Experimental Pain Phenotypes in Individuals With Symptomatic Knee Osteoarthritis: A Neural Network–Based Clustering Approach
Pain Research and Management · 2026-01 · 2026 (1)
DOI: 10.1155/prm/5069475 ↗Kurz gefasst
Fachartikel in der Zeitschrift Pain Research and Management, online erschienen am 17.08.2026. Verfasst von Chiyoung Lee und 10 weiteren Autor:innen. Beteiligte Einrichtung: The University of Arizona College of Nursing, Tucson, Arizona, USA. Hauptthemen laut MEDLINE: Neural Networks, Computer, Osteoarthritis, Knee, Pain, Pain Threshold. Die Arbeit nennt 58 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NINR NIH HHS, NINDS NIH HHS, NIH/NINR.
Abstract (Originalsprache)
Objective
Research has emphasized the “phenotyping” of knee osteoarthritis (OA) pain as a priority to effectively target therapies to individual patients. This study identified pain phenotypes based on experimental pain responses in patients with symptomatic knee OA using machine learning approaches and examined their associations with participant characteristics.
Methods
In this cross‐sectional study, participants ( N = 208; mean age = 67.64 ± 7.51 years) completed demographic, clinical, and psychological questionnaires, followed by a multimodal quantitative sensory testing (QST) battery. For phenotyping, we implemented a two‐layer neural network–based k ‐means algorithm. Participant characteristics were compared across phenotypes using analysis of variance (ANOVA).
Results
Five phenotypes were identified, with marked differences across QST measures: (1) “high pressure pain sensitivity and impaired descending inhibition” ( n = 86), characterized by average responses on most QST measures, particularly low pressure pain threshold (PPTh) and impaired conditioned pain modulation (CPM), (2) “high heat pain sensitivity” ( n = 40), marked by the lowest heat pain thresholds and tolerances, (3) “low pain sensitivity” ( n = 46), exhibiting low sensitivity across most QST measures, (4) “high pain sensitivity and impaired pain modulation” ( n = 26), with the lowest PPTh, highest punctate mechanical pain, greatest temporal summation of mechanical pain, lowest CPM, and highest cold pain sensitivity, and (5) “pain‐resilient” ( n = 10), characterized by low sensitivity across most QST measures, along with the highest PPTh and highest CPM. Phenotypes differed significantly by sex, Kellgren–Lawrence score (index knee), pain severity (Numeric Rating Scale), and OA‐related symptoms (Western Ontario and McMaster Universities Osteoarthritis Index), particularly physical function ( p < 0.05).
Conclusions
This study identified distinct experimental pain phenotypes, each associated with unique demographic and clinical characteristics. These findings may inform the development of interventions tailored to specific pain profiles to improve pain management and functional outcomes in patients with symptomatic knee OA. Trial Registration: ClinicalTrials.gov identifier: NCT04375072
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- The University of Arizona College of Nursing, Tucson, Arizona, USA.
- Global School of Media, College of Information Technology, Soongsil University, Seoul, Republic of Korea, ssu.ac.kr.
- Department of Epidemiology and Biostatistics, The University of Arizona Mel and Enid Zuckerman College of Public Health, Tucson, Arizona, USA.
- The University of Arizona College of Medicine-Tucson, Tucson, Arizona, USA.
- Pediatric Biostatistics Core, Children's Healthcare of Atlanta and Emory University, Atlanta, Georgia, USA.
- Department of Electrical and Computer Engineering, The University of Arizona College of Engineering, Tucson, Arizona, USA.
ORCID-Profile: Chiyoung Lee, Yeri Kim, Seoyoung Kim, Juyoung Park, C Kent Kwoh, Heewon Kim, Hyochol Ahn
Förderung
- NINR NIH HHS
- NINDS NIH HHS
- NIH/NINR
- HEAL National K12 Clinical Pain Career Development Award
Zitieren
APA
Lee, C., Kim, Y., Kim, S., Park, J., Sun, X., Chen, C. X., Kwoh, C. K., Liu, S., Ahn, D., Kim, H., & Ahn, H. (2026). Experimental Pain Phenotypes in Individuals With Symptomatic Knee Osteoarthritis: A Neural Network–Based Clustering Approach. Pain Research and Management, 2026(1). https://doi.org/10.1155/prm/5069475
Vancouver
Lee C, Kim Y, Kim S, Park J, Sun X, Chen CX, et al. Experimental Pain Phenotypes in Individuals With Symptomatic Knee Osteoarthritis: A Neural Network–Based Clustering Approach. Pain Research and Management. 2026;2026(1). doi:10.1155/prm/5069475
BibTeX
@article{Lee2026,
author = {Lee, Chiyoung and Kim, Yeri and Kim, Seoyoung and Park, Juyoung and Sun, Xiaoxiao and Chen, Chen X and Kwoh, C Kent and Liu, Shen and Ahn, Dasol and Kim, Heewon and Ahn, Hyochol},
title = {Experimental Pain Phenotypes in Individuals With Symptomatic Knee Osteoarthritis: A Neural Network–Based Clustering Approach},
journal = {Pain Research and Management},
year = {2026},
volume = {2026},
number = {1},
doi = {10.1155/prm/5069475}
}
EUVIMED