Randomized, placebo‐controlled, phase 3 study of perifosine combined with bortezomib and dexamethasone in patients with relapsed, refractory multiple myeloma previously treated with bortezomib

Abstract Perifosine, an investigational, oral, synthetic alkylphospholipid, inhibits signal transduction pathways of relevance in multiple myeloma (MM) including PI3K/Akt. Perifosine demonstrated anti‐MM activity in preclinical studies and encouraging early‐phase clinical activity in combination with bortezomib. A randomized, double‐blind, placebo‐controlled phase 3 study was conducted to evaluate addition of perifosine to bortezomib‐dexamethasone in MM patients with one to four prior therapies who had relapsed following previous bortezomib‐based therapy. The primary endpoint was progression‐free survival (PFS). The study was discontinued at planned interim analysis, with 135 patients enrolled. Median PFS was 22.7 weeks (95% confidence interval 16·0–45·4) in the perifosine arm and 39.0 weeks (18.3–50.1) in the placebo arm (hazard ratio 1.269 [0.817–1.969]; P = .287); overall response rates were 20% and 27%, respectively. Conversely, median overall survival (OS) was 141.9 weeks and 83.3 weeks (hazard ratio 0.734 [0.380–1.419]; P = .356). Overall, 61% and 55% of patients in the perifosine and placebo arms reported grade 3/4 adverse events, including thrombocytopenia (26% vs 14%), anemia (7% vs 8%), hyponatremia (6% vs 8%), and pneumonia (9% vs 3%). These findings demonstrate no PFS benefit from the addition of perifosine to bortezomib‐dexamethasone in this study of relapsed/refractory MM, but comparable safety and OS.


K E Y W O R D S
Akt inhibition, bortezomib, multiple myeloma, perifosine, proteasome inhibition

INTRODUCTION
The treatment of multiple myeloma (MM) has been transformed over the past two decades with the introduction of novel targeted agents, such as proteasome inhibitors, immunomodulatory drugs, and monoclonal antibodies, and as a result of our increasing understanding of the complex disease biology of MM and the signaling pathways of importance [1][2][3][4][5]. Proteasome inhibitors and immunomodulatory drugs have emerged as backbone therapeutics for MM treatment as a result of the substantial efficacy demonstrated alone and in combination in different disease settings [6]. This activity arises due to these mechanisms of action affecting multiple critical signaling pathways of importance [7,8]. However, as MM progresses, patients can develop resistance to these and other commonly used agents. Therefore, an ongoing unmet need for patients with relapsed and/or refractory MM (RRMM) is for novel targeted agents that inhibit specific pathways of relevance and potentially synergize with or overcome resistance to the mechanisms of action of the proteasome inhibitors and the immunomodulatory drugs [9][10][11].
One of the pathways of interest for targeting in the treatment of MM has been the phosphoinositide-3-kinase (PI3K)/Akt signaling cascade, the activation of which is induced by interactions between MM cells and bone marrow stromal cells within the bone marrow microenvironment [12,13]. The induction of this and other signaling pathways results in MM proliferation, survival, and drug resistance [13], making it a rational therapeutic target in MM. Perifosine is an investigational, oral, synthetic alkylphospholipid that inhibits or modifies signal transduction pathways of relevance in MM including PI3K/Akt, nuclear factor-B, and c-Jun N-terminal kinase (JNK) cascades [13]. Perifosine demonstrated potent anti-MM activity in preclinical studies [14][15][16][17][18], including enhanced cytotoxicity in combination with bortezomib based on synergism between mechanisms of action [19].
Consequently, perifosine was investigated clinically in MM and demonstrated encouraging activity in patients with MM when combined with bortezomib and lenalidomide [20,21]. In a phase 1/2 study in 73 patients with RRMM, perifosine in combination with bortezomib, with or without added dexamethasone, resulted in a rate of minimal response or better of 41%, including rates of 65% in patients who had relapsed following prior bortezomib treatment and 32% in bortezomib-refractory patients [21]. Based on these promising early-phase study results, we conducted a randomized, double-blind, placebo-controlled phase 3 study to evaluate the benefit of adding perifosine to bortezomib-dexamethasone in MM patients who had previously relapsed after a bortezomib-based regimen.

Study design
This was a double-blind, placebo-controlled, randomized phase 3 study (NCT01002248). Patients were randomized in a 1:1 ratio to receive perifosine 50 mg orally, once-daily, or matching placebo, plus intra-  confidence interval (CI) was reported. The hazard ratio (HR) and its 95% CI were also reported. The secondary endpoint of OS was analyzed similarly to PFS. Other data were summarized using descriptive statistics.

Patients
At the data cut-off for the first interim analysis (March 12, 2013), a total of 135 patients had been enrolled and randomized -69 to the perifosine arm and 66 to the placebo arm ( Figure 1). Patient demographics and disease characteristics at randomization were similar between treatment arms (Table 1), except for a numerically higher proportion of patients aged <65 years in the perifosine arm versus the placebo arm (61% vs 42%). Prior treatment exposure and relapsed/refractory status were balanced between arms.

Efficacy
The primary endpoint of PFS was assessed in the intent-to-treat population, with the exception of one patient in the placebo arm. This patient was not evaluable for time-to-event outcomes; subsequent to randomization, the patient was found not to meet eligibility criteria based on M-protein levels required by the protocol, and so they were excluded  Figure 2A.
An additional 18 and 11 patients, respectively, achieved a minimal response or stable disease, giving clinical benefit rates (stable disease or better) of 46% and 44%.

