Figure 3

Curve estimation of dynamic patterns between moderate and severe groups. (A) Optimal curve fitting between conventional CT score and time from symptom onset (d) in moderate and severe groups with the equations: \(y=0.460*x-0.0156*{x}^{2}+0.000128*{x}^{3}\) (r² = 0.608, p < 0.001), and \(y=1.41*x-0.0448*{x}^{2}+0.000389*{x}^{3}\) (r² = 0.822, p < 0.001), respectively; (B) Optimal curve fitting between the percent of pulmonary lesions (%) calculated by deep learning-based quantification and time from symptom onset (d) in moderate and severe groups with the equations: \(y=0.413*x-0.0148*{x}^{2}+0.000127*{x}^{3}\) (r² = 0.319, p < 0.001), and \(y=2.89*x-0.0912*{x}^{2}+0.000794*{x}^{3}\) (r² = 0.661, p < 0.001), respectively. (C) Optimal curve fitting between the percent of pulmonary GGO lesions (%) calculated by deep learning-based quantification and time from symptom onset (d) in moderate and severe groups with the equations: \(y=0.321*x-0.0114*{x}^{2}+0.0000977*{x}^{3}\) (r² = 0.331, p < 0.001), and \(y=3.30*x-0.0704*{x}^{2}+0.000606*{x}^{3}\) (r² = 0.670, p < 0.001), respectively. (D) Optimal curve fitting between the percent of pulmonary consolidation lesions (%) calculated by deep learning-based quantification and time from symptom onset (d) in moderate and severe groups with the equations: \(y=0.0911*x-0.00338*{x}^{2}+0.0000297*{x}^{3}\) (r² = 0.202, p < 0.001), and \(y=0.590*x-0.0208*{x}^{2}+0.000188*{x}^{3}\) (r² = 0.462, p < 0.001), respectively.