Ruxolitinib Discontinuation in Myelofibrosis Leads to Increased Morbidity Burden, Underscoring Unmet Need

John O. Mascarenhas, MD

A United States population-based outcome analysis of real-world patients with myelofibrosis who had discontinued ruxolitinib (Jakafi) demonstrated an increase in morbidity burden and identified the risk factors of survival outcomes, according to John O. Mascarenhas, MD.

"Patients who discontinue ruxolitinib have been shown in several studies to have poor outcomes as it relates to survival, with a median survival that's approximately 1 year; that is concerning, because it suggests that we really don't have effective strategies for most of these patients once they have enjoyed success with ruxolitinib and then stopped treatment," said Mascarenhas. "It really highlights a need to understand the disease biology and to develop therapies either in combination with ruxolitinib to extend that duration or benefit, or to salvage patients' outcomes after they come off ruxolitinib."

For the analysis, Mascarenhas and colleagues identified a total of 290 patients with myelofibrosis between 2006 and 2015 who were treated with and then discontinued ruxolitinib. Only those with ≥90 days of medical history prior to index diagnosis and part A, B, and D enrollment at the time of index diagnosis were included. Investigators evaluated for morbidities in the 30 days following ruxolitinib initiation, prior to discontinuation, and after discontinuation. The incidence of cytopenias and efficacy outcomes were assessed from baseline.

The median age of those included in the analysis was 68 years, with an equal proportion regarding gender. Additionally, the median time to ruxolitinib discontinuation was 284 days, with a median follow-up after discontinuation of 70.9 days.

Most of the patients included in the analysis were diagnosed with anemia and more than 30% of patients received red blood cell transfusions during the 30-day time frame prior to and following discontinuation of ruxolitinib. After treatment discontinuation, half of the patients went on to develop cytopenias. Following discontinuation, the median treatment progression-free survival (PFS) and overall survival (OS) were 6.0 months versus 11.1 months, respectively.

Investigators found that the age at ruxolitinib discontinuation (HR, 2.071; 95% CI, 1.320-3.248), Charlston Comorbidity Index score (HR, 1.172; 95% CI, 1.093-1.257), and gender (HR, 1.620; 95% CI, 1.108-2.369) increased the risk of treatment progression or death.

"One of the reasons why this study is important is that it emphasizes the point that therapies evaluated in the second line, following ruxolitinib failure, really need to start considering OS as an important end point, because it's poor," said Mascarenhas. "This has been demonstrated, and patients want to live longer obviously with a good quality of life."

In an interview with OncLive, Mascarenhas, an associate professor of medicine, hematology, and medical oncology at the Icahn School of Medicine at Mount Sinai; director of the Adult Leukemia Program and leader of Clinical Investigation within the Myeloproliferative Disorders Program at Mount Sinai; and a member of the Tisch Cancer Institute, discussed patient characteristics and outcomes following ruxolitinib discontinuation in myelofibrosis and research efforts being made to address a large unmet need in the space.

OncLive: Could you begin by providing some background on the impact of ruxolitinib in the myelofibrosis treatment paradigm?

Mascarenhas: In 2011, ruxolitinib was approved by the FDA for myelofibrosis, and then in 2014, for polycythemia vera; this clearly was a large step forward for myelofibrosis treatment. Prior to ruxolitinib, we really didn't have effective tools to address symptom burden and splenomegaly, which are 2 important aspects of the disease process.

With the commercial availability of ruxolitinib, quality of life, and even outcomes, of patients with myelofibrosis clearly changed. Patients definitely feel better; they have less spleen-related complaints; they can eat more, there's reversal of cachexia, and thus, an improvement in performance; and they are living longer than they would in the absence of ruxolitinib.

That being said, we also realize that all those benefits come with a price tag. There's an element of myelosuppression—cytopenias, anemia, and thrombocytopenia can occur—but otherwise, it's a relatively well tolerated drug. We also now know that there's some degree of infectious risk, whether it be [herpes] zoster, urinary tract infections, or pneumonia. Still, the benefit outweighs the risk of the drug.

Unfortunately, for many patients, that benefit is not long lived. The median time to discontinuation is approximately 3 years, and 85% of patients discontinue by 5 years. Although [ruxolitinib] is a huge step forward in the treatment paradigm for myelofibrosis, there are still clear unmet needs. For example, patients with low platelets at onset, so less than 50,000 [per microliter] or those who have significant anemia and transfusion requirements are 2 [groups that are not] effectively addressed by ruxolitinib. Another very urgent unmet need would be [patients] after ruxolitinib discontinuation in whom survival outcomes are poor.

What was the inspiration behind conducting this retrospective analysis?

Several studies have been conducted to try and understand what happens to patients after ruxolitinib fails; again, this can look different for different people, so even the concise definition of that is not agreed upon. As such, we set out to evaluate that further by looking at 3 claims databases. The idea was to look at this from a little bit of a different angle; [rather than examining this] from the specific patient perspective, we looked at it from a claims-based angle.

The 3 databases included a commercial database, an electronic health record that was linked to a claims database, and then the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database. We assessed patients of various ages. Obviously, the SEER Medicare linked database included those who were 65 years and older, and then the other 2 databases included patients under 65 years, so we really took a broad range of patients. Notably, this was linked with prescriptions and ICD codes rather than the physician's direct experience with the patient at their institution.

The analysis involved a 9-year period, ranging from 2006 to 2015 and the criteria to be included in this study to have had at least 90 days of medical history, a diagnosis of myelofibrosis—primary or secondary as by the inclusion of ICD codes—and having a prescription for ruxolitinib in the database. A total of 290 patients who fit that description were identified across these 3 databases; these were nonoverlapping patients and the median age was about 68 years. Additionally, there was even distribution between men and women and the median time [to] ruxolitinib [discontinuation] was approximately 284 [days]. Patients had a median follow-up of about 70 days of data after [ruxolitinib discontinuation].

