-Meier survival analysis and multivariable Cox proportional hazards modeling underscores the essential role of controlling for baseline patient heterogeneity in observational data interpretation. The substantial variations in patient demographics and clinical features between treatment cohorts encompassing tumor dimensions, age patterns, gender distribution, histological grading, staging parameters, systemic therapy utilization, and sociodemographic factors illustrate the intricate, multidimensional nature of clinical therapeutic decision-making. Such variations may represent evolving therapeutic philosophies, institutional treatment preferences, or individual patient considerations that guide treatment sequence selection.
The observed variability in therapeutic efficacy across distinct patient populations indicates that individualized treatment sequencing strategies may be appropriate. The differential benefits demonstrated among various demographic and clinical categories correspond with the increasing understanding that neuroblastoma represents a heterogeneous group of malignancies characterized by diverse biological characteristics and therapeutic responsiveness [25–27]. The racial interaction phenomenon identified in our analysis is particularly noteworthy and merits additional exploration into possible biological or socioeconomic determinants that may impact treatment effectiveness.
These results must be considered within the framework of multiple study constraints. The SEER registry provides limited detailed information regarding radiation dosimetry, treatment volumes, delivery techniques, surgical resection margins, and molecular tumor characteristics that could affect therapeutic outcomes. Moreover, the database fails to document treatment-associated adverse effects, which represent critical factors in pediatric patient populations with extended post-treatment life expectancies. Additionally, temporal changes in therapeutic approaches throughout the study interval may introduce confounding into the observed relationships.
Despite these constraints, this investigation provides meaningful contributions to the continuing discussion regarding optimal multimodal therapy sequencing in neuroblastoma management. Our results underscore the significance of incorporating patient-specific and tumor-related factors into treatment planning decisions and emphasize the necessity for prospective randomized investigations to establish definitive evidence for optimal treatment ordering.
Subsequent research efforts should prioritize the incorporation of comprehensive molecular and genetic characterization to enhance patient risk stratification and therapeutic selection processes. Furthermore, investigations evaluating patient-reported outcomes and long-term treatment-related morbidities associated with alternative treatment sequences would offer valuable supplementary data to inform clinical decision-making processes. The creation of advanced imaging-based predictive markers for preoperative radiation response could additionally assist in identifying patient populations most likely to derive benefit from this therapeutic approach.
Limitation
A major limitation of this study is that the SEER database lacks detailed treatment information, including radiation dosimetry, treatment volumes, delivery techniques, surgical resection margins, and molecular tumor characteristics, all of which could influence treatment outcomes but cannot be adequately considered in the analysis.
Conclusion
Using Mendelian randomization methodology, we observed a statistically significant association between genetically predicted interleukin-7 levels and neuroblastoma risk, though this finding should be interpreted cautiously given the limited number of instrumental variables employed. Analysis of SEER registry data revealed complex relationships between radiation therapy timing and survival outcomes, with adjusted and unadjusted analyses yielding conflicting results.
Author contributions
G.R. conceived the study design, conducted all Mendelian randomization analyses and statistical computations, and drafted the manuscript. Y.C. contributed to data collection and validation, assisted with clinical data interpretation, and participated in manuscript preparation. Q.G. supervised the project, provided critical guidance, reviewed and revised the manuscript, and approved the final version for submission.
Funding
Study on the immune function of NK cells in patients with hepatoblastoma and its relationship with prognosis (AHWJ2021a025).
Data availability
Inquiries regarding additional details should be directed to the corresponding author.
Declarations
Ethics approval and consent to participate
Not available.
