Longitudinal Mutational Analysis in Hydroxycarbamide-Resistant/Intolerant Essential Thrombocythemia Treated on the Majic-ET Study

MAJIC-ET is a phase 2 study comparing ruxolitinib (RUX) to best available therapy (BAT) in hydroxycarbamide resistant/intolerant (HC-RES/INT) essential thrombocythaemia (ET). The presence of additional non-driver mutations at baseline, in particular TP53 and SF3B1, were predictive for increased risk of disease transformation in this cohort (EHA-4276). Now, we expand the mutation analysis up to 5 years to evaluate clonal evolution after prolonged treatment and correlate with clinical outcome.

Driver mutation (JAK2/CALR/MPL) allele burdens were quantified using targeted next-generation sequencing and non-driver mutation analysis was performed using an ISO accredited Illumina TruSeq Custom Amplicon Panel, including 31 gene mutation hotspots & exons (~36,000 bp, 287 amplicons).

Data was censored in January 2017 with 47.3% (n=52) randomised to BAT & 52.7% (n=58) to RUX. Overall, 49.1% were JAK2-mutated, 30% CALR-mutated, 4.5% MPL-mutated and 16.4% triple-negative (TN). At 12 months (m), molecular responses (MR) occurred rarely with complete MR (CMR) in 2 patients (pts) and partial MR (PMR) in 3 pts. MR were exclusively seen in RUX-treated pts. 54% (50/92) of those with driver mutations had a >12m allele burden analysis. Median sample time was 48m after trial entry (24 - 60m). There was no significant change in allele burden in the overall cohort. A new JAK2 V617F mutation (allele burden of 6%) was detected at 36m in a "TN" pt at baseline. Of the 5 pts achieving a molecular response at 12m, 2 of these subsequently lost their molecular response; a JAK2-mutated pt achieving CMR at 12m but at 52m had an allele burden of 4% (now PMR) and a CALR-mutated pt with PMR (allele burden 65 to 9%) at 12m, switched from RUX due to toxicity and subsequently lost their MR at 60m (allele burden of 44%), associated with clonal evolution (gain of a new ASXL1 mutation). New MRs occurred in only 2 pts; PMR in a MPL-mutated pt on pegylated-interferon and in a CALR-mutated pt on RUX. 4 pts with a baseline allele burden of <20% had a ≥50% reduction at latest time point (TP); interferon-treated (n=2) and RUX-treated (n=2).

Analysis of non-driver mutations was performed on latest TP samples in 51 pts (46%) at a median of 40m after trial entry (6 - 60m). Additional non-driver mutations were present in 30% (33/110) of the cohort at baseline (EHA-4276). In 94% of pts with non-driver mutations at baseline, these were still detected at the latest TP (median 38m, 12 - 52m). There was a <10% allele burden change in 92%. However, a ≥10% increase in allele burden occurred in a TP53-mutated pt who transformed to acute leukaemia: allele burden at baseline of 9% to 39% at transformation. New non-driver mutations were frequently acquired; 9/51 pts (18%) with a total of 15 mutations detected; TET2 [5], ASXL1 [3], TP53 [3], SF3B1 [2], U2AF1 [1], SETBP1 [1]. Retrospective review of the paired baseline sample showed 93% of these mutations present at levels below the pre-specified threshold for variant calling. Mutations were equally acquired in JAK2- or CALR-mutated pts with no mutations acquired in MPL-mutated or TN pts. New mutations occurred on RUX (5/9, 56%) at a similar frequency to BAT (4/9, 44%). There was no association between acquisition of a subsequent new mutation and presence of non-driver mutation at baseline or haematological response at 12m. In transformed pts (12/110), 75% (9/12) had additional non-driver mutations at baseline. Of those with later samples for analysis (n=8) at the time of transformation, no new mutations were detected. New mutation acquisition was not associated with disease duration, age, gender or prior treatments and did not influence 2-year overall survival.

We report for the first time a comprehensive longitudinal mutational analysis of HC-RES/INT ET patients within the context of a clinical trial. In this cohort, we found that MR was rare, both early and late in the course of treatment. Initial MR was often not sustained and may signify clonal evolution. As the presence of SF3B1 and TP53 mutations at trial entry predicted for disease progression, we anticipate that the new SF3B1 and TP53 mutations identified at later TP will also increase risk of subsequent transformation during longer follow-up. These data highlight the clinical utility of longitudinal molecular analysis of patients with HC-RES/INT ET.