We explored the shape of the exposure-response relationship of arsenic-related lung cancer and the interaction between arsenic and tobacco use.A total of 3278 tin miners with at least 10 years of arsenic exposure were enrolled since 1992 and followed up for 27 years. After excluding radon-exposed miners and former smokers, 1620 miners were included into the sub-cohort. Lung cancer risks were estimated by modeling total exposure and intensity of arsenic exposure.The cohort experienced 73,866 person-years and 414 lung cancer cases. Firstly, the ERR/mg/m3-year was 0.0033 (95% CI: 0.0014-0.0045) in arsenic concentration <3 mg/m3 and 0.0056 (95% CI: 0.0035-0.0073) in arsenic concentration ≥3 mg/m3. After adjusting for cumulative arsenic exposure, and the ERR/mg/m3 increased with increasing intensity (0.129 (95% CI: 0.039, 0.189)). Secondly, an unique aspect of this population was the early age at first arsenic exposure for workers. Results showed that lung cancer incidence risk from exposed in childhood (<13 years) was non-significantly greater than those in other age groups (13-17 and ≥ 18 years). Finally, the most likely joint effects of inhaled arsenic and tobacco use was sub-multiplicative.This study enlightened us that for fixed cumulative arsenic exposure, higher concentration over shorter duration might be more deleterious than lower concentration over longer duration. Substantial reductions in the lung cancer burden of smokers exposed to arsenic could be achieved by reductions in either exposure.Copyright © 2021 Elsevier Inc. All rights reserved.