Improving Interpretability of Radiology Report-based Pediatric Brain Tumor Pathology Classification and Key-phrases Extraction Using Large Language Models
Bibliographic record
Abstract
Radiology reports are crucial for bridging the expertise of radiologists and other clinicians. Machine Learning models trained on these reports have shown promising performance in various downstream clinical tasks, such as predicting the necessity of future follow-up procedures, based on past radiology reports. However, for clinicians to adopt these models and for radiologists to validate the results, interpretability of the model is essential. In this study, we train BERT models on radiology reports to classify pediatric brain tumor pathologies. These large language models enable accurate report-level classification, without the need for costly word-level annotations. To identify and extract keywords and key-phrases related to distinct pathologies from radiology reports, we used a modified Term Frequency-Inverse Document Frequency to determine phrase importance based on prevalence and attributions scores. We achieved an overall multiclass Area Under Receiver Operating Characteristic Curve (AUROC) of 79.57% using ClinicaiBERT. Moreover, the per-class AUROC values were 86%, 71.2%, and 81.5%, for ‘Pilocytic Astrocytoma’, ‘Low-Grade Astrocytoma’, and ‘Other’ pathologies, respectively. Our explainability analysis identified hypotonia and mesencephalon as the most important terms for ‘Pilocytic Astrocytoma’ and ‘Low-Grade Astrocytoma’, respectively.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.031 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.006 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".