Prospective, Multicenter Clinical Trial of Everolimus as Primary Therapy in Waldenstrom Macroglobulinemia (WMCTG 09-214)

The PI3K/AKT pathway is an important survival signaling cascade that supports the growth and survival of malignant lymphoplasmacytic cells in Waldenstrom macroglobulinemia (1). Activating mutations in MYD88 and CXCR4, found in 95% and 30% of Waldenstrom macroglobulinemia patients, respectively, trigger PI3K/AKT signaling (2–5). Everolimus is an orally administered inhibitor of mTOR, a serine–threonine kinase that is downstream of the PI3K/AKT signaling pathway (6). Everolimus is approved by the FDA for the treatment of several solid malignancy indications and shows in vitro activity against Waldenstrom macroglobulinemia cells (6). In previous work, we evaluated the activity of everolimus in previously treated Waldenstrom macroglobulinemia patients (7). The overall response rate (ORR) using consensus criteria was 73%, and 50% of patients achieved a major response (8). The median progression-free survival (PFS) was 21 months in this study (9). Toxicities were common, with grade 3 or higher adverse events observed in 67% of patients (8, 9). We therefore examined the safety and efficacy of everolimus in previously untreated, symptomatic Waldenstrom macroglobulinemia patients. We performed serial bone marrow biopsies to more fully delineate the impact of therapy on Waldenstrom macroglobulinemia disease burden and also assessed the impact of MYD88 and CXCR4 mutations on treatment outcome. We present herein the first report from this prospective, multicenter study.

Symptomatic Waldenstrom macroglobulinemia patients requiring therapy based on consensus recommendations (10) and who were previously untreated were eligible to enroll. To meet eligibility, patients must have had a platelet count of ≥75 × 10⁹/L, absolute neutrophil count of ≥1.5 × 10⁹/L. Patients with symptomatic hyperviscosity, known history of active or chronic hepatitis B or C, uncontrolled diabetes, or severe or uncontrolled medical conditions that prohibited study participation were excluded.

The study was conducted by the Waldenstrom macroglobulinemia Clinical Trials Group (WMCTG Protocol 09-214). All patients provided informed written consent approved by the local Institutional Review Boards. Intended therapy consisted of self-administered oral everolimus (10 mg/day). Patients were treated until progression or unacceptable toxicity. Each treatment cycle was 4 weeks. Dose reduction for grade 3 or higher adverse events was permitted as follows: 7.5 mg daily, 5.0 mg daily, and 5.0 mg every other day for the first, second, and third dose deescalations, respectively. Dose reescalation was not permitted. Retreatment was permitted once adverse events resolved to grade 1 or less. Filgrastim or transfusional support was permitted to treat hematologic adverse events. Patients were strongly encouraged to use 5 mL of an oral dexamethasone solution (0.5 mg/5mL) to swish and spit up to four times daily for prevention of stomatitis during the first 3 months of everolimus therapy.

Baseline studies consisted of complete blood counts and differential, quantitative serum IgM levels, serum protein electrophoresis, a bone marrow biopsy and aspiration, CT scans of the chest, abdomen and pelvis (CAP), serum electrolytes, liver function tests, amylase, lipase, blood urea nitrogen, creatinine, and serum β₂-microglobulin levels, lipid panel, and glucose. Patients were assessed for efficacy and toxicity on the first day of cycles 2, 3, and thereafter every 12 weeks. A bone marrow biopsy and aspiration, and CT scans of CAP (if extramedullary disease was present at baseline) were required at 24 weeks, and thereafter as clinically indicated, including to confirm complete response (CR) or suspected disease progression.

MYD88 and CXCR4 genotyping was performed for patients enrolled at the Dana Farber Cancer Institute (DFCI; Boston, MA). An allele-specific PCR assay was used for the determination of MYD88^L265P using DNA isolated from CD19-selected bone marrow cells as described previously (1). CXCR4^WHIM mutation status was determined by AS-PCR and Sanger sequencing of CD19-selected bone marrow cells (2).

