Machine Learning: A Multicenter Study on Predicting Lateral Lymph Node Metastasis in cN0 Papillary Thyroid Carcinoma
Authors' conclusion
ML, especially RF, reliably predicts lateral LNM in cN0 PTC patients, outperforming traditional nomograms.
Study context
- Population
- cN0 papillary thyroid carcinoma (PTC) patients diagnosed via fine-needle aspiration biopsy; training/testing sets from hospital A (7:3), validation set from hospital B1,616 hospital A + 243 validation
- Comparison
- Machine learning models vs traditional nomograms for predicting lateral lymph node metastasis (LNM)
- Outcome
- Prediction of ipsilateral lateral and level II, III and IV LNM
Methodological detailsmulticenter, machine learning, external validation set, 10-fold cross-validation, Web-based calculator for each of the 4 dependent variables
- Multicenter
- Machine learning
- External validation set
- 10-fold cross-validation
- Web-based calculator for each of the 4 dependent variables
Performance
As reported in the abstract
Ipsilateral lateral LNM
| Test or model | Accuracy (testing/validation) | AUC (testing/validation) | Specificity (testing/validation) | Sensitivity (testing/validation) |
|---|---|---|---|---|
| Random forest (RF) | 0.773/0.728 | 0.858/0.799 | 0.984/0.935 | 0.757/0.807 |
Random forest (RF)
Accuracy (testing/validation)0.773/0.728
AUC (testing/validation)0.858/0.799
Specificity (testing/validation)0.984/0.935
Sensitivity (testing/validation)0.757/0.807
Further findings
Level II, III and IV LNM
- Similar results for the other dependent variables, RF models relying on distinct but overlapping feature sets
Original publication
Machine Learning: A Multicenter Study on Predicting Lateral Lymph Node Metastasis in cN0 Papillary Thyroid Carcinoma. J Clin Endocrinol Metab. Published 8 February 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →