Multiple Machine Learning Algorithms for Predicting Postoperative Residual Tumor in Pituitary Neuroendocrine Tumors.
- 1Department of Neurosurgery, Xiangya Hospital Central South University, Changsha, China.
Abstract
Aims
To develop and validate an interpretable machine learning (ML) model for predicting radiologically confirmed postoperative residual tumor (RT) in pituitary neuroendocrine tumors (PitNETs).Methods
We retrospectively analyzed 584 surgically treated patients with PitNETs from a single center, using a random 7:3 training-test split and internal cross-validation. The primary endpoint was radiologically confirmed postoperative RT, assessed using MRI obtained within 48 h after surgery and longitudinal follow-up MRI. Candidate variables were selected using LASSO regression, with nine ML algorithms developed. Discrimination, calibration, clinical utility, interobserver agreement, and SHAP-based interpretability were assessed.Results
Postoperative RT occurred in 72 patients (12.3%). Sensitivity analysis showed high agreement between 48-h MRI and follow-up confirmation (96.4%; Cohen's kappa = 0.837). Interobserver agreement was high for retrosellar invasion (kappa = 0.825), internal carotid artery invasion (kappa = 0.960), and Knosp grade (quadratic weighted kappa = 0.994). Eight predictors were retained: maximum sagittal diameter, maximum coronal diameter, age, Knosp grade, internal carotid artery invasion, sphenoid sinus pneumatization, tumor consistency, and retrosellar invasion. In the internal test set, the random forest model achieved high discrimination (AUC = 0.945; accuracy = 0.938), with performance comparable to LightGBM and XGBoost in DeLong testing. LightGBM had the lowest Brier score.Conclusion
Random forest was selected for the web tool after internal evaluation, with discrimination comparable to LightGBM, XGBoost, and SVM. Its generalizability to other centers and future patient cohorts remains unproven because evaluation used only a single-center random split and internal cross-validation. External and temporal validation with calibration assessment is needed before clinical implementation.
Related
- From histopathology to molecular taxonomy: predicting pituitary neuroendocrine tumors invasiveness
- Autophagy in Pituitary Neuroendocrine Tumors (PitNETs): Molecular Mechanisms, Secretory Homeostasis, Therapeutic Resistance, and Translational Perspectives
- Cell-type-specific localization of huntingtin-associated protein 1 in neuroendocrine and neurosecretory systems of the mouse hypothalamic–pituitary axis
- Genomic characterization of aggressiveness in pituitary neuroendocrine tumors
- When Thyrotropin Meets Growth Hormone: The Clinical Challenges of a Mixed GH/TSH-Secreting PIT1-Lineage Pituitary Neuroendocrine Tumour
BibTeX
@article{li2026multiple,
title = {{Multiple Machine Learning Algorithms for Predicting Postoperative Residual Tumor in Pituitary Neuroendocrine Tumors.}},
author = {Mingrui Li and Chenyang Zhao and Guodong Tang and Yang Li and Qing Liu},
journal = {CNS neuroscience \& therapeutics},
year = {2026},
month = nov,
doi = {10.1002/cns.71244},
url = {https://europepmc.org/article/MED/42972894},
note = {CC BY},
}
RIS
TY - JOUR
TI - Multiple Machine Learning Algorithms for Predicting Postoperative Residual Tumor in Pituitary Neuroendocrine Tumors.
AU - Mingrui Li
AU - Chenyang Zhao
AU - Guodong Tang
AU - Yang Li
AU - Qing Liu
PY - 2026
DA - 2026/11/01
JO - CNS neuroscience & therapeutics
DO - 10.1002/cns.71244
UR - https://europepmc.org/article/MED/42972894
AB - Aims To develop and validate an interpretable machine learning (ML) model for predicting radiologically confirmed postoperative residual tumor (RT) in pituitary neuroendocrine tumors (PitNETs).Methods We retrospectively analyzed 584 surgically treated patients with PitNETs from a single center, using a random 7:3 training-test split and internal cross-validation. The primary endpoint was radiologically confirmed postoperative RT, assessed using MRI obtained within 48 h after surgery and longitudinal follow-up MRI. Candidate variables were selected using LASSO regression, with nine ML algorithms developed. Discrimination, calibration, clinical utility, interobserver agreement, and SHAP-based interpretability were assessed.Results Postoperative RT occurred in 72 patients (12.3%). Sensitivity analysis showed high agreement between 48-h MRI and follow-up confirmation (96.4%; Cohen's kappa = 0.837). Interobserver agreement was high for retrosellar invasion (kappa = 0.825), internal carotid artery invasion (kappa = 0.960), and Knosp grade (quadratic weighted kappa = 0.994). Eight predictors were retained: maximum sagittal diameter, maximum coronal diameter, age, Knosp grade, internal carotid artery invasion, sphenoid sinus pneumatization, tumor consistency, and retrosellar invasion. In the internal test set, the random forest model achieved high discrimination (AUC = 0.945; accuracy = 0.938), with performance comparable to LightGBM and XGBoost in DeLong testing. LightGBM had the lowest Brier score.Conclusion Random forest was selected for the web tool after internal evaluation, with discrimination comparable to LightGBM, XGBoost, and SVM. Its generalizability to other centers and future patient cohorts remains unproven because evaluation used only a single-center random split and internal cross-validation. External and temporal validation with calibration assessment is needed before clinical implementation.
SN - 1755-5930
SN - 1755-5949
ER -