Skip to main content
. Author manuscript; available in PMC: 2015 Aug 1.
Published in final edited form as: J Thorac Oncol. 2014 Aug;9(8):1111–1119. doi: 10.1097/JTO.0000000000000235

Table 2.

Performance of Clinical & HD-CTM Models

NSCLC vs. Benign Stage I vs. Benign

Training n = 88 Test n = 41 All n = 129 Training n = 71 Test n = 34 All n=105
Clinical Variables Only* AUC 0.78 (0.68–0.89) 0.75 (0.53–0.97) 0.77 (0.68–0.87) 0.80 (0.68–0.91) 0.76 (0.54–0.97) 0.79 (0.68–0.87)
Sens 0.65 0.91 0.71 0.83 0.85 0.68
Spec 0.88 0.63 0.80 0.71 0.63 0.84

Clinical & HD-CTM AUC 0.88 (0.81–0.96) 0.88 (0.66–0.99) 0.88 (0.82–0.94) 0.88 (0.79–0.97) 0.88 (0.75–1.00) 0.87 (0.82–0.94)
Sens 0.79 0.72 0.82 0.82 0.78 0.83
Spec 0.88 1.00 0.84 0.88 1.00 0.84

AUC = Area under curve (C-statistic) with 95% confidence interval. Sens = Sensitivity; Spec = Specificity.

*

See Table 1 for variables and levels included in the clinical model.

See Supplemental Table 2 for significant variables in each model. Models that significantly add value to the clinical model are in bold.