Modeling and simulation of maintenance treatment in first-line non-small cell lung cancer with external validation

Table 3 Screening of the potential covariates for overall survival using the Cox model

	HR	95 % CI	Score	p	Sign on risk
TTG (weeks)	0.83^a	0.81–0.85^a	151.7	<0.0001	−
Tumor size at randomization (cm)	1.17^b	1.13–1.20^b	51.8	<0.0001	+
Week 8 ECTS	1.12^c	1.10–1.14^c	45.1	<0.0001	+
Female	0.64	0.56–0.74	21.2	<0.0001	−
Never smoked	0.57	0.47–0.68	20.3	<0.0001	−
Asian	0.61	0.50–0.75	12.5	<0.0001	−
Study SATURN	1.33	1.16–1.51	9.2	<0.0001	+
ECOG score >0	1.30	1.14–1.49	7.7	0.0001	+
Age ≥ 55 years	1.23	1.07–1.42	4.2	0.0037	+
Squamous	1.22	1.06–1.40	3.8	0.0060	+
Erlotinib	0.85	0.75–0.96	3.5	0.0082	−
Erlotinib in SD	0.84	0.73–0.97	3	0.0144	−
Age (years)	1.07^d	1.01–1.14^d	2.6	0.0221	+

CI confidence interval, ECTS early change in tumor size, Erlotinib patients who received erlotinib during the first-line treatment run-in phase (all patients were analyzed), Erlotinib in SD patients who received erlotinib during the first-line treatment phase (only those patients who achieved stable disease during first-line treatment run-in phase were analyzed), HR hazard ratio, p obtained by likelihood ratio test, Score log(likelihood ratio), Sign on risk + (−) indicates that this variable is associated with increased (decreased) risk, TTG time to tumor regrowth, ^aHR for increase of every 8 weeks of TTG; ^bHR for increase of every 2 cm of tumor size at randomization; ^cHR for increase in every 10 % of tumor size as compared to the randomization; ^dHR for increase of every 10 years of age

ISSN: 1471-2407