Skip to main content

Table 8 Stabilized multivariate cox regression using bootstrapping and ridge penalization for key biomarkers

From: Prognostic and diagnostic value of PVR gene and protein levels, serum amylase, and urinary IGFBP-7 and TIMP-2 biomarkers in multiple myeloma

Parameter

Original HR (95% CI)

Bootstrapped HR (95% CI)*

Ridge Regression HR (95% CI)*

p-value

PVR Gene (High)

12.2 (1.03–361.2)

11.8 (1.2–45.1)

10.9 (1.1–42.3)

0.042

Serum Amylase

11.5 (1.36–41.6)

10.7 (1.4–39.8)

11.1 (1.3–40.2)

0.038

IGFBP-7 (High)

11.9 (2.13–321.6)

12.1 (2.3–58.4)

11.5 (2.1–55.6)

0.041

  1. Bootstrapped hazard ratios were calculated using 1,000 iterations. Ridge regression was applied to penalize model complexity and stabilize estimates in the context of a small sample with limited survival events. HR Hazard Ratio, CI Confidence Interval