ABSTRACT: BACKGROUND: Arsenic exposure increases the risk of non-cancerous and cancerous diseases. In the Antofagasta region in Chile, an established relationship exists between arsenic exposure and the risk of cancer of the bladder, lung and skin. Platinum-based drugs are first-line treatments, and many works recognise selenium as a cancer-fighting nutrient. We characterised the short-term urinary excretion amounts of arsenic, selenium and platinum in 24-h urine samples from patients with lung cancer and those with cancer other than lung treated with cisplatin or/and carboplatin. As - Se - Pt inter-element relationships were also investigated. RESULTS: The amounts of platinum excreted in urine were not significantly different between patients with lung cancer and those with other cancers treated with cisplatin, despite the significant variation in platinum amounts supplied from platinum-based drugs. In general, the analytical amounts of excreted selenium were greater than those for arsenic, which could imply that platinum favours the excretion of selenium. For other types of cancers treated with drugs without platinum, excretion of selenium was also greater than that of arsenic, suggesting an antagonist selenium-anti-cancer drug relationship. CONCLUSIONS: Regards the baseline status of patients, the analytical amounts of excreted Se is greater than those for As, particularly, for cisplatin chemotherapy. This finding could imply that for over the As displacement Pt favours the excretion of Se. The analytical amounts of excreted Se were greater than those for As, either with and without Pt-containing drugs, suggesting an antagonist Se-anti-cancer drug relationship. However, it seemed that differences existed between As - Se - Pt inter-element associations in patients treated for lung cancer in comparison with those treated for cancer other than lung. Therefore, knowledge obtained in this work, can contribute to understanding the arsenic cancer mechanism and the As - Se - Pt inter-element association for lung cancer and other types of cancer, which in some cases respond at a linear mathematical model.