The primary endpoint was determination of ORR. Response determinations were made using consensus criteria adapted from the Sixth International Workshop on Waldenstrom macroglobulinemia (11). Secondary endpoints included determination of time to progression and assessment of safety and tolerability of everolimus. Sample size was predicated on an expected ORR of ≥70% and a minimal acceptable response rate of 50% based on assumptions derived from our previous published experience with everolimus (8, 9). PFS was defined as the time between initiation of therapy and date of progression, death, or last follow-up. Patients without disease progression (including those taken off study for toxicity) were censored at the date of their last evaluation. For categorical univariate analyses, the Kaplan–Meier method for incomplete observations was used to estimate PFS curves, which were compared using the log-rank test. A Wilcoxon rank-sum test was used for analysis of pre- and posttherapy continuous variables. For categorical variables, a two-tailed Fisher exact test was used. P values ≤0.05 were considered statistically significant. All graphics and calculations were obtained using STATA 13.1 (StataCorp LP).

Thirty-three patients were enrolled, and their baseline characteristics are shown in Table 1. The first patient started on therapy on December 9, 2009, and the last patient started therapy on June 13, 2011. The last patient came off study for disease progression on July 14, 2016. The median number of treatment cycles administered was 10.5 (range, 1–64.6). All participants are off treatment.

Median IgM levels for all 33 patients declined from 4,440 mg/dL (range, 959–10,256 mg/dL) at baseline to 1,360 (range, 146–7,100 mg/dL) at best response (P < 0.0001). Pretherapy, 23 of 33 (69.7%) patients had a serum IgM level ≥3,000 mg/dL; following treatment, 8 of 33 (12.9%) patients had a serum IgM level ≥3,000 mg/dL (P = 0.0004). Among 24 patients with serial bone marrow biopsies, the median bone marrow involvement declined from 75% (range, 2%–95%) to 52.5% (range, 6%–95%) at best response (P = 0.03). Discordance between serum IgM levels and bone marrow disease burden was common (Fig. 1). For 24 patients who had a repeat bone marrow biopsy by week 26, the median change in serum IgM was −47.9% (range, 3.8% to −88.5%), whereas the synchronous change in bone marrow disease burden was −13.3% (range −36.8%–200%; r = 0.23; P = 0.26). In 8 (33.3%) of these patients, the bone marrow disease burden had increased, with a median increase of 43% (range, 25%–200%), whereas the serum IgM level decreased by a median of 41% (range, −75.7%–3.8%). No serum IgM flare was observed.

Following treatment, the median hemoglobin level for all patients rose from 10.8 to 12 g/dL (P = 0.001) at best response. Categorical responses were as follows: very good partial response (VGPR; n = 1), partial response (PR; n = 19), minor response (n = 4), for an ORR and major response (PR or better) rate of 72.7% [95% confidence interval (CI), 57.5%–87.9%] and 60.6% (95% CI, 43.9%–77.3%), respectively. No patient achieved a CR. Overall and major response rates were not impacted by the International Scoring System for Waldenstrom macroglobulinemia score (12).

MYD88 and CXCR4 somatic mutation status was evaluable in 21 patients. Twenty (95.2%) genotyped participants expressed MYD88^L265P. The overall and major response rates in these patients were 71.4% and 52.4%, respectively. One patient with MYD88 wild type did not respond. CXCR4^WHIM mutations were present in 4 of 21 (19.2%) genotyped participants, all of whom were MYD88^L265P mutated. The ORR for patients with CXCR4 wild type and WHIM mutations were 81.3% and 50%, respectively (P = 0.25). Major response attainment occurred in 62.5% and 25% of patients with CXCR4 wild type and WHIM mutations, respectively (P = 0.56). Among responders, the median time to at least a minor response was 1 month (range, 1–9 months), whereas the median time to a major response was 2 months (range, 1–27 months).

With a median follow-up of 13.1 months (range, 1.6–64.6 months), all patients were alive. All patients are off treatment. The Kaplan–Meier curve for PFS for all patients appears in Fig. 2. The median time to progression was 21 months (95% CI, 11–39 months). By univariate analysis, neither baseline bone marrow disease burden >50% versus ≤50% (P = 0.95), serum IgM >4,000 versus ≤4,000 mg/dL (P = 0.59), nor hemoglobin >11 versus ≤11 g/dL (P = 0.16) levels were associated with increased risk for progression. Dose deescalation (<10 mg/day vs. 10 mg/day) was also not associated with increased risk of progression (P = 0.34). Categorical response attainment was associated with progression risk. Patients achieving a PR or better had a longer PFS versus those with less than a PR (33 vs. 5 months, respectively; HR = 4.350; 95% CI, 3.9–113 months, P = 0.0011).

