EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Machine Learning–Driven Integration of Clinical and Multiomics Biomarkers for Risk Stratification in Primary Squamous Cell Carcinoma of the Thyroid

Huanxin Shi, Lina Huang, Ziyang Zeng

International Journal of Clinical Practice · 2026-01 · 2026 (1)

DOI: 10.1155/ijcp/3193727 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Clinical Practice, online erschienen am 12.08.2026. Verfasst von Huanxin Shi und 2 weiteren Autor:innen. Die Arbeit nennt 14 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objective

Primary squamous cell carcinoma of the thyroid (PSCCT) represents a rare, aggressive malignancy with insufficiently characterized molecular underpinnings and a lack of robust prognostic instruments. This investigation aimed to derive novel prognostic biomarkers and to develop a composite risk‐stratification model by fusing multiomics data with standard clinical parameters through advanced machine learning methodologies.

Methods

We performed a retrospective analysis on a cohort of 250 PSCCT patients sourced from the SEER registry (1975–2018). Clinically relevant variables—encompassing age, sex, tumor grade, SEER stage, and treatment received—were merged with transcriptome‐imputed biomarkers identified via artificial intelligence (AI)–powered bioinformatics analysis. A machine learning–enhanced Cox proportional hazards model was implemented to ascertain determinants of overall survival (OS) and disease‐specific survival (DSS). The interpretability of the model was increased through SHapley Additive exPlanations (SHAP) to quantify and rank feature importance.

Results

Univariate evaluation indicated significant associations of tumor grade, extrathyroidal extension, SEER stage, and surgical approach with both OS and DSS (all p values < 0.05). Multivariate Cox regression analysis validated tumor grade and SEER stage as independent prognostic variables for these survival outcomes. Leveraging machine learning for feature selection, we subsequently delineated a suite of transcriptomic biomarkers that augment traditional clinical predictors. The final AI‐integrated prognostic model showed improved accuracy over established staging systems in forecasting 3‐year survival (concordance index: 0.79 compared to 0.68) and efficiently segregated patients into discrete risk categories.

Conclusion

We have established and validated an interpretable, machine learning–based prognostic framework for PSCCT that synergistically integrates clinical and multiomics biomarkers. This methodology advances individualized outcome forecasting and can inform clinical decision‐making for this uncommon but lethal cancer. While tumor grade endures as a pivotal prognostic factor, AI‐refined biomarker panels provide a promising strategy to optimize risk assessment and personalize therapeutic interventions in PSCCT.

Zitieren

APA

Shi, H., Huang, L., & Zeng, Z. (2026). Machine Learning–Driven Integration of Clinical and Multiomics Biomarkers for Risk Stratification in Primary Squamous Cell Carcinoma of the Thyroid. International Journal of Clinical Practice, 2026(1). https://doi.org/10.1155/ijcp/3193727

Vancouver

Shi H, Huang L, Zeng Z. Machine Learning–Driven Integration of Clinical and Multiomics Biomarkers for Risk Stratification in Primary Squamous Cell Carcinoma of the Thyroid. International Journal of Clinical Practice. 2026;2026(1). doi:10.1155/ijcp/3193727

BibTeX
@article{Shi2026,
  author = {Shi, Huanxin and Huang, Lina and Zeng, Ziyang},
  title = {Machine Learning–Driven Integration of Clinical and Multiomics Biomarkers for Risk Stratification in Primary Squamous Cell Carcinoma of the Thyroid},
  journal = {International Journal of Clinical Practice},
  year = {2026},
  volume = {2026},
  number = {1},
  doi = {10.1155/ijcp/3193727}
}
Kostenlos starten →