Developing Machine Learning‐Based Prognostic Models for Nonmelanoma Skin Cancer Patients With Comorbid Diabetes: A Retrospective Cohort Study
Entwicklung von auf Machine Learning basierenden prognostischen Modellen für Patient:innen mit Nicht-Melanom-Hautkrebs und Komorbidität Diabetes: retrospektive Kohortenstudie
Cancer Medicine · 2026-11 · 15 (11)
DOI: 10.1002/cam4.72050 ↗Auf einen Blick
Bei 1.325 Patient:innen mit Nicht-Melanom-Hautkrebs und Typ-2-Diabetes wurden Risikofaktoren für das Überleben analysiert. Das RSF-Modell zeigte die beste Vorhersageleistung.
Fragestellung
Welche Risikofaktoren beeinflussen das Überleben von NMSC-Patient:innen mit T2DM, und wie gut sagen verschiedene Modelle die Prognose voraus?
Methode
Retrospektive Kohortenstudie mit Daten aus Hong Kong; Vergleich von CoxPH, LASSO-Cox, RSF und Survival Tree; SHAP-Analyse; Risiko-Score mit AutoScore-Survival.
Ergebnisse
Als ungünstige Prädiktoren wurden höheres Alter bei Krebsdiagnose, männliches Geschlecht, Rauchen, längere T2DM-Dauer, erhöhtes Kreatinin und HbA1c sowie niedrigere Nüchternglukose identifiziert. Der AutoScore-Survival-Score erreichte im Testset einen C-Index von 0,712; AUC 0,677, 0,697 und 0,729 nach 1, 3 und 5 Jahren.
Schlussfolgerung der Autor:innen
Älteres Alter bei Krebsdiagnose, männliches Geschlecht, längere Diabetesdauer, Rauchen, erhöhtes Kreatinin und HbA1c sowie niedrigere Nüchternglukose waren ungünstige Prognosefaktoren. Das RSF-Modell war am genauesten; der Risiko-Score war klinisch interpretierbar.
Kernaussagen
- 1.325 NMSC-Patient:innen mit T2DM wurden analysiert.
- RSF zeigte die höchste Vorhersageleistung.
- AutoScore-Survival: C-Index 0,712 im Testset.
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
Kurz gefasst
Kohortenstudie in der Zeitschrift Cancer Medicine, online erschienen am 09.11.2026. Verfasst von Junjie Huang und 8 weiteren Autor:innen. Beteiligte Einrichtung: The Jockey Club School of Public Health and Primary Care, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong SAR, China. Hauptthemen laut MEDLINE: Diabetes Mellitus, Type 2, Machine Learning, Non-Melanoma Skin Neoplasms. Die Arbeit nennt 66 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Themen: Künstliche Intelligenz in der Pflege, Onkologische Pflege, Diabetes mellitus
Abstract (Originalsprache)
Background
Nonmelanoma skin cancer (NMSC) is one of the most common malignancies worldwide, with increasing incidence. Type 2 diabetes mellitus (T2DM) is a prevalent comorbidity that may adversely affect cancer prognosis through shared metabolic and inflammatory pathways. This study aims to identify risk factors influencing the survival of NMSC patients with T2DM, construct and evaluate the performance of different predictive models, and develop a comprehensive risk score to accurately predict the survival probability of these patients.
Methods
We collected data from the Hong Kong Hospital Authority Data Collaboration Laboratory, including 1325 NMSC patients with T2DM between 2000 and 2020. Cox proportional hazards (CoxPH), least absolute shrinkage and selection operator‐Cox model (LASSO‐Cox), random survival forest (RSF), and survival tree were employed to construct prediction models. SHapley Additive exPlanations (SHAP) analysis was conducted to identify key risk factors and attribute model outputs. A risk score system was developed using the AutoScore‐Survival package.
Results
We identified seven significant predictors for poorer overall survival among patients with T2DM, including higher age at cancer diagnosis, male sex, smoking, longer duration of T2DM, elevated creatinine and glycated hemoglobin (HbA1c), and lower fasting glucose. The RSF model showed the highest predictive performance among the evaluated machine learning models. The tuned AutoScore‐Survival risk score system demonstrated good predictive performance (concordance index (C‐index) = 0.712 in the test set), and was based on the following variables: age at cancer diagnosis, duration of T2DM, creatinine, insulin treatment, stroke, and HbA1c. The areas under the curve (AUCs) at 1, 3, and 5 years in the test set were 0.677, 0.697, and 0.729, respectively.
Conclusion
Older age at cancer diagnosis, male sex, longer diabetes duration, smoking, elevated creatinine and HbA1c, and lower fasting glucose were identified as adverse prognostic factors for NMSC patients with T2DM. The RSF model outperformed other algorithms in predictive accuracy. The proposed risk score offers a clinically interpretable and effective means of predicting survival among NMSC patients with T2DM.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- The Jockey Club School of Public Health and Primary Care, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong SAR, China.
- Centre for Health Education and Health Promotion, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong SAR, China.
- Clinical Research Center & Big Data Center, The Seventh Affiliated Hospital, Sun Yat-Sen University, Shenzhen, Guangdong, China.
- Information Technology and Health Informatics Division, Hospital Authority, Hong Kong SAR, China.
- Department of Computer Science and Engineering, Faculty of Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China.
ORCID-Profile: Junjie Huang, Martin C S Wong
Zitieren
APA
Huang, J., Zhong, C. C., Li, Z., Yang, Z., Jiang, Y., Yuan, J., Poon, J., Dou, Q., & Wong, M. C. S. (2026). Developing Machine Learning‐Based Prognostic Models for Nonmelanoma Skin Cancer Patients With Comorbid Diabetes: A Retrospective Cohort Study. Cancer Medicine, 15(11). https://doi.org/10.1002/cam4.72050
Vancouver
Huang J, Zhong CC, Li Z, Yang Z, Jiang Y, Yuan J, et al. Developing Machine Learning‐Based Prognostic Models for Nonmelanoma Skin Cancer Patients With Comorbid Diabetes: A Retrospective Cohort Study. Cancer Medicine. 2026;15(11). doi:10.1002/cam4.72050
BibTeX
@article{Huang2026,
author = {Huang, Junjie and Zhong, Claire Chenwen and Li, Zhaojun and Yang, Zehuan and Jiang, Yu and Yuan, Jinqiu and Poon, Jonathan and Dou, Qi and Wong, Martin C S},
title = {Developing Machine Learning‐Based Prognostic Models for Nonmelanoma Skin Cancer Patients With Comorbid Diabetes: A Retrospective Cohort Study},
journal = {Cancer Medicine},
year = {2026},
volume = {15},
number = {11},
doi = {10.1002/cam4.72050}
}
EUVIMED