T15: Predicting hip Kellgren-Lawrence score¶
Objective:
This task aims to classify the radiology report by the Hip Kellgren-Lawrence score for hip osteoarthritis, ranging from 0 to 4.
Patient Population:
Radiology reports (in Dutch) for hiposteoarthritis were collected at Radboudumc.
Imaging Data:
Not applicable. The task is based solely on textual data — radiology reports written in Dutch.
Test Data:
The test set contains 172 cases.
Reference Standard:
- Reports were manually annotated by a trained investigator.
- Difficult cases were reviewed by a radiologist.
- Labels include Kellgren-Lawrence scores (0–4) and two additional categories: "hip prosthesis" and "not determinable."
- The dataset is imbalanced, and while the scores are ordinal, the additional categories are not.
Evaluation Metrics:
Performance is assessed using unweighted Kappa.
Predictions for the left and right hip are concatenated, and unweighted Kappa is computed on the combined predictions.
Relation to Existing Challenges:¶
- Task 15 is derived from Task018 (Hip Kellgren-Lawrence scoring) of the DRAGON challenge.
- Unlike the DRAGON challenge, UNICORN allows training on platform using only a small set of few-shot examples per task.