JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Tumor volume features predict survival outcomes for patients diagnosed with diffuse intrinsic pontine glioma

  • 0Center for Genetic Medicine Research, Children's National Hospital, Washington, District of Columbia, USA.
Neuro-Oncology Advances +

|

October 22, 2024

|

October 22, 2024

View abstract on PubMed

Summary

This summary is machine-generated.

Magnetic resonance imaging (MRI) features can predict survival in children with diffuse intrinsic pontine glioma (DIPG). Tumor volume and contrast enhancement at diagnosis and after radiation therapy are key indicators for overall survival (OS) outcomes.

Area Of Science

  • Pediatric Oncology
  • Neuro-oncology
  • Medical Imaging

Background

  • Diffuse intrinsic pontine glioma (DIPG) is a lethal pediatric brain tumor.
  • Magnetic resonance imaging (MRI) is crucial for DIPG diagnosis and monitoring treatment response.
  • Predicting overall survival (OS) in DIPG patients is essential for treatment planning.

Purpose Of The Study

  • To identify MRI-derived imaging features that predict overall survival (OS) in children with DIPG.
  • To evaluate the utility of tumor volume and contrast enhancement metrics at diagnosis and post-radiation therapy (RT).
  • To explore the application of machine learning (SVM) for outcome prediction in DIPG.

Main Methods

  • Retrospective analysis of contrast-enhanced T1- and FLAIR/T2-weighted MRI scans from 43 pediatric DIPG patients.
  • Evaluation of imaging features including 3D tumor volume (T<sub>wv</sub>), contrast-enhancing tumor core volume (T<sub>c</sub>), and their ratios (T<sub>c</sub>/T<sub>wv</sub>) at diagnosis and post-RT.
  • Application of Support Vector Machine (SVM) learning to identify feature combinations predicting OS outcomes (shorter or longer than 12 months).

Main Results

  • Features associated with poor OS included contrast-enhancing tumor at diagnosis and specific post-RT T<sub>c</sub>/T<sub>wv</sub> ratios (>15% and >20% ∆T<sub>c</sub>/T<sub>wv</sub>).
  • SVM analysis identified T<sub>c</sub>/T<sub>wv</sub> at diagnosis (74% accuracy) and ∆T<sub>c</sub>/T<sub>wv</sub> <2 months post-RT (75% accuracy) as significant predictors of poor survival.
  • Tumor imaging features evaluated within 4 months of RT demonstrated predictive value for differential OS outcomes.

Conclusions

  • MRI-based tumor imaging features at diagnosis and post-radiation therapy can predict overall survival (OS) in DIPG patients.
  • These findings support the integration of quantitative tumor volume analyses into clinical practice for DIPG.
  • The study highlights the potential for machine learning-based analyses to personalize DIPG treatment strategies based on tumor risk characteristics.

Keywords: