Relationship of JAK2^V617F Allelic Burden (AB) to Demographics, Disease Characteristics, and Response to Therapy in Persist-1, a Randomized Phase III Study of Pacritinib (PAC) Versus Best Available Therapy (BAT) in Patients (pts) with Primary and Secondary Myelofibrosis (MF)

Background: The majority of pts with MF present with a mutation in exon 14 of the JAK2 gene (JAK2^V617F). Studies have suggested that in pts with MF and this mutation, a lower AB is associated with a myelodepletive phenotype and shortened survival (Tefferi, Leukemia 2008; Guglielmelli Blood 2009; reviewed in Vannucchi, Ther Adv Hematol 2011). For example, in a Cox proportional hazard model for factors predictive of overall survival (OS) in pts with primary MF, a JAK2^V617F AB of <25% was associated with a shortened OS (p=0.002; Guglielmelli, Blood 2009). In the phase III trials of ruxolitinib (RUX) therapy in JAK inhibitor-naïve MF, RUX-treated pts had a median decrease in AB of 16% (maximal) in COMFORT-I followed up to 216 weeks and 7% in COMFORT-II at 48 weeks. The median AB at the start of therapy was not reported for COMFORT-I and in COMFORT-II was 84.5% (Cervantes, Blood 2013; Deininger, Blood2015).

Methods: In this communication, we report on clinical correlates of quantitative baseline JAK2 mutational status and the effect of treatment on JAK2^V617Fmutational burden in the PERSIST-1 trial. This trial enrolled JAK inhibitor-naïve pts with primary or secondary MF without restrictions on baseline platelet counts, randomized 2:1 between PAC and BAT (excluding JAK2 inhibitors) arms. Stratifications included Dynamic International Prognostic Scoring System (DIPSS) risk category, and baseline platelet counts (<50,000/μL, 50,000-100,000/μL, >100,000/μL). Following the 24-week primary analysis, crossover was allowed by physician/pt choice and 90/107 BAT-treated pts (84%) crossed over to receive PAC. Baseline characteristics have been previously reported (Mesa, ASCO 2016) and included 203/327 (62%) pts with Primary MF, 42 (13%) with post-essential thrombocythemia MF and 81 (25%) with post-polycythemia vera MF. 320 of 327 pts had informative PCR testing for JAK2^V617F. We report the baseline characteristics of pts with high (≥50%) AB of JAK2^V617F, low (<50%) AB of JAK2^V617Fand those without a detectable mutation at baseline (<0.1%), as well as the effect of PAC and BAT on AB and correlation with outcome measures.

Results: 131 (40.9%) had AB >50% (high AB), 115 (35.9%) had an AB ≥0.1% and ≤50% (low AB), and 74 (23.1%) had AB <0.1% (wild-type JAK2). Median AB values were 80.3% for the high AB group and 30.3% for the low AB group. Baseline characteristics that were significantly different between high and low AB pts are shown in the Table. No differences were noted in the DIPSS risk category or baseline percentage of blasts between pts with low or high AB. There were a higher percentage of pts with a lower AB in the PAC arm (64% vs. 48%; p=0.01). PAC therapy resulted in a median decrement in AB at Week 24 of −15.8% vs −7.9% for pts in the BAT arm (Figure). The median greatest reduction from baseline in AB in PAC-treated pts was −31.6%. Decrease in AB was strongly correlated with spleen volume reduction (p=0.003) and trended towards correlation with improvement in total symptom score (p=0.07) in PAC-treated, but not in BAT-treated pts. Overall survival was shorter in PAC-treated pts with a minimal AB reduction vs pts with greater decreases in AB.

Conclusions: In the PERSIST-1 trial, pts with a JAK2^V617FAB of <50% had substantially different disease and demographic characteristics when compared with those pts with a >50% AB despite similar distribution of DIPSS scores. In this trial, reduction in AB associated with PAC, but not BAT correlated with clinical benefit.

View largeDownload slide

View largeDownload slide

Close modal

View largeDownload slide

View largeDownload slide

Close modal