radgraph 0.1.18

No project description provided

RadGraph

Table of Contents

• Requirements
• RadGraph
• F1-RadGraph
• Processed Annotations
• RadGraph v1

---

Requirements

python_requires=">=3.8",
install_requires=[
    "torch>=2.1.0",
    "transformers>=4.39.0",
    "appdirs",
    "jsonpickle",
    "filelock",
    "h5py",
    "nltk",
    "dotmap",
    "pytest",
],

Testing:

pytest

RadGraph

Usage:

from radgraph import RadGraph
radgraph = RadGraph(model="modern-radgraph-xl")
annotations = radgraph(["No evidence of pneumothorax following chest tube removal."])

Output:

{
  "0": {
    "text": "No evidence of pneumothorax following chest tube removal .",
    "entities": {
      "1": {
        "tokens": "pneumothorax",
        "label": "Observation::definitely absent",
        "start_ix": 3,
        "end_ix": 3,
        "relations": []
      },
      "2": {
        "tokens": "chest",
        "label": "Anatomy::definitely present",
        "start_ix": 5,
        "end_ix": 5,
        "relations": []
      },
      "3": {
        "tokens": "tube",
        "label": "Observation::definitely present",
        "start_ix": 6,
        "end_ix": 6,
        "relations": [
          [
            "located_at",
            "2"
          ]
        ]
      }
    },
    "data_source": null,
    "data_split": "inference"
  }
}

Official package as per:

@inproceedings{delbrouck-etal-2024-radgraph,
    title = "{R}ad{G}raph-{XL}: A Large-Scale Expert-Annotated Dataset for Entity and Relation Extraction from Radiology Reports",
    author = "Delbrouck, Jean-Benoit  and
      Chambon, Pierre  and
      Chen, Zhihong  and
      Varma, Maya  and
      Johnston, Andrew  and
      Blankemeier, Louis  and
      Van Veen, Dave  and
      Bui, Tan  and
      Truong, Steven  and
      Langlotz, Curtis",
    editor = "Ku, Lun-Wei  and
      Martins, Andre  and
      Srikumar, Vivek",
    booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand and virtual meeting",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.findings-acl.765",
    pages = "12902--12915",
    }

F1-RadGraph

Usage:
```python
from radgraph import F1RadGraph
refs = ["no acute cardiopulmonary abnormality",
        "endotracheal tube is present and bibasilar opacities likely represent mild atelectasis",
]

hyps = ["no acute cardiopulmonary abnormality",
        "et tube terminates 2 cm above the carina and bibasilar opacities"
]
f1radgraph = F1RadGraph(reward_level="all", model_type="radgraph-xl")
mean_reward, reward_list, hypothesis_annotation_lists, reference_annotation_lists = f1radgraph(hyps=hyps, refs=refs)

rg_e, rg_er, rg_bar_er = mean_reward

print(mean_reward)

Output:

(np.float64(0.75), np.float64(0.6666666666666666), np.float64(0.6538461538461539))

Over the years, RG_ER has been reported widely in RRG papers.

F1RadGraph as per:

@inproceedings{delbrouck-etal-2022-improving,
    title = "Improving the Factual Correctness of Radiology Report Generation with Semantic Rewards",
    author = "Delbrouck, Jean-Benoit  and
      Chambon, Pierre  and
      Bluethgen, Christian  and
      Tsai, Emily  and
      Almusa, Omar  and
      Langlotz, Curtis",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2022",
    month = dec,
    year = "2022",
    address = "Abu Dhabi, United Arab Emirates",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2022.findings-emnlp.319",
    pages = "4348--4360",
}

Processed Annotations

import json
from radgraph import get_radgraph_processed_annotations, RadGraph

report = """
A right-sided chest tube is present with the distal end near the right lung apex.  
Right central line ends at lower SVC.  Innumerable, bilateral, nodular opacities are similar.  
The size of the pneumothorax at the right lung apex is smaller whereas at the right lower lateral chest wall and at the right lung base is overall unchanged. 
Spinal hardware device is present at lower thoracic and upper lumbar region. Increase retrocardiac density representing left lower lung volume loss,
moderate left and mild right pleural effusions are stable.
"""
model_type = "modern-radgraph-xl"
radgraph = RadGraph(model_type=model_type)
annotations = radgraph(
    [report]
    )

processed_annotations = get_radgraph_processed_annotations(annotations)


for annotation in processed_annotations["processed_annotations"]:
    located_at = f" [Location: {', '.join(annotation['located_at'])}]" if annotation["located_at"] else ""
    suggestive_of = f" [Suggestive of: {', '.join(annotation['suggestive_of'])}]" if annotation["suggestive_of"] else ""
    tag = f" [Tag: {annotation['tags'][0]}]"
    print(f"Observation: {annotation['observation']}{located_at}{suggestive_of}{tag}")

Output:

Observation: tube distal end [Location: chest] [Tag: definitely present]
Observation: central line [Location: right, lower svc] [Tag: definitely present]
Observation: innumerable nodular opacities similar [Location: bilateral] [Tag: definitely present]
Observation: size pneumothorax unchanged [Location: right lower lateral chest wall] [Tag: definitely present]
Observation: smaller [Tag: definitely present]
Observation: hardware device [Location: spinal] [Tag: definitely present]
Observation: increase density stable [Location: retrocardiac] [Suggestive of: density suggestive of loss] [Tag: definitely present]
Observation: loss [Location: volume] [Tag: definitely present]
Observation: moderate mild effusions stable [Location: left right pleural] [Tag: definitely present]

RadGraph v1

radgraph = RadGraph(model="radgraph")

@inproceedings{NEURIPS DATASETS AND BENCHMARKS2021_c8ffe9a5,
 author = {Jain, Saahil and Agrawal, Ashwin and Saporta, Adriel and Truong, Steven and Duong, Du Nguyen Duong Nguyen and Bui, Tan and Chambon, Pierre and Zhang, Yuhao and Lungren, Matthew and Ng, Andrew and Langlotz, Curtis and Rajpurkar, Pranav and Rajpurkar, Pranav},
 booktitle = {Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks},
 editor = {J. Vanschoren and S. Yeung},
 pages = {},
 publisher = {Curran},
 title = {RadGraph: Extracting Clinical Entities and Relations from Radiology Reports},
 url = {https://datasets-benchmarks-proceedings.neurips.cc/paper_files/paper/2021/file/c8ffe9a587b126f152ed3d89a146b445-Paper-round1.pdf},
 volume = {1},
 year = {2021}
}