<p>Mixed venous oxygen saturation (SvO₂) is a valuable indicator of the balance between oxygen delivery and consumption in the body. However, the invasiveness and risk of complications limit its use. This study aimed to test the feasibility of continuously monitoring variations in SvO₂ from photoplethysmography (PPG) measurements of sternal intramedullary oxygen saturation. In an animal model, including anesthetized pigs (<i>n</i> = 15, mean 33.1 (SD 2.0) kg), alterations in SvO₂ were instituted by separate sessions: incremental decrease in arterial oxygen saturation using hypoxic gas mixtures (N₂/air), and hemorrhage of 20% plus 20% of estimated blood volume. Estimations of non-invasive sternal oxygen saturation (nSsO₂) were derived from calibrating PPG measurements against SvO₂ sampled from the pulmonary artery. Individual linear correlations were found between SvO₂ and PPG measurements. nSsO₂ and SvO₂ co-varied, except in severe hypovolemia. A concordance rate between nSsO₂ and SvO₂ was determined to 87% (95% CI 79.1–95.0). In this study, we demonstrated linear correlations between SvO₂ and PPG-measurements on the sternum, enabling estimation of nSsO₂, which trends with changes in SvO₂. The results support further technical development of non-invasive monitoring of SvO₂ from sternal measurements.</p> Graphical abstract <p>Non-invasive sternal oxygen saturation (nSsO₂) is derived from photoplethysmographic measurements of sternal intramedullary blood flow (SQV), correlated against mixed venous oxygen saturation (SvO₂). The performance of nSsO₂ to detect changes in SvO₂ is subsequent analyzed with a concordance analysis.</p> <p></p>