BACKGROUND: Cell-free, circulating tumour DNA (ctDNA) is an established measure of minimal residual disease
however, it is not utilised in melanoma management. We investigated whether ctDNA measurements could predict survival outcomes during adjuvant targeted therapy or placebo treatment in stage III melanoma, thereby identifying patients at high risk and low risk of recurrence. METHODS: Analytically validated mutation-specific droplet digital PCR assays were used to measure BRAF FINDINGS: Baseline plasma samples were available for 597 of 870 patients (331 male patients and 266 female patients) and samples for assessing the ctDNA positivity rate at landmark follow-up timepoints of 3 months, 6 months, 9 months, and 12 months after treatment initiation were available for 94 of 870 patients. Additionally, samples were available from 118 of 870 patients within a 2-month timeframe before or after clinical or radiographic recurrence. Median follow-up for the biomarker analyses was 60 months (IQR 39-66) in the combination therapy group and 58 months (21-66) for the placebo group. ctDNA was detectable in 79 (13%) of 597 baseline samples. ctDNA positivity rate and mutant copies per mL plasma were significantly higher in patients with higher disease substages. As a binary variable, ctDNA detection was associated with worse recurrence-free survival (placebo group: median 3·71 months [95% CI 2·39-6·89] vs 24·41 months [17·28-43·13]
hazard ratio [HR] 2·91 [95% CI 1·99-4·25], p<
0·0002)
combination therapy group: median 16·59 months [95% CI 12·02-26·80] vs 68·11 months [50·36-not reached]
HR 2·98 [1·95-4·54], p<
0·0002) and overall survival (placebo group: median 33·90 months [13·96-not reached] vs not reached
HR 3·35 [2·01-5·55], p<
0·0002)
combination therapy group: median 40·31 months [24·90-not reached] vs not reached
HR 4·27 [2·50-7·27], p<
0·0002) in the placebo group and combination therapy groups. Baseline ctDNA was more strongly associated with survival outcomes than IFNG gene expression or tumour mutational burden. Patients with adverse longitudinal ctDNA kinetics (molecular relapse or persistently positive) had markedly shorter median recurrence-free survival (8·31 months [95% CI 5·39-12·20] and 5·32 months [2·79-not reached], respectively) compared with patients with favourable kinetics (ie, undetectable after positive baseline result: 19·25 months [16·39-not reached]
and durable undetectable: not reached [38·44-not reached], p<
0·0002). INTERPRETATION: Droplet digital PCR measurements of ctDNA to assess minimal residual disease before adjuvant targeted therapy and during follow-up can identify patients at high risk of early recurrence. Additional studies using ctDNA measurements to guide therapeutic interventions might lead to improvements in the management of resected stage III melanoma. FUNDING: Novartis.