Grade 2 or higher toxicities that were at least possibly related to protocol therapy are presented in Table 2. The most common nonhematologic treatment-related grade ≥2 toxicities included mucositis (27.3%), infection (21.2%), rash (21.2%), fatigue (18.2%), and pneumonitis (18.2%). Treatment-related hematologic toxicities that were at least grade ≥2 included anemia (27.3%), neutropenia (18.2%), and thrombocytopenia (15.2%). Dose reduction due to adverse events occurred in 9 patients, with dose reduction to 7.5 mg (n = 5) and 5.0 mg (n = 2) daily, and 5.0 mg (n = 2) every other day. Treatment was discontinued in 32 patients, with reasons for discontinuation as follows: nonresponse (n = 6), progressive disease (n = 15), pneumonitis (n = 5), withdrawal of consent (n = 3) that included one patient for recurring grade 2 stomatitis, noncompliance (n = 2), and prolonged study drug hold for unrelated infection.

Following discontinuation of everolimus, rapid increases in serum IgM levels were common. In 7 patients, serum IgM levels showed at least a doubling. In these patients, serum IgM levels increased from a median of 1,410 (range, 306–2,880 mg/dL) to 4,670 (range, 2,510–5,910 mg/dL) at a median of 39 days (range, 18–95 days) following discontinuation of everolimus. Four of these patients underwent plasmapheresis, including two for symptomatic hyperviscosity, and two to prevent a rituximab-related IgM flare with subsequent therapy.

At baseline, median serum IgA and IgG levels were 99 and 884 mg/dL, respectively. Following therapy, at last individual patient assessment, median serum IgA and IgG levels declined to 44 and 447 mg/dL, respectively (P = 0.49 and P = 0.23, respectively).

In this prospective, multicenter study, we examined the single-agent activity of everolimus in symptomatic, untreated Waldenstrom macroglobulinemia patients. Everolimus inhibits mTOR, a serine–threonine kinase that contributes to PI3K/AKT-directed growth and survival signaling in Waldenstrom macroglobulinemia (6, 7). Activating mutations in both MYD88 and CXCR4 trigger PI3K/AKT signaling (1, 2). Our findings showed that everolimus was associated with overall and major responses in 72.7% and 60.6%, respectively, using consensus criteria (11). However, the frequent finding of IgM discordance complicated response interpretation, as consensus criteria for response in Waldenstrom macroglobulinemia primarily rely on changes in serum IgM levels. IgM discordance has been observed with other therapeutics used to treat Waldenstrom macroglobulinemia patients, including rituximab and ofatumumab that can increase, whereas bortezomib and ibrutinib can decrease serum IgM levels independent of changes in bone marrow tumor burden (13–17). A bystander effect for the “IgM flare” by rituximab has been proposed wherein immune cells release IL6 through the interactions with the Fc domain of rituximab prompting IgM release by Waldenstrom macroglobulinemia cells (18). The BTK substrate STAT5A regulates IgM secretion in Waldenstrom macroglobulinemia cells, and its selective inhibition by ibrutinib has been proposed to contribute to its discordant findings (17). However, the PI3K/AKT pathway is not a known contributor to STAT5A signaling, and other on-target signaling events, as well as off-target effects that contribute to IgM production and secretion could be impacted by everolimus.

The impact of MYD88 and CXCR4 mutations on response activity was also investigated. Genotyping was performed for 21 patients (all at DFCI). Although the study numbers are small for a meaningful analysis, the one patient with MYD88 wild type was a nonresponder, while the overall and major response rates were lower in CXCR4 mutated patients. Wild-type MYD88 and CXCR4^WHIM mutation status were also associated with lower overall and major clinical responses to ibrutinib (17). In preclinical studies, Waldenstrom macroglobulinemia cells transduced with CXCR4^WHIM receptors showed resistance to everolimus as well as ibrutinib following SDF-1a stimulation (5, 19). The use of CXCR4 antagonists sensitized Waldenstrom macroglobulinemia cells transduced with CXCR4^WHIM receptors to the effects of everolimus and ibrutinib in these studies. Combination studies with CXCR4 antagonists may be of interest with everolimus, as well as other agents, such as ibrutinib, whose responses are impacted by CXCR4 mutations.

