Skip to main content
. Author manuscript; available in PMC: 2010 Apr 1.
Published in final edited form as: Clin Endocrinol (Oxf). 2008 Aug 21;70(4):606–614. doi: 10.1111/j.1365-2265.2008.03389.x

Table 4.

Results from multivariable prediction models of tumours—cut-points, model fit statistics, and receiver operating characteristic curve statistics among individuals in the study population with complete data on all considered factors (n=39).

Factors(#) Model Factors in model Cut-point c-statistic Correctly predicted Sensitivity (95% CI) Specificity (95% CI) False Positives False Negatives
0.943 85.7 % 78.6 % 88.9 % 3 3
1 A 1. Peripheral T (nmol/L) 4.51 (49.2–95.3%) (70.8–97.6%)
2 B 1. Peripheral T (nmol/L) 4.51 0.950 92.7 % 92.9 %
(66.1–99.8%)
92.6 %
(75.7–99.1%)
2 1
2. Peripheral F (nmol/L) 455.24
3 C 1. Peripheral T (nmol/L) 4.51 0.959 92.7 % 92.9 %
(66.1–99.8%)
92.6 %
(75.7–99.1%)
2 1
2. Peripheral F (nmol/L) 455.24
3. Peripheral A (nmol/L) 14.48
2+* D 1. Peripheral T ≥4.51 nmol/L ~ 0.982 97.6 % 100 %
(76.8–100.0%)
96.4 %
(81.0–99.9%)
0 1
2. Peripheral F >281.42 nmol/L ~
3. Interaction
*

best-fit model