What did this analysis reveal?

We were able to show that there's a certain degree of cytopenias in patients [who] start ruxolitinib, but that progressively gets worse during their treatment, and particularly after they come off treatment; I believe this is the first time this has been shown in this setting. For example, when looking at anemia, thrombocytopenia, and neutropenia, the incidence 30 days prior to starting ruxolitinib was 36%, 12%, and 2%, respectively. However, when you look at 30 days prior to stopping ruxolitinib, the [incidence of] anemia rose to 45%, to 12% for thrombocytopenia, and to 18% for neutropenia. If you look at these numbers 30 days after stopping ruxolitinib, those numbers rise to 53% for anemia, 23% for thrombocytopenia, and 10% for neutropenia. These data demonstrate a pattern of worsening cytopenias, which is obviously concerning. [This] includes the 33% of patients who became red blood cell transfusion dependent on ruxolitinib; one-third of the patients [were dependent] at the time of discontinuation.

There is a significant burden and morbidity associated with this disease and being on this drug; unfortunately, at the time of discontinuation, it is significant. If you look at end points that are very meaningful, like median OS, it was approximately 11 months [in this analysis]. We also did an analysis looking at time to median treatment PFS, and that was about 6 months. These data speak to the fact that outcomes are not good and that we don't really have effective treatment available to alter that course. We also know that cytopenias, with death as a competing event, occurred in about 75% of patients at 18 months following discontinuation; as such, there's a significant morbidity and mortality aspect. If you look at potential risk factors, when we did variable analyses, we saw that age over 65 and an increase in the Charlston Comorbidity Index or worsening of performance status is associated with an increased hazard for death.

The study highlights the fact that across a large claims database, in a patient population that's validated and has adequate data, whether [patients] are above 65 or [younger] than 65, the outcomes are poor. There's this high morbidity in terms of cytopenias and the mortality rate is poor, as well. These data just highlight the fact that there's an unmet need here for additional therapies.

That's why studies with, for example, imetelstat, the Geron compound that was run in a phase 2 study, [are needed]. What was always interesting about that study was, if you look at the survival outcome, the median survival for the active dose, which was 9.4 mg/kg given every 3 weeks, was 30 months; that's much more favorable than what we're seeing [with ruxolitinib], which is about 1 year. Of course, that was a study that didn't have a control arm per say. The control arm [evaluated] the low dose, and the low-dose arm had [a median survival of] 19 months. [Those results] suggested that there is a survival benefit [with that approach]. However, that was an open-label phase 2 study, so the next study that needs to be done is a randomized phase 3 study of imetelstat compared with best available therapy, with a primary end point of OS. I believe it such a study would mark the first time such an important end point would have been the focus in myelofibrosis.

Beyond imetelstat, what other novel agents are being examined in this patient population?

The field is very competitive now in terms of finding alternatives for patients who come off ruxolitinib, as well as commercial alternatives. It has to be noted that fedratinib (Inrebic) is an FDA-approved therapy with a broad, open label that is indicated for patients with a diagnosis of myelofibrosis and a platelet count of greater than or equal to 50 x 10⁹L. If you look at the JAKARTA-2 study, which was a phase 2 study that evaluated fedratinib as a second-line option for patients in whom ruxolitinib failed, that agent proved to be effective with response rates with splenomegaly, for example, of 30%. You could clearly achieve responses and salvage responses as it relates to spleen and symptom burden with fedratinib with the understanding that it is also a FLT3 inhibitor, so there are some GI toxicities. One needs to be aware of those toxicities and proactively manage that aspect [of treatment]. Fedratinib provides another JAK inhibitor option to patients within the community.

Beyond that, there's a lot of interesting work that's informed by laboratory studies. Some of these efforts include examining the addition of a drug like CPI-0610 to ruxolitinib in patients who are having a suboptimal response, and then drugs like navitoclax, which is a BCL-2 inhibitor. Many [agents are] in the pipeline that are being developed in an attempt to salvage responses either as a single agent or in combination with the goal of [managing] spleen symptoms, [improving] quality of life, and hopefully, [increasing] OS.

What is your take-home message to your colleagues?

Enroll. My recommendation always is to enroll patients, because myelofibrosis is a rare disease. We know that 2 approved therapies are commercially available and those should be used [in these patients]. However, the reality is that there is still a lot of unmet need in this chronic, aggressive hematologic malignancy. As such, if you have patients who have low platelets, I would strongly consider enrollment [on a clinical trial], if possible. I do realize that it's not always possible and that many logistics are involved.

However, for patients with low platelets, the PACIFICA study, for example, is evaluating pacritinib as a novel JAK inhibitor. For patients with transfusion dependence, the MOMENTUM study is evaluating momelotinib. CPI-0610 will eventually roll into a phase 3 study [for examination in the] up front setting for treatment-naïve patients. Investigators will compare the agent in combination to single-agent ruxolitinib in JAK inhibitor—naïve patients. As I said, I would also like to see imetelstat move into a phase 3 study for survival; it would be an important study. Several studies capture many unmet needs, and if access is provided to patients to [enroll on] those studies, that's really important; that's how the agents will eventually get to the commercial space.

Mascarenhas J, Mehra M, He J, et al. Patient characteristics and outcomes after ruxolitinib discontinuation in patients with myelofibrosis [published ahead of print March 31, 2020]. J Med Econ. doi:10.1080/13696998.2020.1741381