The median PFS observed in this study of untreated patients was similar (21 months) to that observed in our previous, multicenter study of everolimus in previously treated Waldenstrom macroglobulinemia patients. Patients with major responses exhibited longer PFS than those with minor responses or stable disease (i.e., 33 vs. 5 months). Only one patient attained a VGPR, and none a CR precluding an analysis on the impact of deeper (i.e., VGPR or better) categorical attainment on PFS. The lack of CR attainment observed in this trial is not uncommon in Waldenstrom macroglobulinemia and may reflect the broad prosurvival signaling cascades induced by MYD88 and/or CXCR4-activating mutations that include canonical NFκB, PI3K/AKT, and MAPK/ERK signaling (4, 5).

Drug-related toxicity was responsible for discontinuation of everolimus in 9 (27%) patients, including 5 for pneumonitis. In total, 6 patients (18%) experienced treatment-related pneumonitis that resulted in hospitalization for 3 patients. The incidence of grade 2 or higher treatment-related pneumonitis was higher in this study versus that observed by us in our study with everolimus monotherapy in previously treated Waldenstrom macroglobulinemia patients (18% vs. 8%). Treatment-related pneumonitis has also been reported with idelalisib and may represent a class effect for therapeutics that target PI3K/AKT signaling. A higher incidence of autoimmune-related events, including pneumonitis, has also been observed with idelalisib in first-line versus previously treated CLL patients (20). The impact of previous treatment status on Treg immune cell function may have contributed for differences in autoimmune activity between first-line and previously treated patients exposed to idelalisib and may be relevant in Waldenstrom macroglobulinemia patients undergoing everolimus therapy (20). Steroids were effective in the treatment of everolimus-related pneumonitis, which resolved in all patients. Oral mucositis, a known side effect of everolimus, was less pronounced (no grade ≥3 events) in this study compared with our experience in previously treated patients and likely reflected the routine use of an oral swish and spit dexamethasone solution during the first 3 months of everolimus therapy as a preventative. Despite this measure, mucositis occurred at a grade 2 level in 27% of patients and contributed to early study withdrawal for one patient. Minimal systemic steroid absorption was likely associated with the oral swish and spit dexamethasone solution. Cytopenias were also commonly observed adverse events, particularly grade 2 or higher anemia that was seen in 27% of patients.

The study findings show that everolimus is active in previously untreated Waldenstrom macroglobulinemia patients. Discordance between serum IgM levels and bone marrow disease burden is remarkable, and cessation of everolimus can produce rapid rebounds in serum IgM levels. Toxicity resulted in premature discontinuation of therapy in 27% of patients and included pneumonitis that appeared more common in this study compared with our experience in previously treated Waldenstrom macroglobulinemia patients. The risks and benefits of everolimus should carefully be weighed against other available treatment options for the primary therapy of Waldenstrom macroglobulinemia. Current NCCN and Waldenstrom macroglobulinemia consensus guidelines support use of everolimus in relapsed/refractory Waldenstrom macroglobulinemia disease and appear reasonable in view of the risk/benefit identified for everolimus in this first-line study (21, 22).

No potential conflicts of interest were disclosed.

Conception and design: S.P. Treon, I.M. Ghobrial

Development of methodology: S.P. Treon

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.P. Treon, C. Tripsas, L.T. Heffner, H. Eradat, A.Z. Badros, G. Yang, J.J. Castillo, J. Matous, I.M. Ghobrial

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.P. Treon, K. Meid, C. Tripsas, Z.R. Hunter, J. Gustine, J.J. Castillo, I.M. Ghobrial

Writing, review, and/or revision of the manuscript: S.P. Treon, K. Meid, C. Tripsas, L.T. Heffner, H. Eradat, A.Z. Badros, J.J. Castillo, I.M. Ghobrial

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.P. Treon, K. Meid, C. Tripsas, L. Xu, C.J. Patterson, J. Gustine

Study supervision: S.P. Treon, A.Z. Badros

Other (accrual of patients): J. Matous

This work was supported by Novartis Inc., that provided the study drug and research funding, Peter Bing, the Edward and Linda Nelson Fund for WM Research, the D'Amato Fund for Genomic Discovery, and the Bauman Fund for Waldenstrom's Macroglobulinemia Research.