Expert Expands on Implications of Molecular Targets in Myelofibrosis

Raajit K. Rampal, MD, PhD

ASXL1, SRAF2, IDH1/2, EXH2, and TP53 were not found to have an impact on overall survival (OS) or relapse-free survival (PFS) in patients with myelofibrosis (MF), according to a retrospective analysis of pre-transplant samples. However, U2AF1 or DNMT3A mutations did correlate with worse outcomes in patients with MF following stem cell transplant (SCT).

“Collectively, I think all of these findings are telling us that mutations definitely have an impact in how patients do after they get a transplant, but what is clear from all of these data is that we need larger data sets to come up with more definitive conclusions,” said Raajit K. Rampal, MD, PhD.

In the analysis, investigators evaluated pre-transplant samples to evaluate the association between mutations and patient outcomes in an effort to retrieve more information on the prognostic value of mutational profiling in patients with MF who undergo SCT. Based on prior studies, mutations such as ASXL1,SRAF2, IDH1/2, EXH2, and TP53 were associated with increased risk of progression in patients with MF.

Data were collected from 101 patients with MF who underwent transplant. Investigators sequenced 585 genes on pre-transplant samples. The 5-year OS and RFS rates were 52% and 51.1%, respectively.

In an interview with OncLive, Rampal, a hematologic oncologist at Memorial Sloan Kettering Cancer Center, discussed the retrospective analysis of pre-transplant samples from patients with MF to determine the value of mutational profiling in predicting outcomes after transplant.

OncLive: What was the rationale for conducting this study?

Rampal: We know that in myeloproliferative neoplasms, we have been able to utilize genomics for prognostication increasingly over the last few years. There have been a number of studies that have pointed at mutations having an impact on prognosis, both in terms of patient’s risk of having, for example, thrombotic events, but also in terms of MF patients having a risk of progression. This was codified very recently over the development of several genomic-based prognostic scoring systems for patients with MF.

Historically, we have had to use clinical variables only, like a patient’s age or whether or not they are anemic or if they have a high white [blood cell] count, in order to give them some sort of prognostic score, but now there have been a couple of scoring systems that have successfully integrated genomics. In particular, we know about some of the mutations that have a negative prognostic impact on these clinical variables [and have] come up with a comprehensive scoring system.

The question to us that arose based on these historic observations was really that we haven’t had a way to predict how patients do in terms of their transplant outcomes that have employed genomics. The goal of our study was to essentially retrospectively examine a cohort of MF patients who had undergone transplant for whom we have had sequencing data available to ask the question of are there mutations that do segregate out in terms of patient outcomes, whether it be positive or negative, after they have had a transplant. That was our overall rationale for this study.

How was this study designed?

The study utilized patient samples and information from the MPN Research Consortium (MPNRC). These were samples that were collected and tissue information that was collected as part of the clinical trial that was run by the MPNRC, which was the MPN-101 study, that was really a study looking at the impact of donors and conditioning regimens. We had samples from those patients, but we also went in and collected further samples from member institutions of the MPNRC including Memorial Sloan Kettering Cancer Center, the Princess Margaret Hospital in Toronto, and the Moffitt Cancer Center in Florida, all of whom are members of MPNRC.

Essentially, what we did is we looked at this group of patients, which was over 100 patients, and looked at their outcomes in terms of relapse post-transplant and death after transplant. We also performed next-generation sequencing on all of these cases from samples that were obtained prior to the transplant. This was a mutation profile of several hundred genes. We performed an analysis to look and see if certain mutations correlated with outcomes. In particular, we started by looking at mutations that are known to be high risk in MF patients for the risk of progression, and these include mutations IDH1/2, EZH1/2, ASXL1, and SRSF2. Those have been known to carry a poor prognosis in patients with MF. We examined the impact of these mutations individually and also collectively.

What were you able to find in this analysis?

What we were able to find was that these mutations didn’t appear to have a negative prognostic impact on patients who were undergoing transplant, which is an interesting finding in the sense that we know that the patients with MF who have these mutations tend to have a worse prognosis. To us, that at least suggested the idea that these mutations might be overcome, or rather, the negative prognostic impact mutations, might be overcome by transplant.

Now, since there were a number of mutations that we did identify, we also wanted to have looked at mutations that occurred at a rate greater than a 5% frequency to determine where there are some other mutations that might actually have prognostic value when we think about the outcomes of patients.As it turns out, we identified that there was another splicing factor called U2AF1, which in fact the presence of which portended a worse transplant outcome in patients who had it. This, to us, was an important finding because it did identify another group of patients who have these mutations who might not have as good an outcome if they undergo a SCT.

Why is identifying these biomarkers in patients with MF important?

There’s a couple of things conceptually that make studies that look at biomarkers important. One is that it can help us with clinical decision making. That being said, it’s only a part of the conversation one has with a patient, but it’s important that when we talk to a patient about something like transplant or even going on a new therapy, we talk to them as best we can about the risks and benefits of undergoing that therapy. For something like transplant, we know that there are certainly a substantial number of risks of undergoing a transplant. If we feel confident that we are likely to have a better outcome by going to transplant versus not going to transplant, then we would favor that treatment approach for a patient. However, if we had data that said even if we do the transplant, the outcome is not likely to be as good or it might be equal to that if we didn’t do the transplant, then that would change our decision making. It’s important for us to have biomarkers to improve our ability to prognosticate what is the likely outcome for a patient. It helps us, and it helps the patient in our collective decision making about what the best treatment option is for that patient. I think that’s the applicability of these types of biomarkers in the clinical setting now and today.

Beyond that, there is also the potential opportunity to devise new treatments. We are learning more and more about these mutations. We know, for example, that the IDH1/2mutations are now targetable with FDA-approved drugs and that other mutations like the splicing factor mutations such as SRSF2are being targeted preclinically and clinically; there is an ongoing trial currently that is using a novel drug to try and target splicing factors.What that might allow us to do, hopefully, is to say that we know certain patients are at high risk because they have, for example, an IDH2mutation, and we know that we have a drug that is approved for IDH2-mutated patients. Could we do a clinical study to use targeted therapies to try to improve their overall outcome? That, to me, is sort of a near-term idea that we can use biomarkers to help us really define high-risk patients and to also identify potentially novel treatments that we might offer those patients to improve their outcomes. To me, those are the 2 biggest aspects that are important when doing biomarker studies.

What is the take home message from this research?

I think there are a couple of important points to be made. There are a couple of recent publications that have looked at the impact of mutations in MF transplant outcomes, and there is disagreement among them, which is not uncommon when we do this type of science. We know that there is an important paper that was put out by the German group which looked at developing a scoring system for MF transplant patients. They concluded that things like the ASXL1mutation are actually prognostic for a poor outcome for transplant, whereas data that has been published by the group at the City of Hope is somewhat more similar to our data where they identified that the U2AF1mutation appears to have negative prognostic significance.

Collectively, I think all of these findings are telling us that mutations may certainly have an impact in how patients do after they get a transplant, but what is clear from all of these data is that we need larger data sets to come up with more definitive conclusions. I think that each study has 100 or up to 300 patients, but what we really need to do is combine a larger number of patients with sequencing data and clinical data to develop a more robust model. This, I think, is the start of trying to understand how these things impact transplant, but now we are going to move beyond the first step to build bigger and more robust cohorts to really understand this.