linuszvideos-boop

linuszvideos-boop / gc_recurrence

Public
19
0
100% credibility
Found May 17, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

This repository contains the analysis pipeline for a published medical research study that predicts gastric cancer recurrence after surgery. The project trains six machine learning models on patient data, handles missing laboratory values, and validates predictions on independent patient groups. Researchers can use this code to reproduce the study's findings or apply the same methods to their own patient datasets. The pipeline produces performance charts, calibration curves, and statistical metrics while following strict medical reporting guidelines (TRIPOD+AI and PROBAST). Patient data is not included due to privacy protections.

How It Works

1
🔬 You discover a cancer prediction study

A researcher shares a study about predicting gastric cancer return after surgery using machine learning, and you want to explore it further.

2
📋 You learn what the study does

The study trains six different prediction models on patient data to guess whether cancer will come back, then tests them on separate patient groups.

3
📊 You see how well it works

The best model achieved 83% accuracy in predicting recurrence, with clear charts showing calibration and feature importance.

4
🧬 You prepare your own patient data

You organize your hospital's gastric cancer patient records into a spreadsheet with the same format, ready for analysis.

5
You choose how to run the analysis
🚀
Run the standard analysis

Launch the prediction pipeline with recommended settings to generate all charts and metrics automatically.

🔧
Customize your analysis

Adjust how missing values are handled or enable hyperparameter tuning for your specific dataset.

6
📁 Your results are generated

The pipeline creates prediction charts, performance metrics, and comparison tables saved to your computer.

🎯 You have validated predictions for your patients

Your analysis is complete with reproducible results following medical research standards, ready for clinical review or publication.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 19 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is gc_recurrence?

gc_recurrence is a Python machine learning pipeline that predicts postoperative gastric cancer recurrence. It trains six classifiers (Random Forest, XGBoost, LightGBM, CatBoost, Logistic Regression, and MLP) on clinical data, then rigorously evaluates them using ROC-AUC, calibration analysis, decision curves, and SHAP feature importance. The recommended model achieves an external validation ROC-AUC of 0.829 with strong calibration performance. It handles missing lab values through iterative imputation and applies Platt scaling for probability calibration.

Why is it gaining traction?

This stands out because it is a complete, reproducible analysis pipeline rather than just model code. It follows TRIPOD+AI and PROBAST reporting standards, which are increasingly required for medical AI publications. The pipeline includes sensitivity analyses out of the box, handles small external validation cohorts gracefully, and generates publication-ready figures. The command-line interface makes it straightforward to run primary analyses, switch imputation methods, or enable hyperparameter tuning with simple flags.

Who should use this?

Clinical researchers building prognostic models for cancer outcomes will find the most value here. Medical data scientists working with small datasets and missing laboratory values can adopt the imputation and calibration strategies. Academic researchers publishing in journals requiring TRIPOD+AI compliance have a ready-made framework. Bioinformaticians seeking reproducible ML workflows for clinical applications will appreciate the structured approach.

Verdict

gc_recurrence is a well-documented, standards-compliant pipeline for clinical ML research, but the 1.0% credibility score and 19 stars indicate it is early-stage with limited community validation. If you are working in gastric cancer prediction or need TRIPOD+AI compliant reporting, it is worth exploring as a reference implementation. Validate thoroughly before clinical deployment.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.