Article Text

Download PDFPDF

OP0190 A MACHINE LEARNING MODEL THAT PREDICTS RA PROGRESSION FROM UNDIFFERENTIATED ARTHRITIS -KURAMA AND ANSWER COHORT STUDY-
Free
  1. T. Fujii1,2,
  2. K. Murata2,3,
  3. H. Onizawa3,
  4. A. Onishi3,
  5. K. Murakami4,
  6. M. Tanaka3,
  7. W. Yamamoto5,
  8. K. Nagai6,
  9. A. Yoshikawa6,
  10. Y. Etani7,
  11. Y. Okita8,
  12. N. Yoshida9,10,
  13. H. Amuro10,
  14. T. Okano11,
  15. Y. Ueda12,
  16. R. Hara13,
  17. M. Hashimoto9,
  18. T. Okano12,
  19. A. Morinobu14,
  20. S. Matsuda2
  1. 1Kyoto University, Department of Advanced Medicine for Rheumatic Diseases, Kyoto, Japan
  2. 2Kyoto University, Department of Orthopaedic Surgery, Kyoto, Japan
  3. 1Kyoto University, Department of Advanced Medicine for Rheumatic Diseases, Kyoto, Japan
  4. 4Kyoto University, Center for Cancer Immunotherapy and Immunobiology, Kyoto, Japan
  5. 5Kurashiki Sweet Hospitap, Department of Health Information Management, Kurashiki, Japan
  6. 6Osaka Medical and Pharmaceutical University, Department of Internal Medicine (Ⅳ), Takatsuki, Japan
  7. 7Osaka University, Department of Orthopaedic Surgery, Suita, Japan
  8. 8Osaka University, Department of Respiratory Medicine and Clinical Immunology, Suita, Japan
  9. 9Osaka Metropolitan University, Department of Clinical Immunology, Osaka, Japan
  10. 10Kansai Medical University, First Department of Internal Medicine, Hirakata, Japan
  11. 11Osaka Metropolitan University, Department of Orthopedic Surgery, Osaka, Japan
  12. 12Kobe University Graduate School of Medicine, Department of Rheumatology and Clinical Immunology, Kobe, Japan
  13. 13Nara Medical University, Department of Orthopaedic Surgery, Kashihara, Japan
  14. 14Kyoto University, Department of Rheumatology and Clinical Immunology, Kyoto, Japan

Abstract

Background Early diagnosis and treatment of rheumatoid arthritis (RA) improve clinical outcomes. Undifferentiated arthritis (UA) is arthritis that does not fit a specific diagnosis. Half of the UA undergo spontaneous remission, while 30% of cases develop RA. Therefore, in UA, identifying patients at high risk for developing RA and providing close monitoring for those patients is required for early diagnosis and treatment [1]). However, predicting the evolution of UA to RA is still difficult.

Objectives Machine learning, including deep learning, which is comparable to and in some cases surpasses the performance of human experts, is broadening its application in medicine. This study aims to build a machine-learning model that predicts the development of UA to RA.

Methods For model training, a total of 322 UA patients in KURAMA cohort were analyzed (Table 1). For variables to train models, we chose 24 clinical features, which are easy to obtain in daily clinical practice. The target variable was the final diagnosis. We built models using Random forest (RF), XGBoost (XGB), Logistic regression (LR), and Deep neural network (DNN) and compared their performances. For model validation, we used data of 88 UA cases in ANSWER cohort (Table 1).

Results We trained models using 24 clinical parameters at the first clinical visit, performed 10-fold cross-validation, and evaluated model performance by averaging accuracy and AUC. The performance of the models was 73.5%, 74.2%, 74.5%, and 85.1% in precision and 0.760, 0.734, 0.748, and 0.895 in AUC for RF, XGB, LR, and DNN, respectively. DNN showed the highest performance. We then applied the DNN model to external validation data from ANSWER cohort and found that the prediction accuracy was 80.0%.

Conclusion Using parameters available in clinical practice, we developed a DNN model that effectively predicted RA development in internal and external UA datasets. Applying a machine learning approach might enable identifying patients at high risk of RA progression and improve the clinical management of UA patients.

Reference [1] de la Calle-Fabregat C, Niemantsverdriet E, Cañete JD, Li T, van der Helm-van Mil AHM, Rodríguez-Ubreva J, Ballestar E. Prediction of the Progression of Undifferentiated Arthritis to Rheumatoid Arthritis Using DNA Methylation Profiling. Arthritis Rheumatol. 2021 Dec;73(12):2229-2239. doi: 10.1002/art.41885. Epub 2021 Nov 2. PMID: 34105306.

