Table 7 Optimal threshold values for the LogIgG and corresponding anti-S1 IgG titers based on the combined QIC approach.

From: Cracking the code of a correlate of protection against SARS-CoV-2 breakthrough infection in cancer patients

Period

Optimal threshold (95% CI)

Sensitivity (95% CI)

Specificity (95% CI)

Threshold before and after period H (log10) (95% CI)

Threshold (BAU/mL) (95% CI)

Sensitivity (95% CI)

Specificity (95% CI)

E

2.332 (0.754, 3.171)

0.846

(0.563, 0.966)

0.525

(0.484, 0.567)

1.310

(1.190, 2.128)

20.420

(15.488, 134.122)

0.538

(0.292, 0.767)

0.796

(0.760, 0.827)

F

1.198

(1.038, 3.305)

0.625

(0.304, 0.862)

0.785

(0.743, 0.822)

1.310

(1.190, 2.128)

20.420

(15.488, 134.122)

0.625

(0.304, 0.862)

0.758

(0.715, 0.797)

G

2.144

(2.027, 3.84)

0.374

(0.285, 0.472)

0.862

(0.823, 0.892)

1.310

(1.190, 2.128)

20.420

(15.488, 134.122)

0.242

(0.169, 0.336)

0.897

(0.863, 0.924)

H

3.379

(1.930, 4119)

0.741

(0.550, 0.869)

0.392

(0.330, 0.458)

3.640

(3.220, 4.010)

4365.160

(1659.587, 10232.930)

0.815

(0.627, 0.921)

0.261

(0.208, 0.323)

I

3.634

(2.494, 4.125)

0.731

(0.536, 0.864)

0. 627

(0.573, 0.578)

3.640

(3.220, 4.010)

4365.160

(1659.587, 10232.930)

0.731

(0.536, 0.864)

0.627

(0.573, 0.678)

  

Overall performance

0.445

(0.373, 0.520)

0.717

(0.697, 0.736)

  1. Presentation of the optimal threshold values for the log-transformed SARS-CoV-2 anti-S1 IgG values and corresponding anti-S1-IgG antibody titers based on the combined QIC approach when modelling the probability of SARS-CoV-2 breakthrough infection in relation to antibody titers and time period (i.e., time period E (reference) up to time period I), though maximizing the Youden index for each of the periods separately. Different colors are used to indicate the periods for which the same threshold is assumed. For the column “Threshold before and after period H” we explicitly estimate threshold values in the period before and after circulation and dominance of Omicron infections in the population. The 95% CIs for sensitivity and specificity are Wilson Score intervals for proportions. For the thresholds, we rely on bootstrap-based 95% percentile CIs based on 1000 nonparametric bootstrap samples.
Back to article page