Multilabel Text Classification For Respiratory Disease

Description

The PHS-BERT Respiratory Disease Classifier Model is a specialized text classification system, engineered to accurately identify and categorize textual mentions of four prominent respiratory diseases: Asthma, Chronic Obstructive Pulmonary Disease (COPD), Emphysema, and Chronic bronchitis. More detailed information about classes as follows:

Asthma: A classification indicating textual mentions explicitly or implicitly referring to Asthma, a condition characterized by chronic inflammation of the airways, leading to episodes of wheezing, shortness of breath, chest tightness, and coughing. Example: “I can’t take part in the marathon due to my persistent asthma issues.

Chronic Obstructive Pulmonary Disease (COPD): This category encapsulates text referring to COPD, a progressive lung disease that engenders obstructed airflow from the lungs. Symptoms include breathing difficulty, cough, mucus production, and wheezing. Example: “COPD makes it incredibly hard for my dad to walk long distances without becoming breathless.”

Emphysema: Text that signifies mentions of Emphysema falls into this classification. Emphysema, a subset of COPD, involves the gradual damage of the air sacs (alveoli) in the lungs, impeding the outward flow of air and causing breathlessness. Example: “Ever since being diagnosed with emphysema, climbing stairs has become a significant challenge.”

Chronic Bronchitis: Any textual content that points toward Chronic Bronchitis is categorized here. Chronic bronchitis is a form of COPD characterized by a chronic cough and mucus production due to the long-term inflammation of the bronchial tubes. Example: “The incessant coughing from chronic bronchitis keeps me awake most nights.”

Predicted Entities

Astham,COPD, Emphysema, Chronic bronchitis, Other/Unknown, No

Live Demo Open in Colab Download Copy S3 URI

How to use

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

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

word_embeddings = WordEmbeddingsModel.pretrained("embeddings_clinical", "en", "clinical/models")\
    .setInputCols(["document", "token"])\
    .setOutputCol("embeddings")

sentence_embeddings = SentenceEmbeddings()\
    .setInputCols(["document", "embeddings"]) \
    .setOutputCol("sentence_embeddings") \
    .setPoolingStrategy("AVERAGE")

multiclassifierdl = MultiClassifierDLModel.pretrained("multiclassifierdl_respiratory_disease", "en", "clinical/models")\
    .setInputCols(["sentence_embeddings"])\
    .setOutputCol("predicted_class")\
    .setThreshold(0.999)

clf_pipeline = Pipeline(
    stages=[
        documentAssembler,
        tokenizer,
        word_embeddings,
        sentence_embeddings,
        multiclassifierdl
])


data = spark.createDataFrame([
        ["""The patient takes inhalers for COPD management, weight loss medications, and disease-modifying antirheumatic drugs (DMARDs) for rheumatoid arthritis."""],
        ["""The patient was on Metformin for DM2, mood stabilizers for Bipolar II Disorder, and inhaled corticosteroids for Asthma."""],
        ["""The patient was diagnosed with Chronic Bronchitis after a series of pulmonary function tests."""],
        ["""Chest CT imaging revealed significant bullae and airspace enlargement, consistent with a diagnosis of emphysema."""],
    ]).toDF("text")


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

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

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

val wordEmbeddings = WordEmbeddingsModel.pretrained("embeddings_healthcare_100d", "en", "clinical/models")
    .setInputCols(Array("document", "token"))
    .setOutputCol("embeddings")

val sentence_embeddings = new SentenceEmbeddings()\
    .setInputCols(Array()"document", "embeddings")) \
    .setOutputCol("sentence_embeddings") \
    .setPoolingStrategy("AVERAGE")

val multiclassifierdl = MultiClassifierDLModel.pretrained("multiclassifierdl_respiratory_disease", "en", "clinical/models")\
    .setInputCols("sentence_embeddings")\
    .setOutputCol("predicted_class")\
    .setThreshold(0.999)

