Fig. 2: ctDNA detection before and on treatment.
A levels of ctDNA detection at baseline in patients with ctDNA detected (n = 39, median = 31.0, IQR [9.0-557.0]). Errors bars are presented as median +/- IQR. B, C correlation between baseline ctDNA levels with serum LDH level (Pearson coefficient; (B); n = 63, exact p-value:3.10−8) and with diameter of the largest metastasis according to TNM (Wilcoxon test; (C); M1a: n = 36; M1b: n = 22 and M1c: n = 6). Error bars are presented as median +/- IQR. M1a: median = 0.1, IQR [0.1-9.0]; M1b: median = 18.0, IQR [8.5-499.3]; M1c: median = 431.0, IQR [146.8-1570]. Two-sided statistical tests were used. Exact p-value for comparison between M1a and M1b is 9.6.10−5. D Sankey plot showing the dynamic changes of ctDNA detection in patients under tebentafusp. E swimmer plot showing the dynamic changes of ctDNA levels in each patient. Top panel E presents the patients with ctDNA detected at baseline while bottom panel presents patients without ctDNA detected at baseline or patients with no available ctDNA analysis at baseline. Symbols in black and white represent the samples with positive and negative ctDNA detection. Red, yellow and green bars represent the duration of treatment (with colors representing the best response in each patient) while gray bars represent the follow-up after tebentafusp discontinuation. ND: not detected; LDH: lactate dehydrogenase; MAF: mutant allele frequency; NA: not available; TNM: tumor node metastasis classification. Source data are provided as a Source Data file.
