BMC Medical Imaging

Table 2 The value of Different Radiomics Models in Training cohort and Validation cohort

From: Radiomics analysis based on CT for the prediction of pulmonary metastases in ewing sarcoma

				Train						Validation
	AUC	ACC	95%CI	Sensitivity	Specificity		AUC	ACC	95%CI	Sensitivity	Specificity
Pre-CT
GP	0.865	0.798	0.805–0.914	0.803	0.792		0.805	0.793	0.635–0.924	0.867	0.714
LR	0.842	0.781	0.779–0.895	0.754	0.811		0.800	0.689	0.635–0.913	0.667	0.714
PLS-DA	0.834	0.771	0.769–0.891	0.721	0.830		0.819	0.724	0.656–0.928	0.733	0.714
CTE
GP	0.831	0.754	0.765–0.890	0.754	0.736		0.824	0.689	0.657–0.948	0.733	0.786
LR	0.834	0.789	0.761–0.891	0.787	0.792		0.833	0.759	0.678–0.944	0.800	0.714
PLS-DA	0.796	0.737	0.723–0.856	0.738	0.736		0.800	0.586	0.637–0.919	0.733	0.857
ComB
GP	0.868	0.781	0.808–0.917	0.705	0.868		0.829	0.689	0.678–0.933	0.800	0.786
LR	0.866	0.798	0.807–0.913	0.852	0.736		0.852	0.724	0.709–0.956	0.800	0.857
PLS-DA	0.845	0.754	0.781–0.898	0.721	0.755		0.848	0.793	0.697–0.957	0.800	0.857

Note. AUC area under the receiver-operating characteristic curve, ACC accuracy, GP gaussian process, LR logistic regression, PLS-DA partial least-squares discrimination analysis, CT computed tomography, CTE CT-enhanced, ComB, the combination of CT and CTE features

Back to article page

ISSN: 1471-2342

Contact us

General enquiries: journalsubmissions@springernature.com