Cascaded-Recalibrated Multiple Instance Deep Model for Pathologic-Level Lung Cancer Prediction in CT Images

Table 1

Predictive performance of our attribute-specific models and the radiologists’ on nine semantic attributes, evaluated by mean absolute error (MAE). The abbreviation “IB,” which indicates the interobserver variation, is calculated based on all possible pairs of scores given by radiologists as in [23, 27].

MAE	Sub	Int	Cal	Sph	Mar	Lob	Spi	Tex	Mal
IB	0.70	0.02	0.20	0.88	0.67	0.83	0.67	0.22	0.90
Attribute-specific model	0.65	0.09	0.24	0.51	0.54	0.51	0.42	0.41	0.44

The bold values mean the superiority to others.