Safety
All 135 patients were evaluable for safety. In total, 60 (87%) of patients in the perifosine arm and 50 (75.8%) of patients in the placebo arm reported at least one AE; the most common events are summarized in Table 3. Among these patients, 43 (62%) and 38 (58%) reported at least one grade ≥3 AE; the most common individual grade 3/4 AE events in the perifosine versus placebo arms included thrombocytopenia (26% vs 14%), anemia (7% vs 8%), hyponatremia (6% vs 8%), and pneumonia (9% vs 3%; Table 3). Rates of grade 3 neuropathy AEs were limited, including two patients (3%) on each arm with peripheral neuropathy and only one patient (2%) on the placebo arm having a grade 3 AE of peripheral sensory neuropathy; no grade 4 neuropathy AEs were reported. However, four patients (6%) on the perifosine arm and one patient (2%) on the placebo arm discontinued treatment due to an AE of neuropathy or neuralgia. Overall, 13 (19%) and 9 (14%) patients on the perifosine and placebo arms, respectively, discontinued treatment due to an AE. Other AEs resulting in treatment discontinuation on the perifosine arm were pneumonia (n = 2), and thrombocytopenia, oral pain/toothache, meningioma, bronchitis, asthenia, diarrhea, and ocular hyperemia (each n = 1), and on the placebo arm were respiratory failure/sepsis, nausea, congestive heart failure, fatigue, diarrhea/asthenia/thrombocytopenia, back pain, abdominal pain, and muscular weakness/pain in extremity (each n = 1).
A total of four patients (6%) on the perifosine arm and three patients (5%) on the placebo arm died during the study reporting period due to reasons other than disease progression, including respiratory infection, suicide, sudden death (treatment-emergent grade 5 AE), and myocardial infarction (each n = 1) on the perifosine arm, and respiratory failure/sepsis (treatment-emergent grade 5 AE), congestive heart failure, and pneumococcal sepsis (treatment-emergent grade 5 AE) (each n = 1) on the placebo arm.

DISCUSSION
The findings from the first planned interim analysis of this randomized, discontinued following this first planned interim analysis. Thus, with limited follow-up, further interpretation of these findings is restricted.
In particular, OS data were immature at the time of this analysis, with only 25% and 29% of patients in the perifosine and placebo arms, respectively, having died. Median OS was numerically longer in the perifosine arm, but these data were based on the tail-ends of the Kaplan-Meier distribution curves and must therefore be interpreted with caution.
The response rates reported in the present study appeared low compared to data from the previous phase 1/2 study of perifosine plus bortezomib-dexamethasone in RRMM [21]; in the previous study, an overall response rate of 45% was reported, compared to 20% in the present study, and all patients achieved stable disease or better, compared to 46%. However, these previous data were based on a cohort of only 20 bortezomib-relapsed patients who had received a median of 4 prior therapies, suggesting ongoing sensitivity to drug therapy in these patients. By contrast, the data from the present study suggest that the enrolled patients had relatively resistant disease, with almost a quarter stratified as refractory to other, non-bortezomib-based prior therapies. Additionally, the discrepancy between the phase 2 and phase 3 experiences may have been due to the small number of patients in the previous open-label phase 2 study; this may not have been optimal for subsequent phase 3 exploration -a larger, randomized phase 2 study or an adaptive study design may have been preferable and may potentially have provided response data more similar to those reported from the present study. It is also notable that the response rate in the placebo arm (27%) appeared lower compared to data from a phase 2 study of bortezomib ± dexamethasone retreatment in patients with relapsed MM (40%) [24]. studies related to the PI3K/Akt signaling cascade were prospectively planned; such analyses may have helped identify a patient population who did benefit from the combination [26].
Reflecting the previous phase 1/2 study of perifosine plus bortezomib-dexamethasone [21] the triplet regimen appeared tolerable at the selected dose of perifosine; indeed, it might be queried, given the limited activity seen with the triplet regimen, whether the perifosine dose was optimal in this study, given that no dose-limiting toxicities were reported using a 100 mg dose of perifosine in the previous phase 1/2 study [21]. However, as noted in that report, toxicity was generally greater in this higher-dose cohort, and this impacted the duration of treatment; thus, the 50 mg dose was selected for the phase 2 component of the previous study and for the present study [21]. In the present study, no specific safety concerns were observed were also among the most common grade 3/4 AEs in the previous study [21]. Overall, the low rates of treatment discontinuation and the convenience of the oral approach of the triplet supported the real-world strategy underlying the combination's development [27], as well as the equipoise in the design, given the favorable performance of the control group and the OS data in the experimental arm subsequent to interim analysis.
Subsequent to the discontinuation of this study, perifosine is no longer being investigated as a potential novel targeted agent for patients with RRMM. Reflecting these findings, a similar lack of OS benefit was seen with the addition of perifosine to capecitabine in patients with metastatic colorectal cancer in the phase 3 X-PECT study [28].
Meanwhile, additional therapeutic approaches targeted at other signaling cascades of known importance are being explored in RRMM [9,44], warranting further evaluation.
In conclusion, perifosine at the dose tested did not improve response rates or outcomes in combination with bortezomibdexamethasone in patients with RRMM who have relapsed following previous bortezomib-based treatment. Although the rational combination of Akt pathway inhibition with perifosine and proteasome inhibition with bortezomib demonstrated synergistic anti-MM activity in preclinical investigation [19], this was not reflected in the phase 3 clinical setting, despite earlier-phase clinical studies showing promise.

ACKNOWLEDGMENTS
The authors would like to thank the Research Nurses, staff, and especially the patients and their families for their participation in this study.
The authors would also like to acknowledge the writing assistance of Abbreviations: AE, adverse event; SOC, system organ class; Vd, bortezomib-dexamethasone. * In addition, one patient on the perifosine arm and two patients on the placebo arm had grade 5 AEs, as described in the text.