Risk Stratification Tools for Thyroid Cancer: A Systematic Review of Models Combining Ultrasound, Cytology, and...

Area of Science:

• Endocrinology
• Oncology
• Diagnostic Imaging

Background:

• Rising thyroid cancer incidence necessitates improved diagnostic tools.
• Current risk stratification systems (RSTs) integrate clinical, ultrasonographic, and cytological features.
• Evaluating the performance and clinical utility of multimodal RSTs is crucial for surgical decision-making.

Purpose of the Study:

• To systematically review and critically evaluate risk stratification tools (RSTs) for thyroid nodules.
• To assess the diagnostic performance and clinical utility of multimodal RSTs in supporting surgical decisions.
• To identify the most effective and generalizable RSTs for thyroid nodule assessment.

Main Methods:

• Systematic review of PubMed, Embase, and Cochrane databases up to 04/13/2026.
• Inclusion of studies evaluating multivariable risk prediction models for adult thyroid nodules, incorporating clinical, ultrasound, and cytology data, validated by histology.
• Risk of bias assessment using the PROBAST+AI tool.

Main Results:

• Five distinct RSTs were identified: Thyroid Nodule App (TNAPP), McGill Thyroid Nodule Score (MTNS), CUT Score, Memorial Sloan Kettering Cancer Centre (MSKCC) nomogram, and Thyroid Prediction Score (TiPS).
• TiPS exhibited the highest sensitivity (96.2%) and specificity (97.5%) with AUC >0.9; CUT Score also performed strongly (AUC >0.9).
• TNAPP underperformed; MTNS and MSKCC lacked robust validation; most models had limited generalizability due to single-center, retrospective data.

Conclusions:

• Multimodal RSTs can enhance thyroid nodule risk stratification, especially for indeterminate cytology cases.
• Methodological limitations and lack of external validation currently hinder widespread clinical utility.
• Prospective evaluation in diverse populations is needed to establish the most effective and generalizable RSTs; current use should supplement clinical judgment.