Abstract: TH-PO867
New Surrogate Marker of CKD Progression and Mortality in Medical Word Virtual Space: Prospective Cohort Study
Session Information
- CKD: Epidemiology, Risk Factors, Prevention - I
November 03, 2022 | Location: Exhibit Hall, Orange County Convention Center‚ West Building
Abstract Time: 10:00 AM - 12:00 PM
Category: CKD (Non-Dialysis)
- 2201 CKD (Non-Dialysis): Epidemiology‚ Risk Factors‚ and Prevention
Authors
- Kanda, Eiichiro, Kawasaki Ika Daigaku, Kurashiki, Okayama, Japan
- Epureanu, Bogdan I., University of Michigan, Ann Arbor, Michigan, United States
- Adachi, Taiji, Kyoto Daigaku, Kyoto, Japan
- Kashihara, Naoki, Kawasaki Ika Daigaku, Kurashiki, Okayama, Japan
Background
Chronic kidney disease (CKD) leads to end-stage renal disease (ESRD) or death. A new surrogate marker reflecting its pathophysiology has been needed for CKD therapy.
Methods
In this study, we developed a virtual space unifying data in the medical literature and that of actual CKD patients and created a surrogate marker of CKD progression and mortality using natural language processing.
Results
A virtual space of medical words was constructed from the CKD-related literature (n=165,271) using natural language processing, in which CKD-related words (n=106,612) composed a network (Figure 1). The data of CKD patients of a prospective cohort study for three years (n=26,433) were transformed into the space and linked with the network on the basis of information-geometry theory. We let the relationship between a patient and the outcome (ESRD or death) in the network be a surrogate marker of the outcome. The network satisfied the definitions of vector keeping their medical meanings. Riemannian metrics highly accurately predicted the primary outcomes; C-statistics, 0.911. Cox proportional hazards models with spline showed that the high Riemannian metrics were associated with high hazard ratio of the primary outcomes (p<0.0001). Moreover, the risk of the primary outcome in high-Riemannian-metric group was 21.92 (95% CI: 14.77, 32.51) times higher than that in the low-Riemannian-metric group.
Conclusion
The medical-word virtual space reflects the real-world patient data. And the Riemannian metrics between a patient and the outcome is a new surrogate marker for CKD therapy.