Osteoarthritis Initiative — Cartilage Thickness Maps and Scoring Dataset

Overview

This dataset provides knee cartilage thickness maps (CTh-Maps) derived from MRI data of the Osteoarthritis Initiative (OAI) cohort, together with their corresponding cartilage osteoarthritic severity scores (CTh-Score, range 0–100) computed by the automatic framework described in the article. 

It is published in conjunction with the article:

Dataset Content

    OAI_CTh-Maps_CTh-Score_Dataset_v1/   
    ├── CTh_maps_00m/          # Cartilage thickness maps — baseline (0 months)   
    ├── CTh_maps_12m/          # Cartilage thickness maps — 12 months   
    ├── CTh_maps_24m/          # Cartilage thickness maps — 24 months   
    ├── CTh_maps_36m/          # Cartilage thickness maps — 36 months   
    ├── CTh_maps_48m/          # Cartilage thickness maps — 48 months   
    ├── CTh_maps_72m/          # Cartilage thickness maps — 72 months   
    ├── CTh_maps_96m/          # Cartilage thickness maps — 96 months   
    └── df_cth_score_v1.csv    # CTh-Score table for all subjects and timepoints   

CTh-Maps (CTh_maps_*/)

Each subdirectory contains 2D cartilage thickness maps stored as NumPy files (.npy).

File naming convention:

{ID}_{SIDE}_v1_CTh_map_{Timepoint}.npy   

Example: 9000099_LEFT_v1_CTh_map_00m.npy

Array format: float64, shape (170, 80) — a 2D projection map of cartilage thickness values (in mm). Background pixels (same for all) are encoded as NaN.

CTh-Score table (df_cth_score_v1.csv)

A CSV file containing the cartilage osteoarthritic severity score for each knee and timepoint.

Usage

    import numpy as np   
    import pandas as pd   
        
    # Load a cartilage thickness map   
    cth_map = np.load("CTh_maps_00m/9000099_LEFT_v1_CTh_map_00m.npy")   
    print(cth_map.shape)   # (170, 80)   
        
    fig, ax = plt.subplots(1, 1, figsize=(12, 5))   
        
    # Original image   
    im = ax.imshow(cth_map, cmap='jet_r')   
    ax.set_title('CTh-Map')   
    plt.colorbar(im, ax=ax, fraction=0.046, pad=0.04)   
        
    fig.tight_layout()   
    plt.show()   
        
    # Load the CTh-Score table   
    df = pd.read_csv("df_cth_score_v1.csv")   
    print(df.head())   

Citation

If you use this dataset, please cite the associated article:

    Automatic framework for evaluating osteoarthritic cartilage severity: high-resolution cartilage thickness mapping and scoring.   

    European Radiology (2026). https://doi.org/10.1007/s00330-026-12414-w   

OAI Data Access

The CTh-Maps were derived from images of the Osteoarthritis Initiative (OAI). Access to the original OAI imaging and clinical data requires registration at:
https://nda.nih.gov/oai/