JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Tailoring nonsurgical therapy for elderly patients with head and neck squamous cell carcinoma: A deep learning-based approach

  • 0Heilongjiang University of Chinese Medicine, Harbin, China.
Medicine +

|

September 17, 2024

|

September 17, 2024

View abstract on PubMed

Summary

This summary is machine-generated.

Deep learning models can personalize chemotherapy for elderly head and neck cancer patients. The Self-Normalizing Balanced individual treatment effect model best identified who benefits from chemoradiotherapy, improving survival predictions.

Area Of Science

  • Oncology
  • Artificial Intelligence
  • Biostatistics

Background

  • Elderly patients with head and neck squamous cell carcinoma (HNSCC) often have limited treatment options, especially if not candidates for surgery.
  • Personalized treatment selection is crucial for optimizing outcomes in this demographic.

Purpose Of The Study

  • To evaluate deep learning models for personalized chemotherapy selection in elderly HNSCC patients.
  • To quantify the impact of patient characteristics on treatment efficacy.
  • To compare model recommendations against standard guidelines.

Main Methods

  • Utilized inverse probability treatment weighting (IPTW) to address bias.
  • Employed mixed-effects regression to analyze patient characteristics' impact on treatment choice.
  • Compared treatment outcomes based on model recommendations versus actual treatments received, using overall survival as the primary endpoint.

Main Results

  • The Self-Normalizing Balanced individual treatment effect for survival data model demonstrated superior performance in treatment recommendation.
  • This model showed a significant improvement in survival metrics (HR 0.74, RD 9.92%, RMST difference 16.42 months) compared to other models and guidelines.
  • No survival benefit was observed for chemoradiotherapy in patients not recommended for it by the model.

Conclusions

  • The Self-Normalizing Balanced individual treatment effect model effectively identifies elderly HNSCC patients who can benefit from chemoradiotherapy.
  • This AI approach offers personalized survival predictions and treatment recommendations, potentially improving patient outcomes.
  • Further clinical validation is necessary for widespread implementation.