A Simple Tool to Predict Genetic Focal Segmental Glomerulosclerosis

Introduction Genetic focal segmental glomerulosclerosis (GFSGS) is caused by pathogenic variant. In the present study, we aimed to develop and validate a predictive model for pathogenic variant in FSGS patients. Methods Patients with biopsy-proven FSGS from two independent cohorts were recruited. FSGS secondary to obesity, hypertension, etc. were excluded. All the enrolled patients underwent whole exome sequencing (WES). We developed a predictive model for pathogenic variants using multivariate logistic regression in the development cohort, and validated the performance of the model using ROC analysis and calibration curve analysis in the validation cohort. Results We recruited 197 FSGS patients for the development cohort and 155 patients for the validation cohort. In the development cohort, 70 patients had a family history and 127 patients did not have a family history; In the validation cohort, 70 patients had a family history and 85 patients did not. WES was performed on all the patients from the two cohorts. We identified 65 FSGS patients with pathogenic variants (33%) in the development cohort and 83 (53.5%) in the validation cohort. Using multivariate logistic regression, we established a predictive model for pathogenic variants, which included parameters such as male sex, lower eGFR at renal biopsy (≤ 30 ml/min/1.73 m ² ), non-nephrotic syndrome and dominant inheritance [R ²  = 0.62; AUC (95%CI) = 0.91 (0.86–0.96)). We validated the model in the validation cohort, which demonstrated a good performance with an R ² of 0.55 and AUC of 0.93 (95%CI: 0.89–0.98); In the calibration analysis, the predictive model showed a close alignment between predicted and observed risks of GFSGS in the validation cohort (R ²  = 0.96). Based on the predictive model, we established a simple tool that provided estimated risk of pathogenic variant. Conclusion We developed and validated a simple tool including four variables which had good performance for GFSGS prediction.