Table 5.

Comparison results of different models.

Input data type	Models	Acc	AUC
Clinical data only	Clinical model (32)	76.03% ± 10.0	0.753 ± 0.13
Image data only	Radiomics model (48)	67.04% ± 4.9	0.640 ± 0.03
	WANG et al.'s Model (32)	69.5% ± 5.1	0.723 ± 0.06
	SENet-like model (34)	64.5% ± 3.4	0.675 ± 0.06
	CBAM-like model (35)	66.7% ± 4.2	0.684 ± 0.05
	Propose DPA model	70.7% ± 2.4	0.747 ± 0.03
Clinical data and Image data	WANG et al.'s fusion model (33)	80.49% ± 4.3	0.833 ± 0.03
	Propose DPA fusion model	81.2% ±1.3	0.869 ±0.03

Acc, accuracy; AUC, area under the receiver operating characteristics curve. The bold values indicate P values < 0.05 are highlighted. The bold values indicate the best results.