Table 3.

Results of ROC curves assessing the accuracy of clinical and imaging factors in predicting IA versus AIS / MIA

variable	Area	95% CI	Sensitivity	Specificity	Cut-off vale	p value
Tumor size	0.87	(0.83,0.91)	0.86	0.72	1.35	<0.0001
Age	0.59	(0.53,0.65)	0.57	0.59	58.50	0.005816
History of family cancer	0.55	(0.49,0.61)	0.74	0.36	0.66	0.1148
CEA	0.62	(0.56,0.68)	0.54	0.65	1.97	0.00020
GGO type	0.72	(0.66,0.77)	0.71	0.73	0.66	<0.0001
Bubble-like sign	0.58	(0.53,0.64)	0.35	0.83	0.63	0.00100
Five-factor combination	0.91	(0.88,0.94)	0.82	0.87	0.69	<0.0001

Five factors including tumor size, age, serum CEA level, GGO type and Bubble-like sign