pLGG-Immune-Clinicoradiomics

This code repository includes the data and source codes used in the manuscript "Multiparametric MRI Along with Machine Learning Informs on Molecular Underpinnings, Prognosis, and Treatment Response In Pediatric Low-Grade Glioma"

Software Requirements

• CaPTk, v1.8.1 (https://cbica.github.io/CaPTk/)
• Python3
• R v4.3
• MATLAB 2023A (v23.2)
  • Parallel Computing Toolbox
  • Statistics and Machine Learning Toolbox

Hardware Used for this Study

• CUBIC (HPC Cluster) (https://www.med.upenn.edu/cbica/cubic.html)
• AWS/EC2 for batch image pre-processing, segmentation, radiomic feature extraction

MRI Pre-processing and Tumor Segmentation:

• Required MRI sequences: T1, T1CE, T2, FLAIR (ADC optional)
• Pre-processing using BraTS Pre-processing Pipeline, details explained in: https://cbica.github.io/CaPTk/preprocessing_brats.html
• Tumor Segmentation, all details provided in: https://github.com/d3b-center/peds-brain-auto-seg-public
• Skull-stripping to generate a brain mask: https://github.com/d3b-center/peds-brain-auto-skull-strip
• Image normalization:
  • run_rescale.py
• Whole tumor generation:
  • run_wtmask.py
• Radiomic feature extraction, using CaPTk v1.8.1: https://cbica.github.io/CaPTk/ht_FeatureExtraction.html;
  • parameter file for radiomic feature extraction: radiomic_feature_params_20230725.csv
  • sample batch file: SampleBatchFile.csv

Immune Profiling:

• analyses/lgg_xcell_analyses

Radioimmunomic Analysis:

• analyses/Radioimmunomic_Signature

Clinicoradiomic Risk Stratification:

• analyses/Clinicoradiomics

Assessment of transcriptomic pathways associated with clinicoradiomic risk:

• analyses/lgg_risk_analysis