Prognostic Value of Metabolic Parameters and Textural Features in Pretreatment 18F-FDG PET/CT of Primary Lesions for Pediatric Patients with Neuroblastoma

Rationale and Objectives Our study evaluated the prognostic value of the metabolic parameters and textural features in pretreatment 18F-Fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) of primary lesions for pediatric patients with neuroblastoma. Materials and Methods In total, 107 pediatric patients with neuroblastoma who underwent pretreatment 18F-FDG PET/CT were retrospectively included and analyzed. All patients were diagnosed by pathology, and baseline characteristics and clinical data were collected. The four metabolic parameters and 43 textural features of 18F-FDG PET/CT of the primary lesions were measured. The prognostic significance of metabolic parameters and other clinical variables was assessed using Cox proportional hazards regression models. Differences in progression-free survival (PFS) and overall survival (OS) in relation to parameters were examined using the Kaplan–Meier method. Results During a median follow-up period of 34.3 months, 45 patients (42.1%) experienced tumor recurrence or progression, and 21 patients (19.6%) died of cancer. In univariate Cox regression analysis, age, location of disease, International Neuroblastoma Risk Group Staging System (INRGSS) stage M, neuron-specific enolase (NSE), lactate dehydrogenase (LDH), four positron emission tomography (PET) metabolic parameters, and 33 textural features were significant predictors of PFS. In multivariate analysis, INRGSS stage M (hazard ratio [HR] = 19.940, 95% confidence interval [CI] = 2.733–145.491, P = 0.003), skewness (＞0.173; PET first-order features; HR = 2.938, 95% CI = 1.389–6.215, P = 0.005), coarseness (＞0.003; neighborhood gray-tone difference matrix; HR = 0.253, 95% CI = 0.132–0.484, P ＜ 0.001), and variance (＞103.837; CT first-order gray histogram parameters; HR = 2.810, 95% CI = 1.160–6.807, P = 0.022) were independent predictors of PFS. In univariate Cox regression analysis, gender, INRGSS stage M, MYCN amplification, NSE, LDH, two PET metabolic parameters, and five textural features were significant predictors of OS. In multivariate analysis, INRGSS stage M (HR = 7.704, 95% CI = 1.031–57.576, P = 0.047), MYCN amplification (HR = 3.011, 95% CI = 1.164–7.786, P = 0.023), and metabolic tumor volume (＞138.788; HR = 3.930, 95% CI = 1.317–11.727, P = 0.014) were independent predictors of OS. Conclusion The metabolic parameters and textural features in pretreatment 18F-FDG PET/CT of primary lesions are predictive of survival in pediatric patients with neuroblastoma. Our study evaluated the prognostic value of the metabolic parameters and textural features in pretreatment 18F-Fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) of primary lesions for pediatric patients with neuroblastoma. In total, 107 pediatric patients with neuroblastoma who underwent pretreatment 18F-FDG PET/CT were retrospectively included and analyzed. All patients were diagnosed by pathology, and baseline characteristics and clinical data were collected. The four metabolic parameters and 43 textural features of 18F-FDG PET/CT of the primary lesions were measured. The prognostic significance of metabolic parameters and other clinical variables was assessed using Cox proportional hazards regression models. Differences in progression-free survival (PFS) and overall survival (OS) in relation to parameters were examined using the Kaplan–Meier method. During a median follow-up period of 34.3 months, 45 patients (42.1%) experienced tumor recurrence or progression, and 21 patients (19.6%) died of cancer. In univariate Cox regression analysis, age, location of disease, International Neuroblastoma Risk Group Staging System (INRGSS) stage M, neuron-specific enolase (NSE), lactate dehydrogenase (LDH), four positron emission tomography (PET) metabolic parameters, and 33 textural features were significant predictors of PFS. In multivariate analysis, INRGSS stage M (hazard ratio [HR] = 19.940, 95% confidence interval [CI] = 2.733–145.491, P = 0.003), skewness (＞0.173; PET first-order features; HR = 2.938, 95% CI = 1.389–6.215, P = 0.005), coarseness (＞0.003; neighborhood gray-tone difference matrix; HR = 0.253, 95% CI = 0.132–0.484, P ＜ 0.001), and variance (＞103.837; CT first-order gray histogram parameters; HR = 2.810, 95% CI = 1.160–6.807, P = 0.022) were independent predictors of PFS. In univariate Cox regression analysis, gender, INRGSS stage M, MYCN amplification, NSE, LDH, two PET metabolic parameters, and five textural features were significant predictors of OS. In multivariate analysis, INRGSS stage M (HR = 7.704, 95% CI = 1.031–57.576, P = 0.047), MYCN amplification (HR = 3.011, 95% CI = 1.164–7.786, P = 0.023), and metabolic tumor volume (＞138.788; HR = 3.930, 95% CI = 1.317–11.727, P = 0.014) were independent predictors of OS. The metabolic parameters and textural features in pretreatment 18F-FDG PET/CT of primary lesions are predictive of survival in pediatric patients with neuroblastoma.