Description
This pipeline is designed to:
- extract clinical/medical entities
- assign assertion status to the extracted entities
- establish relations between the extracted entities
from clinical texts. In this pipeline, 4 NER models, one assertion model, and one relation extraction model were used to achieve those tasks. Here are the NER, assertion, and relation extraction labels this pipeline can extract.
- Clinical Entity Labels: PROBLEM, TEST, TREATMENT
- Assertion Status Labels: Present, Absent, Possible, Planned, Past, Family, Hypotetical, SomeoneElse
- Relation Extraction Labels: TrAP, TeRP, TrIP, TrWP, TrCP, TrAP, TrNAP, TeCP, PIP
Predicted Entities
PROBLEM, TEST, TREATMENT
How to use
from sparknlp.pretrained import PretrainedPipeline
ner_pipeline = PretrainedPipeline("explain_clinical_doc_generic", "en", "clinical/models")
result = ner_pipeline.annotate("""Patient with severe fever and sore throat. He shows no stomach pain. He maintained on the epidural and PCA for pain control.
After CT, lung tumor located at the right lower lobe. Father with Alzheimer.""")
import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline
val ner_pipeline = PretrainedPipeline("explain_clinical_doc_generic", "en", "clinical/models")
val result = ner_pipeline.annotate("""Patient with severe fever and sore throat. He shows no stomach pain. He maintained on the epidural and PCA for pain control.
After CT, lung tumor located at the right lower lobe. Father with Alzheimer.""")
Results
# NER and Assertion Result
+------------+---------+------------+
|merged_chunk|entities |assertion |
+------------+---------+------------+
|severe fever|PROBLEM |Present |
|sore throat |PROBLEM |Present |
|stomach pain|PROBLEM |Absent |
|the epidural|TREATMENT|Present |
|PCA |TREATMENT|Past |
|pain |PROBLEM |Hypothetical|
|CT |PROBLEM |Past |
|lung tumor |PROBLEM |Present |
|Alzheimer |PROBLEM |Family |
+------------+---------+------------+
# Relation Extraction Result
+--------+-------------+-----------+------------+---------+-------------+-----------+-----------+-------+--------+----------+
|sentence|entity1_begin|entity1_end| chunk1| entity1|entity2_begin|entity2_end| chunk2|entity2|relation|confidence|
+--------+-------------+-----------+------------+---------+-------------+-----------+-----------+-------+--------+----------+
| 0| 13| 24|severe fever| PROBLEM| 30| 40|sore throat|PROBLEM| PIP| 0.9999982|
| 2| 86| 97|the epidural|TREATMENT| 111| 114| pain|PROBLEM| TrAP|0.81723267|
| 2| 103| 105| PCA|TREATMENT| 111| 114| pain|PROBLEM| TrAP|0.99933213|
| 3| 131| 132| CT| PROBLEM| 135| 144| lung tumor|PROBLEM| PIP| 0.999998|
+--------+-------------+-----------+------------+---------+-------------+-----------+-----------+-------+--------+----------+
Model Information
| Model Name: | explain_clinical_doc_generic |
| Type: | pipeline |
| Compatibility: | Healthcare NLP 5.2.1+ |
| License: | Licensed |
| Edition: | Official |
| Language: | en |
| Size: | 1.8 GB |
Included Models
- DocumentAssembler
- SentenceDetectorDLModel
- TokenizerModel
- WordEmbeddingsModel
- MedicalNerModel
- NerConverterInternalModel
- MedicalNerModel
- NerConverterInternalModel
- MedicalNerModel
- NerConverterInternalModel
- MedicalNerModel
- NerConverterInternalModel
- ChunkMergeModel
- AssertionDLModel
- PerceptronModel
- DependencyParserModel
- RelationExtractionModel