Fig. 4: Serial ctDNA monitoring for personalized dynamic risk prediction.
From: Dynamic recurrence risk and adjuvant chemotherapy benefit prediction by ctDNA in resected NSCLC
a Conception of Joint model. b The comparison of model performance between the joint model and cox models (testing datasets). Fivefold cross-validation were repeated for 20 times. The p value is calculated using two-sided Wilcoxon signed-rank test. ns: not significant; *p < 0.05; **p < 0.01; ***p < 0.001. Center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range. AUROC: areas under the receiver-operating characteristics curves. c, d Personalized dynamic risk prediction for patients P062 (c) and P017 (d). From left to right, predictions were calculated accounting for ctDNA that measured previously and were updated when new measure became available. The vertical dotted lines represent the time point of the last ctDNA measurement. To the left of the vertical line is fitted longitudinal trajectory. To the right of the vertical line is the median estimator for recurrence-free probability with 95% pointwise uncertainty band. \({{{{{\rm{Log}}}}}}\,{{{{{\rm{ctDNA}}}}}}\,{{{{{\rm{level}}}}}}={{{{{\rm{ln}}}}}}\left({{{{{\rm{mean}}}}}}\,{{{{{\rm{VAF}}}}}}+{10}^{-6}\right)-{{{{{\rm{ln}}}}}}{10}^{-6}\).
