Scholarlink

Objective <p>This study aims to analyze clinical and laboratory data from patients to identify indicators associated with lymph node metastasis (LNM) in papillary thyroid carcinoma (PTC). Furthermore, we aim to develop a nomogram and a web-based calculator for predicting the risk of LNM.</p> Methods <p>We conducted a retrospective analysis of 1,134 patients who underwent PTC resection between January 2018 and July 2023. The enrolled patients were randomly divided into a modeling set and a validation set in a 7:3 ratio. Lasso-logistic regression was employed to identify independent predictors of LNM in PTC. The dose-response relationship between independent influencing factors and the risk of PTC occurrence was assessed using restricted cubic spline (RCS). Finally, we evaluated the diagnostic value and clinical net benefit of each indicator using receiver operating characteristic (ROC) curves and decision curve analysis (DCA).</p> Results <p>Lasso-logistic regression analysis revealed that ultrasound TI-RADS classification, nodule diameter, nodule number, gender, age, APOB, and CEA are independent predictors of LNM in patients with PTC. RCS analysis indicates that nodule diameter, age, CEA, and APOB levels exhibit both linear or nonlinear relationships with the risk of LNM in PTC patients. A nomogram and a web-based calculator were developed to predict the risk of LNM in PTC based on these characteristic variables. The full model's area under the receiver operating characteristic (AUROC) curve is 0.753 (95% CI: 0.716-0.790), with a mean cross-validation AUROC of 0.738 (95% CI: 0.710-0.768) and a pooled cross-validation AUROC of 0.734 (95% CI: 0.696-0.772). The AUROC value for the validation set is 0.733 (95% CI: 0.682-0.779), the prediction model demonstrates moderate discriminative ability. Calibration assessment using calibration plots, slope, intercept, and the Hosmer–Lemeshow test indicated acceptable agreement between predicted and observed risks. DCA demonstrated higher net benefit across clinically relevant thresholds, and the clinical impact curve (CIC) showed good agreement between predicted high-risk patients and observed events. Furthermore, the model has been transformed into a freely accessible web-based calculator (https://ley120.shinyapps.io/Lymph_Node_Metastasis_in_PTC/).</p> Conclusion <p>TI-RADS classification, Nodule diameter, Nodule number, Gender, Age, APOB, and CEA have moderate diagnostic value in evaluating lymph node metastasis of PTC. This study developed a nomogram with acceptable discrimination for predicting LNM risk in PTC, which may assist clinicians in risk stratification when used alongside other clinical assessments. This model serves as a supplementary tool for risk estimation rather than a definitive diagnostic standard.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development of a clinically applicable prediction tool for lymph node metastasis in papillary thyroid carcinoma using lasso-logistic regression

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月17日

查看全文
  • Wei Zhang,
  • Jichao Zhu,
  • Ying Zhang,
  • Xu Zhang,
  • Ying Dong,
  • Xiao Yu,
  • Yidan Zhang,
  • Kun Wang,
  • Anquan Shang

摘要

Objective

This study aims to analyze clinical and laboratory data from patients to identify indicators associated with lymph node metastasis (LNM) in papillary thyroid carcinoma (PTC). Furthermore, we aim to develop a nomogram and a web-based calculator for predicting the risk of LNM.

Methods

We conducted a retrospective analysis of 1,134 patients who underwent PTC resection between January 2018 and July 2023. The enrolled patients were randomly divided into a modeling set and a validation set in a 7:3 ratio. Lasso-logistic regression was employed to identify independent predictors of LNM in PTC. The dose-response relationship between independent influencing factors and the risk of PTC occurrence was assessed using restricted cubic spline (RCS). Finally, we evaluated the diagnostic value and clinical net benefit of each indicator using receiver operating characteristic (ROC) curves and decision curve analysis (DCA).

Results

Lasso-logistic regression analysis revealed that ultrasound TI-RADS classification, nodule diameter, nodule number, gender, age, APOB, and CEA are independent predictors of LNM in patients with PTC. RCS analysis indicates that nodule diameter, age, CEA, and APOB levels exhibit both linear or nonlinear relationships with the risk of LNM in PTC patients. A nomogram and a web-based calculator were developed to predict the risk of LNM in PTC based on these characteristic variables. The full model's area under the receiver operating characteristic (AUROC) curve is 0.753 (95% CI: 0.716-0.790), with a mean cross-validation AUROC of 0.738 (95% CI: 0.710-0.768) and a pooled cross-validation AUROC of 0.734 (95% CI: 0.696-0.772). The AUROC value for the validation set is 0.733 (95% CI: 0.682-0.779), the prediction model demonstrates moderate discriminative ability. Calibration assessment using calibration plots, slope, intercept, and the Hosmer–Lemeshow test indicated acceptable agreement between predicted and observed risks. DCA demonstrated higher net benefit across clinically relevant thresholds, and the clinical impact curve (CIC) showed good agreement between predicted high-risk patients and observed events. Furthermore, the model has been transformed into a freely accessible web-based calculator (https://ley120.shinyapps.io/Lymph_Node_Metastasis_in_PTC/).

Conclusion

TI-RADS classification, Nodule diameter, Nodule number, Gender, Age, APOB, and CEA have moderate diagnostic value in evaluating lymph node metastasis of PTC. This study developed a nomogram with acceptable discrimination for predicting LNM risk in PTC, which may assist clinicians in risk stratification when used alongside other clinical assessments. This model serves as a supplementary tool for risk estimation rather than a definitive diagnostic standard.