Background: Urolithiasis, a multifactorial disease with high recurrence, lacks robust tools for personalized risk prediction. While genetic susceptibility and biochemical signatures are implicated in pathogenesis, their combined utility remains unexplored. This study aims to evaluate the integration of polygenic risk scores (PRS) and biomarker-derived clusters in personalized urolithiasis risk stratification across ethnically diverse populations. Materials and methods: A population-based cohort study was performed using UK Biobank (UKB) data ( N = 480 098). Sixteen out of 62 hematology and biochemistry markers related to urolithiasis were identified by LASSO regression. K-means clustering categorized participants into distinct biomarker profiles. Findings were validated in a Hong Kong cohort ( N = 6177) from the electronic health record database of Hospital Authority (EHR-HK). Results: In the UKB, individuals with the top PRS quartile had a 41% higher urolithiasis risk (HR = 1.41, 95% CI: 1.34–1.48). Three biomarker-derived clusters were identified, namely cardiovascular-skeletal (C-S), hematal-endocrine (H-E), and inflammatory-metabolic (I-M) clusters. Compared to the C-S cluster, H-E and I-M clusters were associated with 78 and 80% increased risks of urolithiasis. The highest cumulative recurrence rate was observed in patients with I-M cluster, with 19.7, 25.4, and 28.7% at 5, 10, and 15 years after the first episode. Synergistic effects were significant: those with the top PRS quartile combined with the I-M profile faced 2.66-, 3.56-, and 2.94-fold higher incidence, recurrent, and multifocal risk of urolithiasis. Site-specific analyses revealed ureter calculi attributable to the strongest genetic-biomarker interaction (HR = 3.73, 95% CI: 3.24–4.30). Validation in EHR-HK confirmed the utility of biomarker-derived clusters in predicting recurrent urolithiasis OR = 1.58 (1.35–1.85) for I-M cluster; OR = 1.31 (1.13–1.52) for H-E cluster. Conclusions: This cross-cohort study demonstrates that integration of genetic and biomarker profiles identifies high-risk phenotypes related to stone formation. Our framework informs clinicians to prioritize high-risk patients for targeted monitoring and prophylactic therapies. The replication across European and East Asian highlights its utility in diverse settings.
Zhan et al. (Mon,) studied this question.