Skip to main content

Table 4 Validation of the optimal cutoff value generalizability using bootstrap resampling

From: Peripheral blood lymphocytes influence human papillomavirus infection and clearance: a retrospective cohort study

Cutoff values

Bootstrap

HR (95%CI)

C-index

Bootstrap

C-index

Bootstrap

AUC

T cellsp (%)

    

<67.39

1

   

≥ 67.39

1.09(1.01–1.14) *

0.61

0.61

0.76

NK cellsp (%)

    

<22.65

1

   

≥ 22.65

0.91(0.85–0.99) *

0.60

0.60

0.69

CD8 + T cells Model RS

    

<0.95

1

   

≥ 0.95

6.03(1.59–29.03) *

0.63

0.68

0.64

  1. Calculated by univariate Cox regression, receiver operating characteristic analysis and bootstrapping
  2. Bootstrap HR (95% CI): average hazard ratio across 1,000 iterations and average 95% confidence interval for hazard ratio across 1,000 iterations
  3. C-index: concordance index of the original data Cox model
  4. Bootstrap C-index: average concordance index across 1,000 iterations
  5. Bootstrap AUC: average area under the curve across 1,000 iterations
  6. * Statistically significant variables: Variables with a mean hazard ratio (HR) and corresponding mean 95% confidence interval (CI) that are entirely between 0 and 1 or greater than 1
  7. p Lymphocyte percentage value
  8. RS: Risk score, CD8 + T-cell model risk score
Back to article page

Virology Journal

ISSN: 1743-422X