Webinar: Gender and Patient Outcomes in HCL: An HCL Patient Data Registry Analysis
March 24, 2023
Hosted by the Hairy Cell Leukemia Foundation (HCLF) with Dr. Leslie Andritsos from The University of New Mexico and Dr. Naren Epperla from The Ohio State University. Moderated by Anna Lambertson, HCLF Executive Director.
Presentation Materials
View the presentation slides. >>
View the published study discussed during this webinar. >>
Transcript of the Slide Presentation
Anna Lambertson:
My name is Anna Lambertson. I'm the executive director of the Hairy Cell Leukemia Foundation, and it's my honor to interact with so many patients and families across the United States and around the world, and to put together programs like the one today about research. The majority of our funding comes from patients and families, and most of that we're putting towards really important research.
Today's webinar is part of our “What's New in Hairy Cell Leukemia” series. Whenever new data that's relevant to hairy cell leukemia is published and made available, we invite the lead investigators or the lead authors of the paper to present their data and share with us the insights that they've discovered.
Dr. Leslie Andritsos:
Welcome everyone. We're excited to share this research.
Just a quick background on hairy cell leukemia. It's a rare hematological malignancy that makes up less than 1% of all lymphoid neoplasms. There are fewer than 2,000 new cases in the United States each year with a median age of diagnosis of about 50 to 55. Of course, with some people older than that and some people younger.
We don't really understand why, but there's a huge discrepancy in the number of female patients diagnosed with this as opposed to male patients. There are many more male patients. It's about a 4:1 ratio. And until this paper was published, there had not been a study performed aimed at just characterizing what happens to female patients with hairy cell leukemia. So we studied female patients in the hairy cell leukemia patient data registry and looked at the disease characteristics, their responses to treatment, and outcomes when compared to their male counterparts.
The HCL patient data registry is an international multi-center patient data registry of people with hairy cell leukemia. We've put in all the relevant information about the disease characteristics, any treatment people have received, their response to treatment and outcomes, other health outcomes that might be of interest, for example, infections, and then molecular testing, which is becoming more and more important in the diagnosis and prognosis of hairy cell leukemia.
The Hairy Cell Leukemia Foundation sponsored the development of the hairy cell leukemia patient data registry. This was built at the request of hairy cell leukemia patients who wanted to know more about their disease and wanted more research done in this disease.
We only included adults in this study. The female patients were the study populations, and we compared them to the male patients in the registry. And our primary objective was to assess the time from their initial treatment to when they would need treatment again in female patients compared with male patients. And then the secondary objectives were to look at their response rates to that first line treatment and determine, if we could, what would predict the time to next treatment.
We had 357 patients, 265 males and 92 females. The age was basically identical. The race was basically identical. Most of the patients were never-smokers or had quit and were ex-smokers. Just a handful of current smokers. There was the same relative chemical radioactive exposures, which was fairly low but not zero.
And then the majority of the patients, as you would expect, had classical hairy cell leukemia. We only had two female patients with the variant and 12 male patients with the variant to compare, and then some atypical patients that may not have had a classical component of proteins on the surfaces of the cells. The majority of patients have the BRAF mutation as you would expect, and the vast minority had wild type which means no mutation. And then looking at the blood counts, the bone marrows, so the bone marrows were essentially identical. The percentage of hairy cell in the bone marrow was essentially identical and the blood counts were identical between the two groups, except for the hemoglobin was a little bit lower in the female patients and the platelet count was a little bit higher. Again, we don't really know why that would be.
The majority of all the patients got cladribine as their first therapy, and then a smaller number of patients but still basically not statistically different received pentostatin. Some patients, probably the more recent patients on the registry, had gotten cladribine plus rituximab, which is emerging as probably the first treatment of choice in people who don't have a reason to not have that treatment. And then just a handful of vemurafenib and other treatment types.
