Normative Growth Modeling of Cortical Thickness Identify Neuroanatomical Variability and Distinct Subtypes in Brainstem Tumor Patients

Abstract

AbstractBackgroundBrainstem tumors can cause structural brain changes, but the resulting heterogeneity within wholebrain structure is not well-studied. This study examines cortical thickness to identify patterns of structural alterations and explore underlying biological subtypes and their associations with clinical factors.Materials and MethodsThis study involved 124 pediatric brainstem tumor patients, aged 4-18 years. Cortical thickness was measured using CAT12 segmentation of 3D T1-weighted structural MRI. A normative model was established using data from 849 healthy children. Deviations in cortical thickness were estimated, and patients were classified into two subtypes using spectral clustering. Clinical statistical analyses were conducted with SPSS 26.0.ResultsThe normative model revealed significant heterogeneity in cortical thickness deviations, which correlated with tumor size and growth patterns. Focal tumors primarily caused negative deviations (t = 3.14, p = 0.02). There was a significant positive correlation between extreme positive deviations and tumor volume (r = 0.214, p = 0.010), and between extreme negative deviations and progression-free survival (r = 0.39, p = 0.008). Two subtypes were identified: Subtype 1, consisting of diffuse tumors with extreme positive deviations, and Subtype 2, consisting of focal tumors with extreme negative deviations. Subtype and tumor growth pattern significantly influenced duration (p < 0.01). The Kaplan-Meier survival curves for Subtype 1 and Subtype 2 demonstrated a significant difference in survival probabilities over time (p = 0.03).ConclusionOverall, this study identifies two major patterns of cortical thickness changes in brainstem tumor patients, enhancing our understanding of their relationship with cortical morphology. The findings suggest that cortical thickness alterations could serve as valuable biomarkers for predicting progression-free survival, which is crucial for clinical assessment and personalized treatment strategies. This research provides new insights into the physiological mechanisms by which brainstem tumors affect brain structure, supporting more precise clinical interventions and efficacy monitoring in the future.