Fig. 3: Comparison across the subphenotypes in the development sub-cohort. | Nature Communications

Fig. 3: Comparison across the subphenotypes in the development sub-cohort.

From: Identification of predictive subphenotypes for clinical outcomes using real world data and machine learning

Fig. 3

a Kaplan–Meier curves for overall survival (OS) for each subphenotype. The survival probabilities are shown by solid lines, and the shaded areas represent 95% confidence intervals around the curves. P-values are indicated in the figures; the numbers of samples to generate the curves are shown in the figures. b Group sunburst plot of administration rates of the medications across subphenotypes. The groups indexed by letters are determined based on the third level of anatomical therapeutic chemical (ATC3) class. c Chord diagrams of differences in metastasis (left), comorbidities (middle), and abnormal clinical feature classes (right) across subphenotypes. Ribbons indicate the normalized proportion of patients from a subphenotype to the corresponding feature class. d Graph of co-incidence patterns across subphenotypes. Each node within the network represents a specific feature, with its size proportional to the incidence within the particular subphenotype. The edge connecting a pair of nodes indicates the co-incidence of a corresponding feature pair, with the thickness proportional to the co-incidence rate; lines are displayed if the rate exceeds 5%. CHF congestive heart failure, FED fluid electrolyte disorders, CPD chronic pulmonary disease. Source data are provided with this paper.

Back to article page