Interpretability of radiomics models is improved when using feature group selection strategies for predicting molecular and clinical targets in clear-cell renal cell carcinoma: insights from the TRACERx Renal study

Table 3 Logistic regression coefficients, AUROC and p-values for five targets that are robustly predicted by three shape features from the whole tumour ROI using the Proposed pipeline

Target	Logistic regression coefficients			Performance
Target	Sphericity	Flatness	MeshVolume	AUROC	p-value
Renal vein invasion	-1.278	-0.551	0.246	0.862	< 0.01
IVC invasion	-0.830		0.094	0.680	0.048
ITH Index	-0.722		0.528	0.745	0.013
wGII Max	-0.783		0.230	0.737	0.015
Loss 9p21.3	-1.29		0.367	0.814	< 0.01

ISSN: 1470-7330