Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
December 8, 2025Blood

Early detection of advanced-stage DLBCL using random forest: Uncovering large-scale demographic disparities and site-specific risk patterns

View Full Paper
Ask AI
Bookmark
Share

Authors

JSJunho SongACAsfand Yar CheemaJSJessica Santucci

Discussion

Loading...

Member takes

Overview

Machine learning finds site-specific risk factors in advanced stage DLBCL, suggesting enhanced screening methods.

Key Points

  • This research aims to enhance early detection of advanced-stage DLBCL using a Random Forest classifier.
  • Analyzed 126,774 DLBCL patients from the SEER database using advanced machine learning techniques.
  • Developed a Random Forest model with hyperparameter tuning using stratified-split data and class weights.
  • Evaluated model performance using metrics such as AUC, accuracy, sensitivity, and confusion matrix.
  • Random Forest achieved a test AUC of 0.723 and accuracy of 0.678, outperforming traditional models.
  • Feature importance analysis revealed primary site categories and treatment status as key predictors.
  • The model identified 53.1% advanced stage prevalence in high-risk patients versus 14.0% in low-risk patients.

Cite This Study

Song et al. (2025) studied this question.

synapsesocial.com/papers/69362f3d4fa91c937236d4e8https://doi.org/10.1182/blood-2025-7194
View Full Paper
Ask AI
Bookmark
Share