Prenatal prediction of neonatal respiratory morbidity: a radiomics method based on imbalanced few-shot fetal lung ultrasound images

Table 5 The classification performance of RUSBoost with different features on the original imbalanced few-shot dataset

Feature	Training set (mean ± std)						Test set
	bACC	AUC	SENS	SPEC	PPV	NPV	bACC	AUC	SENS	SPEC	PPV	NPV
GA	0.72 ± 0.11	0.80 ± 0.10	0.58 ± 0.21	0.88 ± 0.06	0.60 ± 0.11	0.87 ± 0.08	0.71	0.97	0.45	0.97	0.83	0.84
GA & GDM	0.71 ± 0.15	0.83 ± 0.10	0.72 ± 0.24	0.69 ± 0.15	0.42 ± 0.14	0.89 ± 0.15	0.66	0.83	0.64	0.68	0.41	0.85
Radiomics features extracted from the irregular ROI
GA, GDM & Radiomics	0.77 ± 0.08	0.83 ± 0.13	0.72 ± 0.05	0.82 ± 0.12	0.74 + 0.02	0.82 ± 0.12	0.83	0.87	0.82	0.84	0.64	0.93
Radiomics features extracted from the square ROI
GA, GDM & Radiomics	0.76 ± 0.08	0.81 ± 0.09	0.70 ± 0.07	0.82 ± 0.15	0.55 ± 0.04	0.90 ± 0.09	0.87	0.89	0.91	0.82	0.63	0.96

The best results of each metric are shown in bold, and the worst results are shown in italics. Performance evaluation results obtained by bootstrap K-fold cross-validation in the training set.

ISSN: 1471-2342