AI and Cancer Research ── Brain ── 2026-10-31
Brain tumours: treatment, trials and AI
- Abstract
Introduction: Craniopharyngiomas are rare sellar/suprasellar tumors with long-term postoperative morbidity, including permanent diabetes insipidus (DI) and tumor recurrence. Existing clinical tools offer limited individualized prognostication. We introduce a pilot artificial intelligence (AI) platform to hypothesize postoperative outcomes in craniopharyngioma surgery.
Methods: Patients with histologically confirmed craniopharyngioma who underwent surgical resection between 2008 and 2023 at a single institution were retrospectively analyzed. Clinical, radiographic, and intraoperative features (e.g., tumor dimensions, volume, intraventricular extension, surgical approach, and extent of resection) were modeled using machine learning algorithms with a unified ensemble pipeline to predict permanent DI and tumor recurrence. Explainable AI interpretation using Shapley Additive Explanations (SHAP) quantified individual feature-level contributions, which were integrated into a FastAPI-powered web application for interactive prediction.
Results: 59 patients were used in a stratified cross-validation model training, of whom permanent DI occurred in 65.5% and tumor recurrence was recorded in 28.8%. The blended Ridge-Catboost ensemble achieved the highest performance for permanent DI prediction (AUC: 0.77±0.05, F1: 0.81±0.04), and a blended Naïve Bayes-Catboost ensemble was chosen for recurrence prediction (AUC: 0.78±0.12, F1: 0.67±0.15). SHAP-based feature ranking revealed that pre-operative BMI, age, and endocrine or visual dysfunction were the strongest predictors of DI and recurrence. Surgical approach and postoperative BMI changes further modulated DI risk. The app can be accessed at tinyurl.com/craniopharyngiomas.
Conclusion: In this small, single-institution cohort, ensemble learning and model tuning helped stabilize model accuracy and discriminative power. This exploratory AI prototype for prognostication might eventually support data-driven risk assessment in skull base oncology.
Journal IF-equivalent: 0.1 (OpenAlex 2-year mean citedness, value as of 2026-10-29, retrieved 2026-10-31; not the official Clarivate IF)Reference: Babel S, Patel A, Parasher A, Tabor M, Agazzi S, van Loveren H, et al. Craniopharyngioma.AI: An Explainable Ensemble Learning Framework for Personalized Craniopharyngioma Outcome Prediction. J Neurol Surg B Skull Base. 2026 Oct 29 [Epub ahead of print]. doi:10.1055/a-2992-8671.Checked: Abstract only - Abstract
PURPOSE: This study aimed to develop and validate a classification model for preoperative prediction of WHO grading of meningioma based on spatial habitat radiomics that can be explained by Ki-67 and p53 biomarkers. METHODS: A retrospective cohort of 400 patients (261 WHO grade I; 139 grade II/III) from a single institution was analyzed. Contrast-enhancing tumors were partitioned into three biologically distinct subregions (hypo-, intermediate-, and hyper-enhancement) using a 3D Simple Linear Iterative Clustering algorithm. Radiomics and clinical-radiological features were selected based on dual associations with Ki-67 ≥ 5% and p53 ≥ 5%. Subsequently, three classification models for preoperative prediction of WHO grading were constructed: clinical-radiological model (C-model), whole-tumor radiomics model (CRW-model), and subregion-based radiomics model (CRS-model). Performance was evaluated using AUC in training (n = 280) and test (n = 120) cohorts. RESULTS: Spatial habitat radiomics demonstrated biomarker-specific predictive power: hyper-enhancement subregions achieved the highest Ki-67 prediction AUC (training: 0.81; test: 0.70), while hypo-enhancement subregions excelled in p53 prediction (training: 0.93; test: 0.57). The CRS-model, achieved the highest AUC in both the training cohort (0.89 vs. 0.80 for CRW-model and 0.76 for C-model) and the test cohort (0.76 vs. 0.68 and 0.67, respectively). In the training cohort, the CRS-model significantly outperformed both the CRW-model and C-model after Bonferroni correction (adjusted p = 0.000450 and 7.06 × 10⁻⁷, respectively). In the test cohort, the CRS-model showed numerically higher AUCs, but the differences were not statistically significant after correction (adjusted p = 0.0891 and 0.0918, respectively). CONCLUSIONS: Spatial habitat radiomics may help characterize intratumoral microenvironmental heterogeneity associated with Ki-67 and p53 expression, while providing a biologically interpretable framework for preoperative WHO grading of meningiomas.
Journal IF-equivalent: 3.8 (OpenAlex 2-year mean citedness, value as of 2026-10-02, retrieved 2026-10-03; not the official Clarivate IF)Reference: Xie H, Peng X, Chen X, Yu Y, Yi Z, Meng W, et al. Ki-67/p53-associated angiogenic heterogeneity for preoperative WHO grading of meningiomas: a spatial habitat radiomics study. BMC Med Imaging. 2026 Aug 3;26(1):562. doi:10.1186/s12880-026-02632-5. PMID: 42958086.Checked: Abstract only