JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Nomograms versus artificial intelligence platforms: which one can better predict sentinel node positivity in melanoma patients?

  • 0Surgical Oncology.
Melanoma Research +

|

June 4, 2025

|

June 4, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

Artificial intelligence (AI) platforms show promise in predicting sentinel lymph node biopsy (SNB) positivity for melanoma patients, outperforming traditional nomograms in a recent study. Further AI development could enhance melanoma treatment decisions.

Area Of Science

  • Oncology
  • Medical Informatics
  • Predictive Analytics

Background

  • Nomograms are standard tools in oncology for clinical decision-making, including sentinel lymph node biopsy (SNB) for melanoma.
  • Artificial intelligence (AI) is emerging as a powerful tool for medical predictions.

Purpose Of The Study

  • To compare the predictive accuracy of nomograms and AI platforms for SNB positivity in melanoma patients.
  • To evaluate the real-world performance of established nomograms and AI tools.

Main Methods

  • Retrospective analysis of 62 melanoma patients who underwent SNB (2020-2024).
  • Utilized three open-access nomograms and three public AI platforms to predict SNB positivity.
  • Assessed predictive accuracy using clinical and pathological data, including logistic regression and ROC curves.

Main Results

  • No concordance was found among nomograms or AI platforms (P < 0.001).
  • The Memorial Sloan Kettering Cancer Center (MSKCC) nomogram and ChatGPT showed statistical significance for SNB positivity (P=0.04 and P=0.02, respectively).
  • ChatGPT predictions yielded an Area Under the Curve (AUC) of 0.702, improved to 0.715 when integrated with MSKCC predictions.

Conclusions

  • AI platforms, particularly ChatGPT, demonstrated superior predictive performance for SNB positivity compared to nomograms in this cohort.
  • Future enhancements in AI platforms may improve nomogram validation and lead to more accurate predictive models for melanoma management.

Keywords: