Fig. 1

Utilization of Sputum Samples. Of the 171 samples run on the LSRII that were originally considered (136 non-cancer; 31 cancer; 4 with unconfirmed health status), 168 samples were used for model building and analysis pipeline development. This included four samples for which we did not have a definitive disease status because the addition of unlabeled samples had been shown to be helpful in model building [24]. In addition, 14 samples flagged as ineligible based on cell counts (see below) were also used in the model development to better capture the distribution of the underlying data and help make generalization of the model more robust to sample noise. Three samples could not be used at all due to technical problems during acquisition. One hundred and fifty samples were ultimately used for the model validation phase (122 non-cancer; 28 cancer). Eighteen of the 168 samples were omitted: thirteen included too few cells for an accurate analysis, one included too few alveolar macrophages thereby failing to confirm it as a lung sample, and four samples were excluded because their cohort status could not be confirmed. An independent validation of the automated analysis was performed with 32 new samples. Participants adhered to the same enrollment criteria, and samples were processed with the same protocol as the previous sample set. Although a different flow cytometer (Navios EX) was used to run the second set of samples, the same model and coefficients were used to analyze the data for both instruments