BMS-911543

The management of symptomatic splenomegaly in myelofibrosis (MF) was revolutionized by the discovery of the JAK2V617F mutation and subsequent development of JAK2 inhibitors (JAKi) [1-4]. Currently, four JAKi are approved by the Food and Drug Administration (FDA) for treatment of MF: ruxolitinib, fedratinib, pacritinib, and momelotinib which alleviate symptomatic splenomegaly and constitutional symptoms. In the COMFORT I (ruxolitinib vs. placebo), and COMFORT II studies (ruxolitinib vs. best available therapy), spleen volume reduction of > 35% was achieved in 41.9% and 28% of ruxolitinib-treated patients, respectively [5, 6]. However, the benefits of JAKi therapy are often transient, and in the COMFORT-I and II trials, 55% and 75% of patients had discontinued ruxolitinib at 3-year and 5-year follow-up, respectively [4, 7, 8]. We have previously shown that patients with ruxolitinib-resistant splenomegaly derive limited benefit from switching to fedratinib [9], and involved-field radiotherapy or splenectomy is often advised as a bridge to allogeneic stem cell transplant (ASCT) [2, 10]. In this study, we describe the outcomes of 34 patients with MF that underwent splenectomy after exposure to JAKi therapy. After institutional board review approval, study patients were identified from the Mayo Clinic (USA) clinical database based on the following criteria: (i) diagnosis of MF according to the International Consensus Classification [11]; (ii) documentation of JAKi therapy exposure followed by splenectomy; (iii) availability of information before and after splenectomy. We identified 39 patients with prior JAKi treatment. Two patients were excluded due to primary indication for splenectomy, with one undergoing splenectomy due to trauma and the other for metastatic solid tumor. Three patients were excluded due to lack of follow-up information after splenectomy. Ultimately, 34 patients were included who underwent an open (n = 29) or laparoscopic (n = 5) splenectomy between 2011 and 2024. Transfusion independence was defined as more than 12 weeks without a red blood cell transfusion and hemoglobin ≥ 1.5 g/dL. [12] Major bleeding event was defined by criteria from the International Society on Thrombosis and Haemostasis [13]. Clinical and laboratory variables at the time of MF diagnosis, at splenectomy, and 3 and 12 months post-procedure were collected. Conventional statistical methods were employed for data analysis (JMP Pro 17.0.0, SAS Institute, Cary, NC, USA). A total of 34 patients with MF (primary MF [PMF, n = 18], post-polycythemia vera MF [post-PV MF, n = 15], and post-essential thrombocythemia MF [post-ET MF, n = 1]); median age; 63 years, 59% males, median spleen size; 25.1 cm (14.7–36) who underwent splenectomy following treatment with one or more JAKi were included. Baseline characteristics and prior JAKi exposure at the time of splenectomy are outlined in Table S1. JAK2V617F mutation was detected in 21 (62%), CALR mutation in 9 (26%), MPL in 1 (3%) case, while 3 (9%) were triple-negative. The majority of patients had received one JAKi (n = 26; 76%), six (18%) received two, and two (6%) patients received three JAKi. Specific JAKi included ruxolitinib in 27 (79%), momelotinib in 5 (15%), fedratinib in 3 (9%), pacritinib in 3 (9%), and BMS 911543 JAKi as part of a clinical trial in 5 (15%) patients. Prior treatment also included splenic radiation in 4 patients (median 0.6 years [0.2–1.2] before splenectomy); preoperative splenic embolization was performed in 4 patients; one patient underwent embolization 10 months prior to splenectomy. Median interval from MF diagnosis to splenectomy was 10.3 years (1.0–35.4) and from initiation of JAKi therapy to splenectomy was 1.8 years (0.03–12.1). Most common indications for splenectomy included symptomatic splenomegaly in 27 (79%) patients, transfusion-dependent anemia in 15 (44%), and in preparation for ASCT in 14 (41%) patients. Overall, 23 (68%) patients discontinued JAKi before splenectomy, whereas the remaining 11 patients continued on JAKi during/after surgery. Postoperative complications were reported in 21 (62%) patients which included major bleeding in 10 (29%); 2 (6%) were fatal events, and 7 (21%) required repeat surgical exploration. Thrombosis occurred in 9 (26%) cases and 5 (15%) patients experienced both major bleeding and thrombotic events. Other complications included infection (n = 5), renal injury (n = 3), and pneumothorax (n = 2). Complication rates were not significantly different among patients with PMF versus post-PV/ET MF (78% vs. 50%; p = 0.09), exposed to one versus two or more JAKi (71% vs. 50%; p = 0.29), in patients who discontinued versus continued JAKi during/after surgery (64% vs. 61%; p = 0.88). At a median follow-up of 1.7 years (0.01–10.2) post-splenectomy, a total of 15 (44%) deaths were documented. Median post-splenectomy survival was 4.3 years, with 30-day and 60-day mortality rates of 6% and 15%, respectively (Figure 1A), and was not impacted by ongoing JAKi therapy during/after splenectomy (not reached vs. 4.1 years; p = 0.44). Causes