Consent for publication
The final version of this manuscript has received unanimous review and approval from all authors.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Fu Y, Zhang N, Cheng J, Qin X, Zhou X, Du X, Wang Y, Wang J, Zhang D. Identification of novel biomarkers and prognostic model for neuroblastoma using Mendelian randomization and transcriptomic analysis. Discov Oncol. 2025;16(1):587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Nhungo CJ, Gilbert DG, Nzowa B, Felix P, Msangi N, Mmbando T, Mtaturu G, Nyongole O, Mkony CA. Giant adrenal neuroblastoma in adults: surgical management and comprehensive review. Clin Case Rep. 2025;13(4):e70453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wen X, Lu X, Xu Y, Liu J, Yang J. Recurrent solitary pleural metastasis originating from neuroblastoma as demonstrated by 123I-MIBG scan. Clin Nucl Med 2025. [DOI] [PubMed]
- 4.Hamouda AEI, Filtjens J, Brabants E, Kancheva D, Debraekeleer A, Brughmans J, Jacobs L, Bardet PMR, Knetemann E, Lefesvre P, et al. Intratumoral delivery of lipid nanoparticle-formulated mRNA encoding IL-21, IL-7, and 4-1BBL induces systemic anti-tumor immunity. Nat Commun. 2024;15(1):10635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lee KJ, Choi D, Tae N, Song HW, Kang YW, Lee M, Moon D, Oh Y, Park S, Kim JH, et al. IL-7-primed bystander CD8 tumor-infiltrating lymphocytes optimize the antitumor efficacy of T cell engager immunotherapy. Cell Rep Med. 2024;5(5):101567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Shin HS, Kim H, Kwon SG, Lee H, Lee JO, Kim YS. Tumor cells ectopically expressing the membrane-bound form of IL-7 develop an antitumor immune response efficiently in a colon carcinoma model. Mol Cells. 2025;48(2):100175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Akhtaruzzaman R, Awad K, Koster A, Varanasi V, Brotto M, Adnan A. A mechanical model for lateral and axial impacts and quantification of effect on viability of SHSY5Y neuroblastoma cells. Sci Rep. 2025;15(1):19353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dholaria H, Naranjo A, Zhang FF, London WB, Diller L, Shamberger RC, Barnewolt CE, Schmidt ML, Maris JM, Cohn SL et al. Long-Term Outcome of Young Infants With Suspected Neuroblastoma following Observation as Primary Therapy: A Report From the Children’s Oncology Group. Pediatr Blood Cancer 2025:e31829. [DOI] [PMC free article] [PubMed]
- 9.Pandher R, Xue C, Gamble LD, Milazzo G, Di Giacomo S, Murray J, Cheung L, Ferrucci F, Palombo M, Purgato S, et al. The cell-permeable iron chelator M606 inhibits MYCN-driven neuroblastoma via an E2F3-mediated response. Proc Natl Acad Sci U S A. 2025;122(23):e2420011122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wu H, Ding M, Zhu J, Mao S, Tang X, Fang S, Liu L, Pan Q, Yue C. Causal relationship between sex Hormone-Binding Globulin and risk of neuroblastoma: A bidirectional Two-Sample Mendelian randomization study. Cancer Epidemiol Biomarkers Prev. 2024;33(6):846–53. [DOI] [PubMed] [Google Scholar]
- 11.Zhang Z, Li D, Xie F, Zhang H. The causal relationship between gut microbiota and neuroblastoma: a bidirectional Mendelian randomization analysis and meta-analysis. Microbiol Spectr. 2024;12(4):e0365623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li Y, Song X, Huang Y, Zhou S, Zhong L. Genetic associations of plasma metabolites with immune cells in hyperthyroidism revealed by Mendelian randomization and GWAS-sc-eQTLs xQTLbiolinks analysis. Sci Rep. 2025;15(1):1377. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ye Z, Liu H, Shi L, Ke X. Investigating causal associations between pork intake and multiple sclerosis using Mendelian randomization: insights from large-scale GWAS data in European populations. Nutr Health 2025:2601060241308918. [DOI] [PubMed]
