Construction and Temporal Validation of a Machine Learning Model for Predicting Clinically Significant (Grade ≥2) Acute Radiation Enteritis After Radiotherapy for Cervical Cancer
Entwicklung und zeitliche Validierung eines Machine-Learning-Modells zur Vorhersage einer klinisch signifikanten (Grad ≥2) akuten Strahlenenteritis nach Radiotherapie bei Zervixkarzinom
Journal of Clinical Medicine · 01.11.2026 · 15 (21) · 8501
DOI: 10.3390/jcm15218501 ↗Auf einen Blick
Bei 363 Patient:innen mit Zervixkarzinom sagte ein logistisches Regressionsmodell eine akute Strahlenenteritis Grad ≥2 am ausgewogensten voraus. Die Validierung blieb jedoch nur mäßig.
Fragestellung
Welche Modelle sagen eine klinisch signifikante akute Strahlenenteritis (CTCAE Grad ≥2) nach pelviner Radiotherapie bei Zervixkarzinom voraus?
Methode
Retrospektive Entwicklung und zeitliche Validierung von 6 Vorhersagemodellen bei 363 Patient:innen; Training, interner Test und temporale externe Validierung. Feature-Auswahl und Tuning nur im Trainingssatz.
Ergebnisse
Das LR-Modell erreichte AUCs von 0,855 (95 %-KI 0,792–0,910) im Training, 0,830 (95 %-KI 0,721–0,918) im internen Test und 0,753 (95 %-KI 0,621–0,864) in der zeitlichen Validierung. Sensitivität dort 0,40 (95 %-KI 0,176–0,619).
Einschränkungen
Prospektive, multizentrische Validierung erforderlich; Sensitivität des LR-Modells in der zeitlichen Validierung nur 0,40.
Schlussfolgerung der Autor:innen
Dosimetrische Parameter kombiniert mit inflammatorisch-nutritiven Indizes zeigen eine vielversprechende Trennschärfe für ARE Grad ≥2. Das LR-Modell ist einfach und interpretierbar, braucht aber weitere Validierung.
Kernaussagen
- LR am ausgewogensten, AUC 0,753 in der zeitlichen Validierung
- RF und LightGBM mit AUC 1,000 im Training, aber schwächer in der Validierung
- Rectal_Dmean, Bowel_V30 und SII waren die wichtigsten Prädiktoren
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
Kurz gefasst
Fachartikel in der Zeitschrift Journal of Clinical Medicine, online erschienen am 01.11.2026. Verfasst von Wenqing Hu und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Oncology Radiotherapy, The First Affiliated Hospital of Anhui Medical University, Hefei 230022, China. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Anhui Province Scientific Research Preparation Pro gram Project.
Abstract (Originalsprache)
Purpose
To develop and temporally validate prediction models for clinically significant acute radiation enteritis, defined as Common Terminology Criteria for Adverse Events (CTCAE) grade ≥2 ARE and hereafter abbreviated SARE to remain consistent with the published literature on this endpoint, in patients with cervical cancer undergoing pelvic radiotherapy.
Methods
A total of 363 patients diagnosed with cervical cancer were retrospectively included and received concurrent chemoradiation at the First Affiliated Hospital of Anhui Medical University. The training set (n = 203, 70%) and the internal test set (n = 87, 30%) were recruited from January 2020 to January 2024; an independent temporal validation cohort (n = 73; 72 with complete outcome and prediction data) was recruited from January 2024 to December 2025. Feature selection (Boruta plus random-forest recursive feature elimination) and hyperparameter tuning were performed exclusively within the training set. Six models were constructed—logistic regression (LR), a multilayer perceptron (MLP), a support vector machine (SVM), random forest (RF), XGBoost and LightGBM—and SHapley Additive exPlanations (SHAP) was used to interpret the final model.
Results
Twelve optimal predictors were selected. The LR model demonstrated the most balanced performance, achieving AUCs of 0.855 (95% CI 0.792–0.910) in the training set, 0.830 (95% CI 0.721–0.918) in the internal test set, and 0.753 (95% CI 0.621–0.864) in the temporal external validation cohort. RF and LightGBM fitted the training data perfectly (AUC 1.000) but fell to 0.762 and 0.757 in validation, and XGBoost, the SVM and the MLP were lower (0.661, 0.599 and 0.616). In temporal validation LR had the lowest Brier score (0.176) but a sensitivity of only 0.40 (95% CI 0.176–0.619) at the 0.5 threshold. Decision curve analysis showed positive net benefit for LR across threshold probabilities of 0.05–0.80. SHAP ranked Rectal_Dmean, Bowel_V30 and SII as the leading contributors.
Conclusions
Dosimetric parameters combined with composite inflammatory–nutritional indices show promising discrimination for grade ≥2 ARE. The LR model provides a simple, interpretable tool for pre-radiotherapy risk stratification, but prospective, multicenter validation is required before routine clinical implementation.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Oncology Radiotherapy, The First Affiliated Hospital of Anhui Medical University, Hefei 230022, China.
- Department of Radiation Therapy Center, Beijing Obstetrics and Gynecology Hospital, Capital Medical University, Beijing Maternal and Child Health Care Hospital, Beijing 100006, China.
- Department of Gynecological Oncology, Beijing Obstetrics and Gynecology Hospital, Capital Medical University, Beijing Maternal and Child Health Care Hospital, Beijing 100006, China.
ORCID-Profile: Wenqing Hu, Ming Wang
Förderung
- Anhui Province Scientific Research Preparation Pro gram Project
Zitieren
APA
Hu, W., Song, X., Wang, Y., Zhao, J., Cheng, L., Wang, M., & Lv, Y. (2026). Construction and Temporal Validation of a Machine Learning Model for Predicting Clinically Significant (Grade ≥2) Acute Radiation Enteritis After Radiotherapy for Cervical Cancer. Journal of Clinical Medicine, 15(21), 8501. https://doi.org/10.3390/jcm15218501
Vancouver
Hu W, Song X, Wang Y, Zhao J, Cheng L, Wang M, et al. Construction and Temporal Validation of a Machine Learning Model for Predicting Clinically Significant (Grade ≥2) Acute Radiation Enteritis After Radiotherapy for Cervical Cancer. Journal of Clinical Medicine. 2026;15(21):8501. doi:10.3390/jcm15218501
BibTeX
@article{Hu2026,
author = {Hu, Wenqing and Song, Xiaofeng and Wang, Yuanjing and Zhao, Jiqi and Cheng, Luyao and Wang, Ming and Lv, Yin},
title = {Construction and Temporal Validation of a Machine Learning Model for Predicting Clinically Significant (Grade ≥2) Acute Radiation Enteritis After Radiotherapy for Cervical Cancer},
journal = {Journal of Clinical Medicine},
year = {2026},
volume = {15},
number = {21},
pages = {8501},
doi = {10.3390/jcm15218501}
}
EUVIMED