We have the group of patients that had a complete response, either unconfirmed or just a hematological response, meaning the blood counts recovered completely. And then same thing with the partial remission and an unconfirmed or partial hematological remission, and then stable disease which is basically people who didn't really respond but also didn't have disease progression during that treatment time.
And so again, you will see that the response to treatment was basically identical between the female and male patients. About 82 to 84% of all the patients had a complete response. Partial remissions were identical and stable disease was identical. That is actually very interesting because as we'll show in just a minute, the female patients had a longer remission than the male patients but they had identical upfront responses, which is what I think makes this study so interesting.
Dr. Naren Epperla:
It's important to note in statistical terms, this 0.50 is what we call as median which means majority of the patients. For the males, the median time to next treatment is around eight years, whereas for females, it is over 17 years. That's a very significant difference, because we are looking at almost nine years difference between males and females in regards to the time to next treatment.
(Slide 8)
And on this slide, the green indicates stable disease, red indicates partial remission, and then blue indicates complete remission. And you can clearly see here, patients who are in complete remission had the best outcome, which means they had the longest time to next treatment compared to those who were partial responders or who were in partial remission. And then lastly, those who have stable disease. And you can see here on the bottom, the numbers were pretty good.
Like Dr. Andritsos was alluding to, you had a total of 185 patients who were complete responders or in complete remission relative to 34 patients who were partial responders or in partial remission, and 5 in stable disease. It's really hard to discern too much out of this for the stable disease patients, but we can definitely learn a lot in regards to their response and outcomes for patients in complete remission and patients in partial remission.
(Slide 9)
Then what we decided to do is look at, since in females, the time to next treatment was very long, let's try to see if this holds good when we stratify or divide it based on the age. So this is a graph showing time to next treatment for patients with hairy cell leukemia who are 60 years or younger. You can see here there were 175 male patients in the bottom and 61 female patients. Again, the red indicates female patients and the blue indicates male patients. As you can see, female patients had longer time to next treatment.
(Slide 10)
And then when we looked at the same thing in patients who were over the age of 60, we did not see the difference. As you can see here, there was not much difference between the two groups. And more importantly, the median, which is if you draw the line at 0.50, was not reached in either of the groups. And again, the numbers are shown below with 57 male patients and 14 female patients.
(Slide 11)
Then we took a step further. Patients who were 60 years old or younger had better outcomes. So let's see if you further divide them into those who are 50 years or younger versus 51 to 60. This is the first of the two graphs trying to show the time to next treatment in patients with hairy cell leukemia who are 50 years old or younger. And you can see here, the female patients had significantly longer time to next treatment with the median not reached for female patients. And for male patients, it was around seven years. And the numbers are shown below with 99 male patients and 41 female patients.
One of the reasons why I have been alluding to or referring to the numbers is these are pretty good numbers. It’s always important to see what are the numbers involved. Because if you have really small numbers, it's really hard to interpret the data. But when you have good numbers, that means the data is interpretable and we can learn from it.
(Slide 12)
And the last of the two graphs is when we looked at the patients between 51 and 60 years old, you can see here, the median, we'll start off with the median, for the male patients was around six and a half years. And for the female patients, it was not reached.
Here's an important thing to understand. The P-Value is 0.05. In statistical terms, whenever the P-Value is less than 0.05, we call it as statistically significant. When it is at or over 0.05, we do not call it as statistically significant. Now in just looking at this, obviously the graphs and the curves look very disparate. Definitely the female patients with hairy cell leukemia have a pretty significantly longer time to next treatment compared to the male patients.
However, this value is also taken into account when you look at the numbers. If you look here, there were 17 males and 18 females. When you put everything together, even though this was trending towards a significant difference, it did not reach yet.
Dr. Leslie Andritsos:
When we looked at things that are predictive of time to next treatment, we looked at several different variables. And of course, the big one that stood out right away was that being female was statistically significant as a predictor of longer time to next treatment. When we looked at most of the other things, there was no predictive value whether people were smoking, what their BRAF status was. But it did seem to be significant what their median hemoglobin value was and what their median platelet value was. And of course we don't know why that would be different between the groups.
