Detect Entities Related to Cancer Therapies

Description

This model extracts entities related to oncology therapies using granular labels, including mentions of treatments, posology information and line of therapy.

Definitions of Predicted Entities:

• Cancer_Surgery: Terms that indicate surgery as a form of cancer treatment.
• Chemotherapy: Mentions of chemotherapy drugs, or unspecific words such as “chemotherapy”.
• Cycle_Count: The total number of cycles being administered of an oncological therapy (e.g. “5 cycles”).
• Cycle_Day: References to the day of the cycle of oncological therapy (e.g. “day 5”).
• Cycle_Number: The number of the cycle of an oncological therapy that is being applied (e.g. “third cycle”).
• Dosage: The quantity prescribed by the physician for an active ingredient.
• Duration: Words indicating the duration of a treatment (e.g. “for 2 weeks”).
• Frequency: Words indicating the frequency of treatment administration (e.g. “daily” or “bid”).
• Hormonal_Therapy: Mentions of hormonal drugs used to treat cancer, or unspecific words such as “hormonal therapy”.
• Immunotherapy: Mentions of immunotherapy drugs, or unspecific words such as “immunotherapy”.
• Line_Of_Therapy: Explicit references to the line of therapy of an oncological therapy (e.g. “first-line treatment”).
• Radiotherapy: Terms that indicate the use of Radiotherapy.
• Radiation_Dose: Dose used in radiotherapy.
• Response_To_Treatment: Terms related to clinical progress of the patient related to cancer treatment, including “recurrence”, “bad response” or “improvement”.
• Route: Words indicating the type of administration route (such as “PO” or “transdermal”).
• Targeted_Therapy: Mentions of targeted therapy drugs, or unspecific words such as “targeted therapy”.
• Unspecific_Therapy: Terms that indicate a known cancer therapy but that is not specific to any other therapy entity (e.g. “chemoradiotherapy” or “adjuvant therapy”).

Predicted Entities

Cancer_Surgery, Chemotherapy, Cycle_Count, Cycle_Day, Cycle_Number, Dosage, Duration, Frequency, Hormonal_Therapy, Immunotherapy, Line_Of_Therapy, Radiotherapy, Radiation_Dose, Response_To_Treatment, Route, Targeted_Therapy, Unspecific_Therapy

Live Demo Open in Colab Download 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")

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

ner = MedicalNerModel.pretrained("ner_oncology_therapy", "en", "clinical/models") \
    .setInputCols(["sentence", "token", "embeddings"]) \
    .setOutputCol("ner")

ner_converter = NerConverter() \
    .setInputCols(["sentence", "token", "ner"]) \
    .setOutputCol("ner_chunk")
pipeline = Pipeline(stages=[document_assembler,
                            sentence_detector,
                            tokenizer,
                            word_embeddings,
                            ner,
                            ner_converter])

data = spark.createDataFrame([["The had previously undergone a left mastectomy and an axillary lymph node dissection for a left breast cancer twenty years ago.
The tumor was positive for ER and PR. Postoperatively, radiotherapy was administered to her breast.
The cancer recurred as a right lung metastasis 13 years later. The patient underwent a regimen consisting of adriamycin (60 mg/m2) and cyclophosphamide (600 mg/m2) over six courses, as first line therapy."]]).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("document")
    .setOutputCol("sentence")
    
val tokenizer = new Tokenizer()
    .setInputCols("sentence")
    .setOutputCol("token")
    
val word_embeddings = WordEmbeddingsModel().pretrained("embeddings_clinical", "en", "clinical/models")
    .setInputCols(Array("sentence", "token"))
    .setOutputCol("embeddings")                
    
val ner = MedicalNerModel.pretrained("ner_oncology_therapy", "en", "clinical/models")
    .setInputCols(Array("sentence", "token", "embeddings"))
    .setOutputCol("ner")
    
val ner_converter = new NerConverter()
    .setInputCols(Array("sentence", "token", "ner"))
    .setOutputCol("ner_chunk")

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

val data = Seq("The had previously undergone a left mastectomy and an axillary lymph node dissection for a left breast cancer twenty years ago.
The tumor was positive for ER and PR. Postoperatively, radiotherapy was administered to her breast.
The cancer recurred as a right lung metastasis 13 years later. The patient underwent a regimen consisting of adriamycin (60 mg/m2) and cyclophosphamide (600 mg/m2) over six courses, as first line therapy.").toDS.toDF("text")

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

import nlu
nlu.load("en.med_ner.oncology_therapy").predict("""The had previously undergone a left mastectomy and an axillary lymph node dissection for a left breast cancer twenty years ago.
The tumor was positive for ER and PR. Postoperatively, radiotherapy was administered to her breast.
The cancer recurred as a right lung metastasis 13 years later. The patient underwent a regimen consisting of adriamycin (60 mg/m2) and cyclophosphamide (600 mg/m2) over six courses, as first line therapy.""")

Results

| chunk                          | ner_label             |
|:-------------------------------|:----------------------|
| mastectomy                     | Cancer_Surgery        |
| axillary lymph node dissection | Cancer_Surgery        |
| radiotherapy                   | Radiotherapy          |
| recurred                       | Response_To_Treatment |
| adriamycin                     | Chemotherapy          |
| 60 mg/m2                       | Dosage                |
| cyclophosphamide               | Chemotherapy          |
| 600 mg/m2                      | Dosage                |
| six courses                    | Cycle_Count           |
| first line                     | Line_Of_Therapy       |

Model Information

References

In-house annotated oncology case reports.

Benchmarking

                label   tp   fp   fn  total  precision  recall   f1
         Cycle_Number   78   41   19     97       0.66    0.80 0.72
Response_To_Treatment  451  205  145    596       0.69    0.76 0.72
          Cycle_Count  210   75   20    230       0.74    0.91 0.82
   Unspecific_Therapy  189   76   89    278       0.71    0.68 0.70
         Chemotherapy  831   87   48    879       0.91    0.95 0.92
     Targeted_Therapy  194   28   34    228       0.87    0.85 0.86
         Radiotherapy  279   35   31    310       0.89    0.90 0.89
       Cancer_Surgery  720  192   99    819       0.79    0.88 0.83
      Line_Of_Therapy   95    6   11    106       0.94    0.90 0.92
     Hormonal_Therapy  170    6   15    185       0.97    0.92 0.94
        Immunotherapy   96   17   32    128       0.85    0.75 0.80
            Cycle_Day  205   38   43    248       0.84    0.83 0.84
            Frequency  363   33   64    427       0.92    0.85 0.88
                Route   93    6   20    113       0.94    0.82 0.88
             Duration  527  102  234    761       0.84    0.69 0.76
               Dosage  959   63  101   1060       0.94    0.90 0.92
       Radiation_Dose  106   12   20    126       0.90    0.84 0.87
            macro_avg 5566 1022 1025   6591       0.85    0.84 0.84
            micro_avg 5566 1022 1025   6591       0.85    0.84 0.84