Inclusive and\nAccurate Clinical Diagnostics Using\nIntelligent Computation and Smartphone Imaging

Smartphone-based colorimetry has been widely applied\nin clinical\nanalysis, although significant challenges remain in its practical\nimplementation, including the need to consider biases introduced by\nthe ambient imaging environment, which limit its potential within\na clinical decision pathway. In addition, most commercial devices\ndemonstrate variability introduced by manufacturer-to-manufacturer\ndifferences. Here, we undertake a systematic characterization of the\npotential imaging interferences that lead to this limited performance\nin conventional smartphones and, in doing so, provide a comprehensive\nnew understanding of smartphone color imaging. Through derivation\nof a strongly correlated parameter for sample quantification, we enable\nreal-time imaging, which for the first time, takes the first steps\nto turning the mobile phone camera into an analytical instrument –\nirrespective of model, software, and the operating systems used. We\ndemonstrate clinical applicability through the imaging of patients’\nskin, enabling rapid and convenient diagnosis of cyanosis and measurement\nof local oxygen concentration to a level that unlocks clinical decision-making\nfor monitoring cardiovascular disease and anemia. Importantly, we\nshow that our solution also accounts for the differences in individuals’\nskin tones as measured across the Fitzpatrick scale, overcoming potential\nclinically significant errors in current optical oximetry.