And then definitely, the thing that made the biggest difference in whether people stayed in remission a longer time was having achieved a complete remission with the first treatment. That was the biggest predictor of time to next treatment. And so looking at the P-Values on the multi-variable assessment, meaning looking at more than one value, taking all these things into consideration, the two things that had the biggest predictive value were being female or having a complete response with that first treatment.
I think that this study raises more questions than it answers. There's so much we don't understand about why there would be a difference. Of course, first and foremost, what is the role of male versus female hormones, testosterone versus estrogen? Is there any impact of hormone replacement? For example, were women taking hormone replacement and that somehow impacted how long they were in remission or not?
And then one of our most frequent questions, we get a lot of questions about what happens to women who are pregnant when they are diagnosed with hairy cell leukemia or become pregnant after their diagnosis or in need of treatment? And then finally, what are the long-term survival outcomes of both groups of patients? So does this translate into a better survival for female patients? I will say in most things, if you stay in remission longer, you're likely to have a longer survival. But we actually don't know whether that's true in this case. And so that's something that we'll be looking at as a follow-up to this initial study.
And then I'd like to acknowledge how this all happened. We want to first thank all of the hairy cell leukemia patients, male and female, who have participated in the patient data registry and made this research possible. And of course, the Hairy Cell Leukemia Foundation for supporting the development of the patient data registry and serving as a platform for this research. We had a lot of people working on this project, but some of the big ones at Ohio State were of course Mirela Anghelina, Puneet Mathur who is in the Division of Biomedical Informatics, and then Qiuhong Zhao who is a biostatistician who helps us with a lot of our projects. They're all at Ohio State.
And then we have some collaborators at other institutions. I would like to highlight Dr. Versha Banerji who is looking at some of these questions about the impact of hormones, female and male hormones, on either development or progression of lymphoma. And Dr. Clive Zent. He's just been a wonderful collaborator. We couldn't do this without all of these people, and especially our patients.
Transcript of the Question & Answer Session
Anna Lambertson:
I want to make sure that some of the key points are really clear for everybody listening. There's longer time to next treatment for women overall. When you looked at patients under the age of 60, that held true. But then when you looked at older patients, that dramatic difference really wasn't there. Is that correct?
Dr. Naren Epperla:
That brings the question about the impact of sex hormones. And that's essentially why we tried to subdivide in patients who are younger than 60 to younger than 50 and 51 to 60. Again, I'm not trying to make a definite statement here, but it just provides us with a thought that could there be an impact of hormones on what we are seeing? Because we definitely saw the difference very significant when they were younger than 60, even more younger than 50. As you see, younger than 50, it was very significant, 51 to 60, trending towards significance, and then over 60, not significant at all. That is where in my mind at least I'm thinking, would there be an impact of hormones playing a role in this?
Anna Lambertson:
You pointed out that hemoglobin levels and platelet levels did seem to be an indicator for patient response. And one of the questions that came to the Q&A was when were those blood counts taken, when were those numbers noted in that patient's journey?
Dr. Naren Epperla:
So your first question was, was it taken at baseline or after? Baseline characteristics is before the start of the treatment.
(Slide 13)
And then the second part of the question, if you go back to the univariate. When you look at something like median hemoglobin a P-value of 0.008, and if you look at 0.86, what it implies is if you have a higher hemoglobin, you have a higher chance of staying in remission and not needing the next treatment. That is what you learn from that point.
Similarly, here with the platelets. So since it is less than 1, it essentially means the higher the platelets, there is a higher chance that you will stay in remission and not need the next treatment. But again, as Dr. Andritsos mentioned, univariants are very helpful but they only give us so much information.
What is more important is multi-variable where you're controlling for other factors and trying to understand if they still pan out when you control for other factors. When you're controlling for age, when you're controlling for smoking status, when you're controlling for BRAF, are they still significant analysis? Only two variables panned out, one was female sex and the other one was remission to or response to the first line.