of death were known in eleven patients and included post-operative complications (n = 2), disease progression (n = 4), infection (n = 3), graft versus host disease (n = 1), and cardiac arrest (n = 1). Fourteen patients underwent splenectomy in preparation for transplant of which 12 (86%) were successfully bridged to ASCT, at a median of 3 months (1.5–26) after splenectomy. Three of 12 (25%) transplanted patients continued to receive JAKi post-splenectomy until ASCT. Median survival calculated from time of splenectomy was superior in transplanted patients (median not reached versus 2.6 years in non-transplanted; 5-year survival 65% vs. 22%; p = 0.02) (Figure 1B). Importantly, at a median posttransplant follow-up of 2.6 years (0.3–9.8), 7 of 12 (58%) patients remain alive in complete remission, 2 (17%) have experienced relapse, and 3 (25%) succumbed to infection (n = 2) and graft-versus-host disease (n = 1). One patient who experienced morphologic relapse, 1.3 years posttransplant, received donor lymphocyte infusion (DLI) and remains disease-free, 1-year post-relapse. A second patient with molecular relapse at 1 year following transplant also achieved remission after two DLI, and remains disease-free, 1.8 years after relapse. Among 22 patients who did not receive ASCT after splenectomy, 10 (45%) had transfusion-dependent anemia at the time of splenectomy of which 5 (50%) became transfusion-independent (TI) (≥ 12 weeks without transfusion) with median TI response duration of 2.8 years (0.3–7.1). In addition, among 5 non-transplanted patients with baseline thrombocytopenia (< 50 × 109/L [16–42 × 109/L]), 4 (80%) achieved platelet count > 100 × 109/L (median; 170 × 109/L, range: 119–279 × 109/L) within 3 months post-procedure. Median blood counts at splenectomy, 6 weeks to 3 months and 1-year postsplenectomy are depicted in Figure 1C–E. The management of JAKi-resistant splenomegaly poses a significant clinical challenge. Although ASCT is advised in such situations, splenectomy is often considered prior to ASCT, due to concerns regarding graft failure and post-transplant relapse [14]. In a European Society for blood and marrow transplantation (EBMT) analysis of 1195 MF patients transplanted between 2000 and 2017, splenectomy was performed prior to ASCT in 202 (17%), of which prior ruxolitinib exposure was documented in 36 (20%); non-relapse mortality was significantly lower in splenectomized versus non-splenectomized patients, however, risk of relapse was higher, with no significant impact on overall survival [15]. In the particular study, in 333 patients with exposure to ruxolitinib, 36-month overall survival was superior at 71% for patients undergoing splenectomy versus 47% in non-splenectomized patients with spleen size ≥ 15 cm. In the particular study, non-relapse mortality was lower with splenectomy at 24% versus 42%, while risk of relapse was similar [15]. In a historical cohort which included 314 patients with MF that underwent splenectomies between 1976 and 2004, perioperative complications occurred in 87 (28%) including infection (10%), thrombosis (10%), or bleeding (14%), and 7% were fatal events [16]. In this study, splenectomy was associated with significant perioperative complications in approximately two-thirds of patients with a 30-day mortality rate of 6%. Nonetheless, ASCT was feasible in the majority (86%) of patients being prepared for transplant with over half of transplanted patients remaining alive and in complete remission. Similar to our findings, in a prior study which included 12 patients previously treated with ruxolitinib at 30–45 mg daily for at least 3 months, who underwent splenectomy in preparation for ASCT, perioperative morbidity was 50% without any deaths, 8 (75%) patients proceeded to subsequent ASCT; 7 patients remain alive and in complete remission and no relapses after ASCT were observed [17]. Prior studies have also shown improvement in transfusion-dependent anemia and thrombocytopenia following splenectomy which is consistent with our observations [2, 18]. In summary, this study underscores the therapeutic value of splenectomy as a bridge to ASCT, as well as a palliative strategy to ameliorate cytopenias in MF patients previously treated with JAKi. Furthermore, additional investigations are required to identify predictors of postoperative complications, in order to guide management strategies, and also determine the optimal approach to the use of JAKi in the post-splenectomy and peri-transplant setting [19]. P.C. and N.G. designed the study, collected data, performed analyses, and wrote the article. P.S. provided surgical expertise on splenectomies. A.P. contributed patients. A.T. provided expertise on myelofibrosis and current recommendations. All authors reviewed the final draft of the article. N.G. has served on the Advisory Board for DISC Medicine and Agios. Data available through email to the corresponding author. Table S1. Presenting clinical characteristics, mutation details, and prior treatment of 34 patients with myelofibrosis. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.