Edit on GitHub
  • John Snow Labs
  • Jul 25, 2022

Detect Diseases in Medical Text

  • en
  • ner
  • clinical
  • licensed
  • bertfortokenclassification

Description

Chemicals, diseases, and their relations are among the most searched topics by PubMed users worldwide as they play central roles in many areas of biomedical research and healthcare, such as drug discovery and safety surveillance.

This model is trained with the BertForTokenClassification method from the transformers library and imported into Spark NLP. The model detects disease from a medical text

Predicted Entities

DISEASE

Open in Colab Copy S3 URI

How to use

document_assembler = DocumentAssembler()\
    .setInputCol("text")\
    .setOutputCol("document")\

sentence_detector = SentenceDetectorDLModel.pretrained("sentence_detector_dl_healthcare", "en", "clinical/models")\
    .setInputCols(["document"])\
    .setOutputCol("sentence")

tokenizer = Tokenizer()\
    .setInputCols(["sentence"])\
    .setOutputCol("token")

ner_model = MedicalBertForTokenClassifier.pretrained("bert_token_classifier_ner_bc5cdr_disease", "en", "clinical/models")\
    .setInputCols(["sentence", "token"])\
    .setOutputCol("ner")\
    .setCaseSensitive(True)\
    .setMaxSentenceLength(512)

ner_converter = NerConverter()\
    .setInputCols(["sentence", "token", "ner"])\
    .setOutputCol("ner_chunk")

pipeline = Pipeline(stages=[
    document_assembler, 
    sentence_detector,
    tokenizer,
    ner_model,
    ner_converter   
    ])


data = spark.createDataFrame([["""Indomethacin resulted in histopathologic findings typical of interstitial cystitis, such as leaky bladder epithelium and mucosal mastocytosis. The true incidence of nonsteroidal anti-inflammatory drug-induced cystitis in humans must be clarified by prospective clinical trials. An open-label phase II study of low-dose thalidomide in androgen-independent prostate cancer."""]]).toDF("text")

result = pipeline.fit(data).transform(data)

val document_assembler = new DocumentAssembler()
    .setInputCol("text")
    .setOutputCol("document")

val sentence_detector = SentenceDetectorDLModel.pretrained("sentence_detector_dl_healthcare", "en", "clinical/models")
    .setInputCols(Array("document"))
    .setOutputCol("sentence")

val tokenizer = new Tokenizer()
    .setInputCols(Array("sentence"))
    .setOutputCol("token")

val ner_model = MedicalBertForTokenClassifier.pretrained("bert_token_classifier_ner_bc5cdr_disease", "en", "clinical/models")
    .setInputCols(Array("sentence", "token"))
    .setOutputCol("ner")
    .setCaseSensitive(True)
    .setMaxSentenceLength(512)

val ner_converter = new NerConverter()
    .setInputCols(Array("sentence", "token", "ner"))
    .setOutputCol("ner_chunk")

val pipeline = new Pipeline().setStages(Array(document_assembler, 
                                                   sentence_detector,
                                                   tokenizer,
                                                   ner_model,
                                                   ner_converter))

val data = Seq("""Indomethacin resulted in histopathologic findings typical of interstitial cystitis, such as leaky bladder epithelium and mucosal mastocytosis. The true incidence of nonsteroidal anti-inflammatory drug-induced cystitis in humans must be clarified by prospective clinical trials. An open-label phase II study of low-dose thalidomide in androgen-independent prostate cancer.""").toDS.toDF("text")

val result = pipeline.fit(data).transform(data)

import nlu
nlu.load("en.classify.token_bert.bc5cdr_disease").predict("""Indomethacin resulted in histopathologic findings typical of interstitial cystitis, such as leaky bladder epithelium and mucosal mastocytosis. The true incidence of nonsteroidal anti-inflammatory drug-induced cystitis in humans must be clarified by prospective clinical trials. An open-label phase II study of low-dose thalidomide in androgen-independent prostate cancer.""")

Results

+---------------------+-------+
|ner_chunk            |label  |
+---------------------+-------+
|interstitial cystitis|DISEASE|
|mastocytosis         |DISEASE|
|cystitis             |DISEASE|
|prostate cancer      |DISEASE|
+---------------------+-------+

Model Information

Model Name: bert_token_classifier_ner_bc5cdr_disease
Compatibility: Healthcare NLP 4.0.0+
License: Licensed
Edition: Official
Input Labels: [sentence, token]
Output Labels: [ner]
Language: en
Size: 404.2 MB
Case sensitive: true
Max sentence length: 512

References

https://github.com/cambridgeltl/MTL-Bioinformatics-2016

Benchmarking

 label         precision  recall  f1-score  support 
 B-DISEASE     0.7905     0.9146  0.8480    4424    
 I-DISEASE     0.6521     0.8725  0.7464    2737    
 micro-avg     0.7328     0.8985  0.8072    7161    
 macro-avg     0.7213     0.8935  0.7972    7161    
 weighted-avg  0.7376     0.8985  0.8092    7161