Radiomic Survival Predictor (RaSPr)

Radiomic Survival Predictor (RaSPr) is a user-friendly tool that estimates survival risk for glioblastoma patients using advanced analysis of MRI images. By extracting and analyzing a wide range of imaging features such as tumor shape, texture, and intensity, RaSPr provides a non-invasive way to predict patient outcomes.

Features

• Computes the RaSPr Score using eight predefined radiomic features.
• Classifies each row into High or Low risk groups based on a customizable cutoff.
• Command-line interface for ease of automation.
• Optional CSV output with scores and risk group included.
• Seamlessly integrated into the RadGLO web platform for GUI-based execution.
• Recommended: Use command-line version for batch analysis; use RadGLO for single-sample predictions.

---

Dependencies

• R (≥ 3.5.0 recommended)
• tidyverse package

---

Install tidyverse if not already available:

install.packages("tidyverse")

NOTE: Features must be extracted only using PyRadiomics and must be scaled using "scale" function of R.

Feature extraction and normalization

Use this YAML file for extracting features using PyRadiomics

pyradiomics path/to/XYZ_t1.nii.gz path/to/XYZ_seg.nii.gz -p pyradiomics_FE_file.yaml -o path/to/output.txt

Usage

Clone the Repository

Run the following command in your terminal:

git clone https://github.com/CGnTLab/RaSPr.git

Navigate to the Directory

cd RaSPr

1. Via Command Line

Rscript raspr_score.R <input_file.csv> [output_file.csv]

• <input_file.csv>: Path to the input CSV file containing radiomic features.
• [output_file.csv] (optional): Path to save the output CSV file with RaSPr scores and classifications.

---

Example

Rscript raspr_score.R example_run.csv results.csv

---

2. Within R Console

source("raspr_score.R")
calculate_raspr("example_run.csv", "output_file.csv")

• input_file.csv: Path to the input CSV file containing radiomic features.
• output_file.csv : Path to save output in CSV format with RaSPr scores and classifications

---

Developer

Kavita Kundal
CG&T Lab, Indian Institute of Technology Hyderabad

📩 Email: cgntlab@gmail.com