Dysphagia Risk Stratification in Head and Neck Cancer via Two-Stage PRO-Clinical Stacking

Dysphagia risk stratification in head and neck cancer using a two-stage PRO–Clinical stacking model. This repository contains the data-processing pipeline, model training notebooks, and the ensemble code used in our study.

Overview

Patients treated for head and neck cancer are at high risk of developing dysphagia (difficulty swallowing). This project predicts that risk by combining:

1. Stage 1 — an XGBoost model trained on patient-reported outcomes (PROs), such as the MD Anderson Dysphagia Inventory (MDADI) and DIGEST-related measures.
2. Stage 2 — a stacking model that combines the Stage 1 PRO predictions with raw clinical features to produce the final risk stratification.