Machine learning to improve interpretability of clinical, radiological and panel-based genomic data of glioma grade 4 patients undergoing surgical resection

Table 1 Univariate and multivariate Cox regression analysis of OS in the whole cohort (102 GG4 cases)

Characteristic	N	Univariate			Multivariate
Characteristic	N	HR	95% CI	p-value	HR	95% CI	p-value
CW	102
0; untreated	63	–	–		–	–
1; treated	39	0.53	0.34, 0.81	0.003	0.62	0.40, 0.96	0.032
Age	102	1.01	1.0, 1.04	0.2
EOR	102	0.92	0.89, 0.96	< 0.001	0.94	0.91, 0.98	0.002
Localization	102
0; precentral	41	–	–
1; postcentral	21	1	0.58, 1.72	> 0.9
2; temporoinsular	40	1.09	0.70, 1.71	0.7
Ki67	102	1.01	1.00, 1.02	0.14
IDH mutational status	102
0; unmutated	94	–	–
1; mutated	8	0.65	0.32, 1.35	0.3
Side	102
0; left	54	–	–
1; right	48	1.2	0.81, 1.80	0.4
Preoperative volume	102	1.01	1.00, 1.02	0.033	1.01	1.00, 1.02	0.25
Extent of resection_2 categories	102
0; ≤99%	52	–	–
1; ≥100%	50	0.81	0.54, 1.20	0.3
MGMT status	102
0; unmethylated	40	–	–
1; methylated	62	0.44	0.28, 0.68	< 0.001	0.47	0.30,0.74	< 0.001
Gender: female	102	1.18	0.76, 1.82	0.5

CI confidence interval, HR hazard ratio, EOR extent of resection. In bold p-value < 0.05

ISSN: 1479-5876