NLP Analysis of COVID-19 Radiology Reports in Indonesian using IndoBERT

Nunung Nurul Qomariyah; Tianda Sun; Dimitar Lubomirov Kazakov

doi:10.1109/IBIOMED56408.2022.9988223

NLP Analysis of COVID-19 Radiology Reports in Indonesian using IndoBERT

Nunung Nurul Qomariyah, Tianda Sun, Dimitar Lubomirov Kazakov^*

^*Corresponding author for this work

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The presence of COVID-19, a respiratory disease, can be detected through medical imaging, such as Chest X-Ray (CXR) and Computed Tomography (CT) scans. These radiology images can also show how the patient's condition progresses. Radiologists need to provide a written report for each image, so that other clinicians can use it in their decision making. In this study, we applied one of the Natural Language Processing (NLP) models called IndoBERT to analyze radiology reports of COVID-19 patients written in Indonesian. We performed two tasks, clustering to group reports by meaning and understand their content, and text classification to predict one of the five possible outcomes for each patient. We show the most frequent topics in radiology reports, and word scores in each topic. The IndoBERT model was fine tuned on a medical text, ‘Kamus Kedokteran Dorland’ in an attempt to further improve it. This proved unnecessary: on one hand, there were no additional benefits, on the other, the standard model alone achieved a very satisfactory classification accuracy of over 90%.

Original language	English
Title of host publication	Proceedings of the 4th International Conference on Biomedical Engineering (IBIOMED)
Place of Publication	Yogakarta, Indonesia
Publisher	IEEE
Pages	65-70
Number of pages	6
ISBN (Electronic)	9781665460798/22/
DOIs	https://doi.org/10.1109/IBIOMED56408.2022.9988223
Publication status	Published - 18 Oct 2022
Event	4th International Conference on Biomedical Engineering - Yogakarta, Indonesia Duration: 18 Oct 2022 → 19 Oct 2022 http://ibiomed.ugm.ac.id/2022/

Conference

Conference	4th International Conference on Biomedical Engineering
Abbreviated title	IBIOMED
Country/Territory	Indonesia
City	Yogakarta
Period	18/10/22 → 19/10/22
Internet address	http://ibiomed.ugm.ac.id/2022/

Keywords

Natural Language Processing
IndoBERT
COVID-19
MEDICAL OUTCOMES
Radiology

Access to Document

10.1109/IBIOMED56408.2022.9988223

Cite this

@inproceedings{75fae20631604da385c22f58ca91a4ce,

title = "NLP Analysis of COVID-19 Radiology Reports in Indonesian using IndoBERT",

abstract = "The presence of COVID-19, a respiratory disease, can be detected through medical imaging, such as Chest X-Ray (CXR) and Computed Tomography (CT) scans. These radiology images can also show how the patient's condition progresses. Radiologists need to provide a written report for each image, so that other clinicians can use it in their decision making. In this study, we applied one of the Natural Language Processing (NLP) models called IndoBERT to analyze radiology reports of COVID-19 patients written in Indonesian. We performed two tasks, clustering to group reports by meaning and understand their content, and text classification to predict one of the five possible outcomes for each patient. We show the most frequent topics in radiology reports, and word scores in each topic. The IndoBERT model was fine tuned on a medical text, {\textquoteleft}Kamus Kedokteran Dorland{\textquoteright} in an attempt to further improve it. This proved unnecessary: on one hand, there were no additional benefits, on the other, the standard model alone achieved a very satisfactory classification accuracy of over 90%.",

keywords = "Natural Language Processing, IndoBERT, COVID-19, MEDICAL OUTCOMES, Radiology",

author = "Qomariyah, {Nunung Nurul} and Tianda Sun and Kazakov, {Dimitar Lubomirov}",

year = "2022",

month = oct,

day = "18",

doi = "10.1109/IBIOMED56408.2022.9988223",

language = "English",

pages = "65--70",

booktitle = "Proceedings of the 4th International Conference on Biomedical Engineering (IBIOMED)",

publisher = "IEEE",

address = "United States",

note = "4th International Conference on Biomedical Engineering, IBIOMED ; Conference date: 18-10-2022 Through 19-10-2022",

url = "http://ibiomed.ugm.ac.id/2022/",

}

Qomariyah, NN, Sun, T & Kazakov, DL 2022, NLP Analysis of COVID-19 Radiology Reports in Indonesian using IndoBERT. in Proceedings of the 4th International Conference on Biomedical Engineering (IBIOMED). IEEE, Yogakarta, Indonesia, pp. 65-70, 4th International Conference on Biomedical Engineering, Yogakarta, Indonesia, 18/10/22. https://doi.org/10.1109/IBIOMED56408.2022.9988223

TY - GEN

T1 - NLP Analysis of COVID-19 Radiology Reports in Indonesian using IndoBERT

AU - Qomariyah, Nunung Nurul

AU - Sun, Tianda

AU - Kazakov, Dimitar Lubomirov

PY - 2022/10/18

Y1 - 2022/10/18

N2 - The presence of COVID-19, a respiratory disease, can be detected through medical imaging, such as Chest X-Ray (CXR) and Computed Tomography (CT) scans. These radiology images can also show how the patient's condition progresses. Radiologists need to provide a written report for each image, so that other clinicians can use it in their decision making. In this study, we applied one of the Natural Language Processing (NLP) models called IndoBERT to analyze radiology reports of COVID-19 patients written in Indonesian. We performed two tasks, clustering to group reports by meaning and understand their content, and text classification to predict one of the five possible outcomes for each patient. We show the most frequent topics in radiology reports, and word scores in each topic. The IndoBERT model was fine tuned on a medical text, ‘Kamus Kedokteran Dorland’ in an attempt to further improve it. This proved unnecessary: on one hand, there were no additional benefits, on the other, the standard model alone achieved a very satisfactory classification accuracy of over 90%.

AB - The presence of COVID-19, a respiratory disease, can be detected through medical imaging, such as Chest X-Ray (CXR) and Computed Tomography (CT) scans. These radiology images can also show how the patient's condition progresses. Radiologists need to provide a written report for each image, so that other clinicians can use it in their decision making. In this study, we applied one of the Natural Language Processing (NLP) models called IndoBERT to analyze radiology reports of COVID-19 patients written in Indonesian. We performed two tasks, clustering to group reports by meaning and understand their content, and text classification to predict one of the five possible outcomes for each patient. We show the most frequent topics in radiology reports, and word scores in each topic. The IndoBERT model was fine tuned on a medical text, ‘Kamus Kedokteran Dorland’ in an attempt to further improve it. This proved unnecessary: on one hand, there were no additional benefits, on the other, the standard model alone achieved a very satisfactory classification accuracy of over 90%.

KW - Natural Language Processing

KW - IndoBERT

KW - COVID-19

KW - MEDICAL OUTCOMES

KW - Radiology

U2 - 10.1109/IBIOMED56408.2022.9988223

DO - 10.1109/IBIOMED56408.2022.9988223

M3 - Conference contribution

SP - 65

EP - 70

BT - Proceedings of the 4th International Conference on Biomedical Engineering (IBIOMED)

PB - IEEE

CY - Yogakarta, Indonesia

T2 - 4th International Conference on Biomedical Engineering

Y2 - 18 October 2022 through 19 October 2022

ER -