Clinical Major Concepts to UMLS Code Pipeline

Description

This pretrained pipeline maps entities (Clinical Major Concepts) with their corresponding UMLS CUI codes. You’ll just feed your text and it will return the corresponding UMLS codes.

Predicted Entities

Qualitative_Concept, Mental_Process, Health_Care_Activity, Professional_or_Occupational_Group, Population_Group, Group, Pharmacologic_Substance, Research_Activity, Medical_Device, Diagnostic_Procedure, Molecular_Function, Spatial_Concept, Organic_Chemical, Amino_Acid, Peptide_or_Protein, Disease_or_Syndrome, Daily_or_Recreational_Activity, Quantitative_Concept, Biologic_Function, Organism_Attribute, Clinical_Attribute, Pathologic_Function, Eukaryote, Body_Part, Organ_or_Organ_Component, Anatomical_Structure, Cell_Component, Geographic_Area, Manufactured_Object, Tissue, Plant, Nucleic_Acid, Nucleoside_or_Nucleotide, Indicator, Reagent_or_Diagnostic_Aid, Prokaryote, Chemical, Therapeutic_or_Preventive_Procedure, Gene_or_Genome, Mammal, Laboratory_Procedure, Substance, Molecular_Biology_Research_Technique, Neoplastic_Process, Cell, Food, Genetic_Function, Mental_or_Behavioral_Dysfunction, Body_Substance, Sign_or_Symptom, Injury_or_Poisoning, Body_Location_or_Region, Organization, Body_System, Fungus, Virus, Nucleotide_Sequence, Biomedical_or_Dental_Material

Live Demo Open in Colab Download Copy S3 URI

How to use

from sparknlp.pretrained import PretrainedPipeline

resolver_pipeline = PretrainedPipeline("umls_major_concepts_resolver_pipeline", "en", "clinical/models")

result = resolver_pipeline.annotate("""The patient complains of pustules after falling from stairs. She has been advised Arthroscopy by her primary care pyhsician""")

resolver_pipeline = nlp.PretrainedPipeline("umls_major_concepts_resolver_pipeline", "en", "clinical/models")

result = resolver_pipeline.annotate("""The patient complains of pustules after falling from stairs. She has been advised Arthroscopy by her primary care pyhsician""")

import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline

val resolver_pipeline = PretrainedPipeline("umls_major_concepts_resolver_pipeline", "en", "clinical/models")

val result = resolver_pipeline.annotate("""The patient complains of pustules after falling from stairs. She has been advised Arthroscopy by her primary care pyhsician""")

Results

+----------------------+-----------------------------------+---------+
|chunk                 |ner_label                          |umls_code|
+----------------------+-----------------------------------+---------+
|pustules              |Sign_or_Symptom                    |C0241157 |
|stairs                |Daily_or_Recreational_Activity     |C4300351 |
|Arthroscopy           |Therapeutic_or_Preventive_Procedure|C0179144 |
|primary care pyhsician|Health_Care_Activity               |C3266804 |
+----------------------+-----------------------------------+---------+

Model Information

Included Models

• DocumentAssembler
• SentenceDetector
• TokenizerModel
• WordEmbeddingsModel
• MedicalNerModel
• NerConverter
• ChunkMapperModel
• ChunkMapperFilterer
• Chunk2Doc
• BertSentenceEmbeddings
• SentenceEntityResolverModel
• ResolverMerger