Pipeline to Summarize Radiology Reports

Description

This pretrained pipeline is built on the top of summarizer_radiology model.

Copy S3 URI

How to use

from sparknlp.pretrained import PretrainedPipeline

pipeline = PretrainedPipeline("summarizer_radiology_pipeline", "en", "clinical/models")

text = """INDICATIONS: Peripheral vascular disease with claudication.

RIGHT:
1. Normal arterial imaging of right lower extremity.
2. Peak systolic velocity is normal.
3. Arterial waveform is triphasic.
4. Ankle brachial index is 0.96.

LEFT:
1. Normal arterial imaging of left lower extremity.
2. Peak systolic velocity is normal.
3. Arterial waveform is triphasic throughout except in posterior tibial artery where it is biphasic.
4. Ankle brachial index is 1.06.

IMPRESSION: 
Normal arterial imaging of both lower lobes.
"""

result = pipeline.fullAnnotate(text)
import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline

val pipeline = new PretrainedPipeline("summarizer_radiology_pipeline", "en", "clinical/models")

val text = """INDICATIONS: Peripheral vascular disease with claudication.

RIGHT:
1. Normal arterial imaging of right lower extremity.
2. Peak systolic velocity is normal.
3. Arterial waveform is triphasic.
4. Ankle brachial index is 0.96.

LEFT:
1. Normal arterial imaging of left lower extremity.
2. Peak systolic velocity is normal.
3. Arterial waveform is triphasic throughout except in posterior tibial artery where it is biphasic.
4. Ankle brachial index is 1.06.

IMPRESSION: 
Normal arterial imaging of both lower lobes.
"""

val result = pipeline.fullAnnotate(text)

Results


The patient has peripheral vascular disease with claudication. The right lower extremity shows normal arterial imaging, but the peak systolic velocity is normal. The arterial waveform is triphasic throughout, except for the posterior tibial artery, which is biphasic. The ankle brachial index is 0.96. The impression is normal arterial imaging of both lower lobes.

Model Information

Model Name: summarizer_radiology_pipeline
Type: pipeline
Compatibility: Healthcare NLP 4.4.2+
License: Licensed
Edition: Official
Language: en
Size: 937.2 MB

Included Models

  • DocumentAssembler
  • MedicalSummarizer