Outcomes of Patients with Myeloproliferative Neoplasms (MPN) after Interferon-Alpha (IFN) Therapy Discontinuation

Background:We and others have shown that IFN therapy provides several benefits to MPN patients (pts) including molecular responses and possible long-term remissions off cytoreductive therapy (CRT). However, little is known about outcomes after IFN discontinuation, and the clinical relevance of achievement of molecular response is disputed.

Methods:We collected all cases of MPN pts followed in our institution who received IFN therapy since 2000, and who discontinued treatment. Cumulative incidences of relapse, hematological transformation and thrombosis were based on competing risks estimator. Survival curves and event free survival (EFS, i.e. time to first event including transformation, thrombosis and death) curves from IFN start were plotted based on the Kaplan Meier method. Univariate logistic models were used to predict complete hematological response (CHR) without any CRT; Cox models were used to predict EFS, and cause-specific hazard of relapse. Multivariable models were then fitted, with a stepwise selection procedure. All tests were two-sided with p-values ≤0.05 denoting statistical significance.

Results:A total of 149 out of 333 IFN treated MPN pts had discontinued therapy at time of analysis. 78 had polycythemia vera (PV), 61 essential thrombocythemia (ET), and 10 myelofibrosis (MF). Reasons for choosing IFN in these 149 pts were young age in 75 (50%), resistance / intolerance to previous therapies in 40 (27%), pregnancy in 4 and other in 30. Median age at IFN start was 49.5 years [interquartile range, IQR: 39-60] and median [IQR] time since diagnosis of MPN was 2 years [0.6-6.8]. 64 pts had a history of vascular event.

123 pts were JAK2V617F+, 15 had CALR, 5 MPL and 1 JAK2 exon 12 mutations, and 5 were triple-neg. Among 47 pts tested by NGS, 15 had additional mutations.

Best response to IFN was CHR in 123 (83%), partial response (PR) in 16, and failure in 10 pts. Reasons for IFN discontinuation were toxicity in 79 (53%), prolonged hematological CHR in 52 (35%), and other in 18. Median IFN treatment duration was 28.3 months (mos) [10-45]. At discontinuation, 114 (77%) pts were in CHR, 17 (11%) in PR, and IFN failed in 18 (12%).

At IFN start, median mutant allele burdenwere 25% [IQR: 15-48] and 38% [34-43] for JAK2 and CALR resp. At IFN discontinuation, median %JAK2 and %CALR had decreased to 10% [3-25], and 13.5% [6-43], resp.

At last follow up, 109 (73%) pts were alive in CHR for a median time after IFN discontinuation of 42 mos [23-58], including 46 (42%) without CRT after 33 mos [19-53]. Based on a multivariable logistic model, the CHR without CRT was associated with mutant allele burden at IFN discontinuation (OR=0.95, 95%CI, 0.95-0.96, P=0.034), stopping for toxicity reason (OR= 0.82, 95%CI, 0.75-0.89, P= 0.009), and cumulative dose of IFN (OR=1.01/mg, 95%CI, 1.003-1.02, P= 0.012).

Hematological relapse occurred in 49 pts. IFN was restarted in 25 of them and all achieved hematological response, including 72% CHR. Based on a multivariable Cox model, the cause-specific hazard of relapse was decreased in ET (HR= 0.4, 95%CI, 0.2-0.8; P= 0.013) while increased with the baseline WBC count (HR= 1.07, 95%CI, 1.02-1.14, P= 0.011).

Ten pts had a thrombotic event after IFN, with a 5 year cumulative incidence of 10% (95%CI, 4-16%). A total of 11 pts transformed to MF, 3 to MDS / AML, and 2 from ET to PV; the 5-year cumulative incidence of transformation was 3.8% (95%CI, 0.4-7.2%). Six patients had died (4 after transformation). Overall, 25 patients developed a thrombosis, transformation or death. The 5- and 10-year EFS were 92% (95%CI, 87-97) and 72% (95%CI, 62-84), resp. In multivariate model, only time between diagnosis and IFN start (HR=1., 95%CI, 1.002-1.016, P=0.011) and mutant allele burden at IFN discontinuation (HR=1.03, 95%CI, 1.016-1.048, P<0.0001) remained jointly associated with the outcome. Pts with an allele burden ≥ 25% at discontinuation had a higher hazard of event (HR= 3.4, 95%CI, 1.3-9.3, P=0.0145).

Conclusion: This study allowed for the first time identification or several parameters associated with major outcomes after IFN therapy. In particular, a lower mutant allele burden at IFN discontinuation was associated with persistent complete hematologic response without therapy and longer event-free survival, indicating a clinical relevance in obtaining a deeper molecular response. Our results also show that late onset of IFN therapy after MPN diagnosis is associated with poorer outcome.