Prediction of radiosensitivity and radiocurability using a novel supervised artificial neural network

Zeng, Zihang; Luo, Maoling; Li, Yangyi; Li, Jiali; Huang, Zhengrong; Zeng, Yuxin; Yuan, Yu; Wang, Mengqin; Liu, Yuying; Gong, Yan; Xie, Conghua

doi:10.1186/s12885-022-10339-3

BMC Cancer

Table 2 Comparison of predictive performance of models for prognosis in TCGA patients with RT

From: Prediction of radiosensitivity and radiocurability using a novel supervised artificial neural network

Cohorts		ANN-SCGP		LASSO CPH		ANN		CPH		RSF		RSI
Cohorts		C-index (CV1)	C-index (CV2)	C-index (CV1)	C-index (CV2)	C-index (CV1)	C-index (CV2)	C-index (CV1)	C-index (CV2)	C-index (CV1)	C-index (CV2)	C-index (CV1)	C-index (CV2)
LGG	Training	0.958	0.964	0.869	0.840	0.870	0.903	1.000	1.000	0.930	0.944	0.675	0.658
LGG	Testing	0.787	0.815	0.825	0.794	0.753	0.716	OP	OP	0.839	0.798	0.658	0.675
HNSC	Training	0.928	0.937	0.790	0.820	0.770	0.784	1.000	1.000	0.916	0.934	0.524	0.522
HNSC	Testing	0.609	0.576	0.629	0.582	0.570	0.558	0.571	OP	0.608	0.568	0.522	0.524
CESC	Training	0.976	0.972	NA	0.891	0.920	0.881	1.000	1.000	0.927	0.961	0.539	0.548
CESC	Testing	0.698	0.623	NA	0.646	0.802	0.658	OP	OP	0.618	0.626	0.548	0.539
SARC	Training	0.948	0.983	0.919	0.732	0.953	0.905	1.000	1.000	0.934	0.955	0.711	0.559
SARC	Testing	0.715	0.621	OP	0.597	0.651	0.592	OP	0.692	0.682	0.720	OP	OP
STAD	Training	0.890	0.981	0.877	0.913	0.881	0.919	1.000	1.000	0.877	0.944	0.699	0.631
STAD	Testing	0.906	0.730	0.538	0.558	0.761	0.727	0.631	0.521	0.756	0.656	0.631	0.699
UCEC	Training	1.000	0.957	1.000	0.936	1.000	0.936	1.000	1.000	NA	NA	0.783	0.532
UCEC	Testing	0.702	0.826	0.638	0.783	0.702	0.957	OP	OP	NA	NA	0.532	0.783
ESCA	Training	1.000	1.000	0.895	0.971	0.947	0.882	1.000	1.000	0.885	NA	0.842	0.794
ESCA	Testing	0.613	OP	OP	0.868	OP	OP	0.516	0.581	0.516	NA	OP	OP
LUAD	Training	0.990	0.817	0.620	0.875	0.750	0.663	1.000	1.000	0.923	0.942	0.644	0.548
LUAD	Testing	0.692	0.683	0.577	0.481	0.567	0.548	OP	OP	0.567	OP	0.548	0.644
PAAD	Training	0.964	0.888	0.813	NA	0.855	0.838	1.000	1.000	0.928	0.900	0.554	0.600
PAAD	Testing	0.663	0.639	OP	NA	0.625	0.596	OP	OP	0.550	OP	0.600	0.554
LUSC	Training	0.982	0.938	0.754	0.979	0.754	0.729	1.000	1.000	0.930	0.917	0.588	0.563
LUSC	Testing	0.563	0.772	OP	OP	OP	0.667	OP	0.614	0.542	0.526	OP	OP

ANN Artificial neural networks, CPH Cox proportional hazard model, RSF Random survival forest, RSI Radiosensitivity index in Torres-Roca's study, OP the prediction directions of training and testing values were opposite, CV Cross validation, NA Not available

Back to article page

ISSN: 1471-2407

Contact us

Submission enquiries: bmccancer@biomedcentral.com
General enquiries: ORSupport@springernature.com