Table 3.

Dose-response relationships for individual causes of non-breast-cancer mortality that are significantly increased by radiotherapy.

Endpoint	Organ	Dose measure	Number of events	Percentage increase in RR per Gy*	Reference†	Table/figure in reference
Major coronary events	Whole heart	Mean	963	7.4% (95% CI 2.9–14.5)	Darby 201318	Fig. 1
Lung cancer	Lung‡	Mean	475 all studies combined	11% (95% CI 6–19)	EBCTCG 201717	Fig. S8
Oesophageal cancer	Whole oesophagus	Median	156	7.1% (95% CI 1.8–20.6)§	Journy 202019	Table 2

Abbreviations: RR rate ratio; Gy gray; CI confidence interval

^*i.e. excess RR per Gy (lung cancer and major coronary events) or excess odds ratio per Gy (oesophageal cancer). Models are of the form B_s(1 + KX/100) where S denotes a group, or stratum, of individuals for whom the rate at which the endpoint occurs in the absence of radiation exposure is likely to be similar. B_s is the rate at which the endpoint occurs in that stratum in the absence of radiotherapy, X is the dose measure in Gy and K is the percentage increase in the rate ratio or the odds ratio per Gy.

^†See also Supplemental Table 8.

^‡Based on published data meta-analysis of five studies where doses were allocated to individuals based on trial-level or individual patient doses. Organ doses were for both lungs combined in one study, ipsilateral lung in two studies and location of second cancer in two studies.

^§The dose–response relationship based on whole oesophagus dose is listed because it is based on median oesophagus dose, which is assessable for patients being considered for breast cancer radiotherapy.