Description
This specialized biomarker pipeline can;
-
extract biomarker entities,
-
classify sentences whether they contain biomarker entities or not,
-
establish relations between the extracted biomarker and biomarker results from the clinical documents.
In this pipeline, two NER, one text matcher, one sentence classifier, and one relation extraction model were employed to accomplish the designated tasks.
-
Clinical Entity Labels:
Biomarker,Biomarker_Result -
Relation Extraction Labels:
is_finding_of -
Classification Model Labels:
1,0
How to use
from sparknlp.pretrained import PretrainedPipeline
biomarker_pipeline = PretrainedPipeline("explain_clinical_doc_biomarker", "en", "clinical/models")
result = biomarker_pipeline.fullAnnotate("""In the bone- marrow (BM) aspiration, blasts accounted for 88.1% of ANCs, which were positive for CD9 and CD10 on flow cytometry. Measurements of serum tumor markers showed elevated level of Cyfra21-1: 4.77 ng/mL, NSE: 19.60 ng/mL, and SCCA: 2.58 ng/mL. Immunohistochemical staining showed positive staining for CK5/6, P40, and negative staining for TTF-1 and weakly positive staining for ALK.""")
import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline
val biomarker_pipeline = PretrainedPipeline("explain_clinical_doc_biomarker", "en", "clinical/models")
val result = biomarker_pipeline.fullAnnotate("""In the bone- marrow (BM) aspiration, blasts accounted for 88.1% of ANCs, which were positive for CD9 and CD10 on flow cytometry. Measurements of serum tumor markers showed elevated level of Cyfra21-1: 4.77 ng/mL, NSE: 19.60 ng/mL, and SCCA: 2.58 ng/mL. Immunohistochemical staining showed positive staining for CK5/6, P40, and negative staining for TTF-1 and weakly positive staining for ALK.""")
Results
# NER Result
+------------------------+-----+---+----------------+
|chunk |begin|end|label |
+------------------------+-----+---+----------------+
|positive |84 |91 |Biomarker_Result|
|CD9 |97 |99 |Biomarker |
|CD10 |105 |108|Biomarker |
|tumor markers |151 |163|Biomarker |
|elevated level |172 |185|Biomarker_Result|
|Cyfra21-1 |190 |198|Biomarker |
|4.77 ng/mL |201 |210|Biomarker_Result|
|NSE |213 |215|Biomarker |
|19.60 ng/mL |218 |228|Biomarker_Result|
|SCCA |235 |238|Biomarker |
|2.58 ng/mL |241 |250|Biomarker_Result|
|positive |289 |296|Biomarker_Result|
|CK5/6 |311 |315|Biomarker |
|P40 |318 |320|Biomarker |
|negative |327 |334|Biomarker_Result|
|TTF-1 |349 |353|Biomarker |
|weakly positive staining|359 |382|Biomarker_Result|
|ALK |388 |390|Biomarker |
+------------------------+-----+---+----------------+
# RE Result
| | sentence | entity1_begin | entity1_end | chunk1 | entity1 | entity2_begin | entity2_end | chunk2 | entity2 | relation | confidence |
|---:|-----------:|----------------:|--------------:|:-------------------------|:-----------------|----------------:|--------------:|:---------------|:-----------------|:--------------|-------------:|
| 0 | 0 | 84 | 91 | positive | Biomarker_Result | 97 | 99 | CD9 | Biomarker | is_finding_of | 0.993281 |
| 1 | 0 | 84 | 91 | positive | Biomarker_Result | 105 | 108 | CD10 | Biomarker | is_finding_of | 0.998891 |
| 2 | 1 | 151 | 163 | tumor markers | Biomarker | 172 | 185 | elevated level | Biomarker_Result | is_finding_of | 0.900508 |
| 6 | 1 | 172 | 185 | elevated level | Biomarker_Result | 190 | 198 | Cyfra21-1 | Biomarker | is_finding_of | 0.995038 |
| 9 | 1 | 190 | 198 | Cyfra21-1 | Biomarker | 201 | 210 | 4.77 ng/mL | Biomarker_Result | is_finding_of | 0.981873 |
| 10 | 1 | 190 | 198 | Cyfra21-1 | Biomarker | 218 | 228 | 19.60 ng/mL | Biomarker_Result | is_finding_of | 0.541739 |
| 14 | 1 | 213 | 215 | NSE | Biomarker | 218 | 228 | 19.60 ng/mL | Biomarker_Result | is_finding_of | 0.988173 |
| 17 | 1 | 235 | 238 | SCCA | Biomarker | 241 | 250 | 2.58 ng/mL | Biomarker_Result | is_finding_of | 0.995757 |
| 18 | 2 | 289 | 296 | positive | Biomarker_Result | 311 | 315 | CK5/6 | Biomarker | is_finding_of | 0.866368 |
| 19 | 2 | 289 | 296 | positive | Biomarker_Result | 318 | 320 | P40 | Biomarker | is_finding_of | 0.895999 |
| 26 | 2 | 327 | 334 | negative | Biomarker_Result | 349 | 353 | TTF-1 | Biomarker | is_finding_of | 0.994164 |
| 29 | 2 | 359 | 382 | weakly positive staining | Biomarker_Result | 388 | 390 | ALK | Biomarker | is_finding_of | 0.988431 |
# Classification Result
+-----------+-------------------------------------------------------------------------------------------------------------------------------------------+-----+
|sentence_ID| sentence|class|
+-----------+-------------------------------------------------------------------------------------------------------------------------------------------+-----+
| 0| In the bone- marrow (BM) aspiration, blasts accounted for 88.1% of ANCs, which were positive for CD9 and CD10 on flow cytometry.| 1|
| 1| Measurements of serum tumor markers showed elevated level of Cyfra21-1: 4.77 ng/mL, NSE: 19.60 ng/mL, and SCCA: 2.58 ng/mL.| 1|
| 2|Immunohistochemical staining showed positive staining for CK5/6, P40, and negative staining for TTF-1 and weakly positive staining for ALK.| 1|
+-----------+-------------------------------------------------------------------------------------------------------------------------------------------+-----+
Model Information
| Model Name: | explain_clinical_doc_biomarker |
| Type: | pipeline |
| Compatibility: | Healthcare NLP 5.3.0+ |
| License: | Licensed |
| Edition: | Official |
| Language: | en |
| Size: | 2.2 GB |
Included Models
- DocumentAssembler
- SentenceDetectorDLModel
- TokenizerModel
- WordEmbeddingsModel
- MedicalBertForSequenceClassification
- TextMatcherInternalModel
- MedicalNerModel
- NerConverterInternalModel
- MedicalNerModel
- NerConverterInternalModel
- ChunkMergeModel
- PerceptronModel
- DependencyParserModel
- RelationExtractionModel