val clf_pipeline = new Pipeline().setStages(Array(
    documentAssembler,
    tokenizer,
    wordEmbeddings,
    sentence_embeddings,
    multiclassifierdl
))

val data = Seq(Array(
    """The patient takes inhalers for COPD management, weight loss medications, and disease-modifying antirheumatic drugs (DMARDs) for rheumatoid arthritis.""",
    """The patient was on Metformin for DM2, mood stabilizers for Bipolar II Disorder, and inhaled corticosteroids for Asthma.""",
    """The patient was diagnosed with Chronic Bronchitis after a series of pulmonary function tests.""",
    """Chest CT imaging revealed significant bullae and airspace enlargement, consistent with a diagnosis of emphysema.""",
    )).toDS.toDF("text")

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

Results

+----------------------------------------------------------------------------------------------------+--------------------+
|                                                                                                text|              result|
+----------------------------------------------------------------------------------------------------+--------------------+
|The patient takes inhalers for COPD management, weight loss medications, and disease-modifying an...|              [COPD]|
|The patient was on Metformin for DM2, mood stabilizers for Bipolar II Disorder, and inhaled corti...|            [Asthma]|
|       The patient was diagnosed with Chronic Bronchitis after a series of pulmonary function tests.|[Chronic bronchitis]|
|Chest CT imaging revealed significant bullae and airspace enlargement, consistent with a diagnosi...|         [Emphysema]|
+----------------------------------------------------------------------------------------------------+--------------------+

Model Information

References

Trained with the in-house dataset

Sample text from the training dataset

Asthma:The patient was first diagnosed with asthma at the age of 12 following a severe respiratory infection. The patient reports experiencing wheezing, shortness of breath, and chest tightness, consistent with asthma exacerbations. The patient has been prescribed a combination inhaler containing a long-acting beta-agonist and an inhaled corticosteroid to manage and prevent asthma symptoms.

Chronic Obstructive Pulmonary Disease (COPD): Mr. Smith was diagnosed with Chronic Obstructive Pulmonary Disease (COPD) 5 years ago, primarily attributed to his 30-year smoking history. He frequently experiences chronic coughing with mucus production and difficulty in breathing, especially during physical activities, indicative of his COPD. As part of his COPD management, the patient has been advised to use a bronchodilator inhaler regularly and undergo pulmonary rehabilitation to improve lung function and quality of life.

Emphysema: The patient’s emphysema diagnosis was confirmed three years ago after a high-resolution CT scanshowed damage to the alveoli. The patient complains of progressive shortness of breath and an inability to sustain physical exertion, characteristics of emphysema. Oxygen therapy has been recommended for the patient to alleviate the symptoms of emphysema and improve oxygen saturation levels.

Chronic Bronchitis: Mrs. Johnson has a recurring history of chronic bronchitis, often triggered by winter months and viral infections. She presents with persistent coughing that produces yellowish mucus, accompanied by fatigue and chest discomfort, hallmark signs of chronic bronchitis. The treatment plan includes regular use of mucolytic agents, chest physiotherapy, and a short course of bronchodilator therapy to relieve symptoms of chronic bronchitis.

Benchmarking

label                tp	   fp	   fn	   prec	       rec	       f1
Other/Unknown        13	   10	   34	   0.5652174	 0.27659574	 0.37142858
Emphysema            143   23	   38	   0.8614458	 0.7900553	 0.82420754
COPD                 267   27	   52	   0.90816325	 0.8369906	 0.8711256
No                   55	   8	   19	   0.8730159	 0.7432432	 0.8029197
Chronic bronchitis   241   27	   25	   0.8992537	 0.90601504	 0.9026217
Asthma               104   15	   25	   0.8739496	 0.8062016	 0.83870965
Macro-average        823   110   193   0.83017427  0.7265169   0.7748944
Micro-average        823   110   193   0.88210076  0.8100393   0.84453565