- 14.Bai Y, Lei N, Zhang P, Yang Q, Feng F, Zhao Y. Risk factors, prognostic factors and nomograms for distant metastasis in colorectal neuroendocrine neoplasms: a SEER-based study. Transl Cancer Res. 2025;14(3):1576–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ellaithy I, Elshiekh H, Elshennawy S, Elshenawy S, Al-Shaikh B, Ellaithy A. Sepsis as a cause of death among elderly cancer patients: an updated SEER database analysis 2000–2021. Ann Med Surg (Lond). 2025;87(4):1838–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Rong T, Ai C, Yang T, Wu Q, Zhang M. Clinical features and prognostic nomogram development for cancer-specific death in patients with dual primary lung cancer: a population-based study from SEER database. J Cardiothorac Surg. 2025;20(1):190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Liu K, Chang CE, Lloyd S, Tao R, Nguyen T, Zhang ZF, Hashibe M. The risk of type 2 diabetes among older asian, native Hawaiian and Pacific Islander colorectal cancer survivors:a population-based study using SEER-Medicare database. Cancer Epidemiol Biomarkers Prev 2025. [DOI] [PMC free article] [PubMed]
- 18.Loap P, Kirova Y, Dendale R. Radiotherapy of localized orbital mantle cell lymphoma: a SEER database analysis of long-term outcomes. Strahlenther Onkol 2025. [DOI] [PubMed]
- 19.Castelli S, Thorwarth A, van Schewick C, Wendt A, Astrahantseff K, Szymansky A, Lodrini M, Veldhoen S, Gratopp A, Mall MA et al. Management of Busulfan-Induced Lung Injury in Pediatric Patients with High-Risk Neuroblastoma. J Clin Med 2024, 13(19). [DOI] [PMC free article] [PubMed]
- 20.de Las Heras BM, Rubio-Aparicio PM, Rubio-San-Simon A, Moreno L, Mazorra P, Almaraz RL, Lopez ML, Guill JB, Segura V, Bermudez M et al. Management and outcome of children with high-risk neuroblastoma: insights from the Spanish society of pediatric hematology and oncology (SEHOP) neuroblastoma group on refractory and relapse/progressive disease. Clin Transl Oncol 2025. [DOI] [PMC free article] [PubMed]
- 21.Patel A, Im E, Kresak J, Olgaard E, Blatt JE, Lobo BC, Chapurin N. Olfactory neuroblastoma with divergent differentiation: contemporary management of unusual pathology and literature review. Ear Nose Throat J 2024:1455613241299684. [DOI] [PubMed]
- 22.Yan Z, Wu Y, Chen Y, Xu J, Zhang X, Yin Q. A clinical prediction model for distant metastases of pediatric neuroblastoma: an analysis based on the SEER database. Front Pediatr. 2024;12:1417818. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhao X, Xu Z, Feng X. Clinical characteristics and prognoses in pediatric neuroblastoma with bone or liver metastasis: data from the SEER 2010–2019. BMC Pediatr. 2024;24(1):162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zhuo X, Xia L, Tang W, He W. A practical nomogram and risk stratification system for predicting survival outcomes in neuroblastoma patients: a SEER population-based study. J Cancer Res Clin Oncol. 2023;149(13):12285–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Aktas TC, Kizmazoglu D, Aktas S, Gokbayrak OE, Serinan E, Erol A, Altun Z, Yuan H, Olgun HN. Identification of ALK mutation in neuroblastoma on the point of molecular heterogeneity. Technol Cancer Res Treat. 2023;22:15330338231211138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Cai L, DeBerardinis RJ, Xie Y, Minna JD, Xiao G. A comparative study of neuroendocrine heterogeneity in small cell lung cancer and neuroblastoma. Mol Cancer Res. 2023;21(8):795–807. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Finlay JB, Ireland AS, Hawgood SB, Reyes T, Ko T, Olsen RR, Abi Hachem R, Jang DW, Bell D, Chan JM, et al. Olfactory neuroblastoma mimics molecular heterogeneity and lineage trajectories of small-cell lung cancer. Cancer Cell. 2024;42(6):1086–105. e1013. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Inquiries regarding additional details should be directed to the corresponding author.