Clinical Characteristics for Predicting Oral Mucositis in Head and Neck Cancer Patients

A dataset of multimodal data from 108 patients, collected to evaluate machine learning models for predicting radiation-induced oral mucositis. The data includes CT imaging, dose distribution, and clinical features. It was authored by Ling Li and last updated on 2026-04 09.

Use Cases

• Benchmarking traditional ML algorithms (e.g., Extra Trees, Logistic Regression) for multi-class prediction based on clinical and dosimetric features.
• Evaluating deep learning architectures (1D-CNN vs. 3D-CNN) on high-dimensional multimodal data under limited sample size conditions.
• Studying model overfitting and mode collapse phenomena in small-cohort radiomics analyses.
• Developing predictive models for radiation-induced oral mucositis severity based on fused low-dimensional features.

Strengths

• Dataset includes 108 patient records with multimodal data (CT imaging, dose distribution, clinical features).
• Model evaluation employed a robust stratified 5-fold cross-validation framework.
• Inter-rater reliability for the target variable (RIOM grading) was excellent (Cohen’s kappa = 0.82).

Limitations

• Row count and column-level documentation are unknown, which may limit suitability assessment.
• The dataset is very small at 2.2 KB, indicating limited scope and feature detail.
• Data may reflect temporal or source bias inherent to the specific study cohort.

Provenance

Source

figshare, author Ling Li.

Collection Method

Multimodal data collected from patients for a study on predictive modeling.

Time Range

null

Freshness

Last updated 2026-04-09 17:28:59; freshness should be verified.

Geography

null