Table 1.

Baseline patients’ characteristics

Acknowledgements I have no acknowledgments to declare.

Disclosure of Interests Takayuki Fujii Speakers bureau: Asahi Kasei Pharma, Abbvie, Jansen, Tanabe Mitsubishi Pharma, and Eisai., Koichi Murata Speakers bureau: AbbVie GK; Eisai Co., Ltd., Chugai.

Pharmaceutical Co., Ltd.; Mitsubishi Tanabe Pharma Corporation; Pfizer Inc.; Bristol-Myers Squibb;

Asahi Kasei Pharma Corp., Hideo Onizawa Speakers bureau: AbbVie, Asahi Kasei, Astellas Pharma Inc.,

Eisai Co. Ltd., Janssen Pharmaceutical K.K., Mitsubishi Tanabe Pharma Corporation, and Daiichi.

Sankyo Co. Ltd., Akira Onishi Speakers bureau: Pfizer Inc., Bristol-Myers Squibb., Advantest,

Asahi Kasei Pharma Corp., Chugai Pharmaceutical Co. Ltd., Eli Lilly Japan K.K, Ono Pharmaceutical.

Co., UCB Japan Co., Mitsubishi Tanabe Pharma Co., Eisai Co. Ltd., Abbvie Inc., Takeda.

Pharmaceutical Co. Ltd., and Daiichi Sankyo Co. Ltd., Kosaku Murakami Speakers bureau: Eisai Co. Ltd, Chugai.

Pharmaceutical Co. Ltd., Pfizer Inc., Bristol-Myers Squibb, Mitsubishi Tanabe Pharma Corporation,

UCB Japan Co. Ltd, Daiichi Sankyo Co. Ltd., and Astellas Pharma Inc., Masao Tanaka Speakers bureau: AbbVie GK, Asahi Kasei Pharma.

Corporation, Astellas Pharma Inc., Chugai Pharmaceutical Co., Ltd., Daiichi Sankyo Co., Ltd., Eisai.

Co., Ltd., Eli Lilly and Company, Pfizer Inc., UCB Japan Co., Ltd., Janssen Pharmaceutical K.K.,

Kyowa Kirin Co., Ltd., Mitsubishi Tanabe Pharma Corporation, Taisho Pharma Co., Ltd, and Teijin.

Pharma, Ltd., Wataru Yamamoto: None declared, Koji Nagai: None declared, Ayaka Yoshikawa: None declared, Yuki Etani: None declared, Yasutaka Okita Speakers bureau: Chugai Pharmaceutical, Pfizer, and Ono Pharmaceutical., Naofumi Yoshida: None declared, Hideki Amuro: None declared, Tadashi Okano Speakers bureau: Abbvie, Chugai, Eli Lilly, Janssen and Novartis Pharma., Grant/research support from: Abbvie, Asahi Kasei, Chugai, Eisai,

Eli Lilly and Tanabe Mitsubishi., Yo Ueda: None declared, Ryota Hara Speakers bureau: AbbVie, Eisai, Motomu Hashimoto Speakers bureau: Abbvie, Asahi Kasei, Astellas, Ayumi, Bristol.

Meyers, Chugai, EA Pharma, Eisai, Daiichi Sankyo, Eli Lilly, Nihon Shinyaku, Novartis Pharma,

Tanabe Mitsubishi., Takaichi Okano: None declared, Akio Morinobu Speakers bureau: AbbVie G.K., Chugai Pharmaceutical Co. Ltd., Eli Lilly Japan.

K.K., Eisai Co. Ltd., Pfizer Inc., Bristol-Myers Squibb., Mitsubishi Tanabe Pharma Co., Astellas.

Pharma Inc., and Gilead Sciences Japan, Grant/research support from: AbbVie G.K., Asahi.

Kasei Pharma Corp., Chugai Pharmaceutical Co. Ltd., Mitsubishi Tanabe Pharma Co. and Eisai Co. Ltd., Shuichi Matsuda Speakers bureau: Astellas Pharma Inc., Daiichi Sankyo Co.

Ltd., Mitsubishi Tanabe Pharma Corporation, Eisai Co., Ltd., Takeda Pharmaceutical Company Limited,

Chugai Pharmaceutical Co. Ltd, Pfizer Inc., and Asahi Kasei Corporation.

  • Rheumatoid arthritis
  • Undifferentiated connective tissue disease
  • Real-world evidence

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.