Monitoring of Dynamic Changes and Clonal Evolution in Circulating Tumor DNA From Patients With IDH-Mutated Cholangiocarcinoma Treated With Isocitrate Dehydrogenase Inhibitors

Abstract

Purpose: IDH mutations occur in about 30% of patients with cholangiocarcinoma. Analysis of mutations in circulating tumor DNA (ctDNA) can be performed by droplet digital polymerase chain reaction (ddPCR). The analysis of ctDNA is a feasible approach to detect IDH mutations.

Methods: We isolated ctDNA from the blood of patients with IDH-mutated advanced cholangiocarcinoma collected at baseline, on therapy, and at progression to isocitrate dehydrogenase (IDH) inhibitors.

Results: Of 31 patients with IDH1^R132 (n = 26) or IDH2^R172 mutations (n = 5) in the tumor, IDH mutations were detected in 84% of ctDNA samples analyzed by ddPCR and in 83% of ctDNA samples analyzed by next-generation sequencing (NGS). Patients with a low variant allele frequency of ctDNA detected by NGS at baseline had a longer median time to treatment failure compared to patients with high variant allele frequency of ctDNA (3.6 v 1.5 months; P = .008). Patients with a decrease in IDH-mutated ctDNA on therapy by ddPCR compared with no change/increase had a trend to a longer median survival (P = .07). Most frequent emergent alterations in ctDNA by NGS at progression were ARID1A (n = 3) and TP53 mutations (n = 3).

Conclusion: Detection of IDH mutations in ctDNA in patients with advanced cholangiocarcinoma is feasible, and dynamic changes in ctDNA can correspond with the clinical course and clonal evolution.

FIG 1.

(A) There was a correlation (0.8, P < .001) between ddPCR and digital NGS for reported VAF in ctDNA. (B) Kaplan-Meier curves for TTF analyses per baseline quantity of ctDNA (≤ median v > median) in blood samples for IDH mutations detected by ddPCR (VAF ≤ 1.4% v VAF > 1.4 %; P = .21), (C) IDH mutations detected by digital NGS (VAF ≤ 2.2% v VAF > 2.2%; P = .09), (D) and the combined VAF for all alterations detected by digital NGS (VAF ≤ 4.6% v VAF > 4.6%; P = .008). (E) Comparisons of ctDNA samples for the levels of IDH mutation detected by ddPCR in baseline and progression samples from the same patient (P = .049) and (F) the levels of IDH mutation detected by digital NGS in baseline and in progression samples from the same patient (P = .06). (G) Kaplan-Meier curves of changes in VAF of IDH-mutated ctDNA during therapy demonstrated no difference in the median TTF between patients with decrease in IDH-mutated ctDNA on therapy compared with increase or no change (0.9 months; 95% CI, 0 to 2.5 v 0.8 months; 95% CI, 0 to 2.1; P = .29), (H) but there was a trend toward a longer median survival in patients with decrease in quantity of IDH-mutated ctDNA compared with patients with no change or increase in IDH-mutated ctDNA (33.1 months; 95% CI, 13 to 53.2 v 9.3 months; 95% CI, 2.1 to 16.4 months; P = .07). ctDNA, circulating tumor DNA; ddPCR, droplet digital polymerase chain reaction; IDH, isocitrate dehydrogenase; NGS, next-generation sequencing; VAF, variant allele frequency; TTF, time to treatment failure.

FIG 2.

(A) IDH-mutated ctDNA by ddPCR in serially collected plasma during therapy with IDH inhibitors. (B) Heatmap of mutations detected by digital NGS in ctDNA from patients with IDH-mutated cholangiocarcinoma treated with IDH inhibitors at baseline and progression. Red crosses mark samples that failed quality control. (C-G) Dynamic tracking of VAF for detected mutations in ctDNA (ddPCR) from serial plasma samples or tumor tissue (NGS sequencing) from serial biopsies in patients treated with IDH inhibitors. Bx, biopsy; ctDNA, circulating tumor DNA; ddPCR, droplet digital polymerase chain reaction; IDH, isocitrate dehydrogenase; IDHi-1, first IDH1 inhibitor; IDHi-2, second IDH inhibitor; mut, mutant without reported VAF; NGS, next-generation sequencing; PD, progressive disease; Pre-tx, pretreatment; SD, stable disease; VAF, variant allele frequency.