Classification of Noisy Free-Text Prostate Cancer Pathology Reports Using Natural Language Processing

Abstract

2 Gleason score 3+4 Gleason grade: a) Primary pattern 3/5 b) Secondary pattern 4/5 a) Total Gleason score 7/10 Swillens, J. E. M., et al. "Identification of barriers and facilitators in nationwide implementation of standardized structured reporting in pathology: a mixed method study." Snoek, Annefleure, et al. "The impact of standardized structured reporting of pathology reports for breast cancer in the Netherlands.

Deep bidirectional LSTM for disease classification supporting hospital admission based on pre-diagnosis: a case study in Vietnam

“…In contrast, the second method was based on the Convolutional Neural Network (CNN) design. In [ 21 ], the authors showed the feasibility and reliability of using paragraph vectors based on natural language processing models to present and classify pathology reports prostate into high-grade and low-grade prostate cancer. The authors in [ 22 ] presented algorithms for early detection and classification of Alzheimer’s disease based on the bag of visual word approach and clinical features.…”

TCGA-Reports: A machine-readable pathology report resource for benchmarking text-based AI models

“…Another study combined OCR and traditional machine learning techniques to classify tumor grade using a TCGA subset of approximately 500 patients with prostate cancer. 22 Additionally, Allada et al. compared different NLP classification methods for the prediction of seven disease classes within a TCGA subset consisting of roughly 2,000 patients.…”

Modelling digital health data: The ExaMode ontology for computational pathology

“… 19 , 20 Currently, this process is often manual and time-consuming due to the high volumes of data and the lack of a standard shared terminology and structure between institutions. 21 , 22 Indeed, automatic knowledge extraction algorithms need a common and machine-readable reference point to understand the concepts identified in the text. To this end, an ontology for digital pathology can overcome data heterogeneity and integration problems as it provides a shared terminology that can be used to produce high-quality annotated image datasets.…”