. Author manuscript; available in PMC: 2023 Jun 16.

Published in final edited form as: Radiat Res. 2022 May 1;197(5):491–508. doi: 10.1667/RADE-21-00059.1

Table 8.

Site-specific excess relative risk estimates for cancer mortality in relation to weighted colon dose under linear and linear-quadratic dose response models: LSS, 1958–2009

	Males						Females
Site	N	L	L	Q	Q/L	P ^a	N	L	L	Q	Q/L	P ^a
		95% CI	95% CI	95% CI				95% CI	95% CI	95% CI
Stomach	2,105	0.12^b	0.03^c	0.04^c	1.72	0.487	1,790	0.43^b	0.45^c	−0.01^c	−0.01	0.792
		<−0.01 to 0.31	<−0.03 to 0.38	NE				0.19 to 0.73	−0.06 to >1.01	NE
Colon	396	0.52	0^d	0.40^d	Infinite	0.091^d	605	0.39	0^d	0.38^d	Infinite	0.025^e
		0.15 to 1.13		0.14 to 0.82				0.09 to 0.87		0.11 to 0.80
Liver	1,120	0.26	0.26	−0.01	−0.04	0.930	806	0.40	0.11	0.28	2.58	0.204
		0.04 to 0.58	<0.05 to 0.83	<−0.27 to 0.31				0.09 to 0.82	<−0.59 to 0.73	−0.13 to 0.94
Lung	1,488	0.42	0.32	0.06	0.19	0.883	1,004	1.03	0.71	0.21	0.29	0.508
		0.16 to 0.72	<−0.23 to 0.95	<−0.28 to 0.43				0.57 to 1.61	<−0.12 to 1.62	<−0.34 to 0.80

Note. LSS, Life Span Study. N, number of deaths. ERR, excess relative risk. L, linear dose coefficient. Q, quadratic dose coefficient. Q/L, ratio of quadratic to linear dose coefficient or curvature. 95% CI, confidence interval. AIC, Akaike information criterion.

P value for quadratic departure from linearity unless stated otherwise.

Estimates are from the site-specific linear ERR models with parametric background, multiplicative adjustment for smoking, and independent effect modification by sex, sex-specific attained age, age at exposure, and high dose. Estimates are shown for individuals at age 70 following exposure at age 30.

Estimates are from the site-specific linear-quadratic ERR models with parametric background, multiplicative adjustment for smoking, and independent effect modification by sex, sex-specific attained age, age at exposure, and high dose unless stated otherwise. Estimates are shown for individuals at age 70 following exposure at age 30.

Estimates are from site-specific pure quadratic ERR model (i.e., linear dose coefficient constrained to be zero) with parametric background, multiplicative adjustment for smoking, and independent effect modification by sex, sex-specific attained age, age at exposure and high dose. Estimates are shown for individuals at age 70 following exposure at age 30.

With linear dose coefficient constrained to be zero.