Anna Lambertson:
There’s another question in the Q&A about the size of data. Could you quickly speak to the statistical significance of the data that you had and why you were able to carry out a meaningful study using the numbers of patients that you had available?
Dr. Naren Epperla:
The first thing in my mind is when I'm trying to analyze data, we need to have a good number of both the primary variable, which is what we are trying to look at which is females, female patients, and the variable that we are comparing to, which is male patients. Let's say if you had only 10 or 20 female patients and had 200 male patients, we would not be learning anything about it. Because when you have that few numbers, you’re diluting it and the numbers may look very skewed towards one way or the other. But when you have the numbers which are closely approaching 100, let's just say in each group, then that gives us enough power in statistical terms to be able to carry out such an analysis that we did.
And the other thing is because of the numbers and again, thanks for all the patients who are willing to participate in this registry, because of the numbers, we were able to do all the subgroup analysis. I can guarantee you if we had only 50 female patients, which is still a very good number in statistical terms, however, we would not be able to do those subgroup analysis like less than 60, 51 to 60, or less than 50. There's absolutely no way because then you are further decreasing the sample size.
So the power of a huge sample size is not just looking at the main variable which is what we looked at, all of ages included, but also try to learn more about it. Because what we learned from this dataset is perhaps there is an impact of sex hormones playing a role and which will further be substantiated in additional studies to come in the years to follow. But this gives us the first snapshot of maybe that is what is playing a role.
Dr. Leslie Andritsos:
And I would just add to that. This is I think the first study to really show the power of this patient data registry because for rare disease research, these are huge numbers. And that's because of the participation of so many patients with hairy cell leukemia, which is really extraordinary. And this is also a snapshot of there's lots of other types of research that we can do with this registry. And so there will be more to come. And we want to also be sure that we're asking questions that are important to the patients with hairy cell leukemia. I know there's a lot of unanswered questions out there, and we want feedback on whether we're going down pathways that are also important to patients.
Anna Lambertson:
There was a study also carried out in part by you, Dr. Andritsos, that looked at quality of life. And again, it drew from the patient data registry. That topic was mentioned here in the Q&A that yes, time to next treatment is extremely important as are elements of quality of life. And so I'm sure down the road it'll be really interesting to try to look at these data sets, perhaps some in parallel, and see if other conclusions can be drawn as well.
Dr. Leslie Andritsos:
Right. We could potentially merge those. We found in the quality of life study, was that the people with the best quality of life were those in remission, not in need of treatment. However, we cannot assume that because there might be a lot of other things lowering people's quality of life. And so I think that would be a really intriguing next step, is to compare the quality of life outcomes as well.
Anna Lambertson:
Great. There's a couple of people wanting to look again at the treatments that were used. Did you have a slide where you looked at the treatments used for these patients?
Dr. Leslie Andritsos:
Yeah. We looked at the most common treatments and they're listed here, so cladribine, pentostatin, cladribine with Rituxan, vemurafenib or others. And keeping in mind that since this was the patient's first treatment, these had to either be approved as first line therapies, which these all are, or somehow they were able to obtain the treatment, because vemurafenib is not approved as a first line treatment. And so then we looked at that. We had 224 patients that we had the treatment information on. And if you look just across the lines here between males and females, we can see what percentage of each total group got which treatment. And they're pretty much identical. Most of the patients got cladribine. It was about the same between males and females. Same thing with pentostatin. The looks like maybe the female patients got a little bit more cladribine plus rituximab. But again, the numbers are not really high enough to determine that. And there's no statistical significance to any of these differences because they're very small differences. And of course these numbers are so tiny, we can't really say anything about any differences between those.
But by far, the number one treatment was cladribine, and about the same numbers of male and female patients got each type of treatment. I think that's important to talk about because you know might start asking, "Well, maybe women just got better treatment. Maybe they somehow received a superior treatment." But it doesn't look like there was any difference in the treatment chosen between male and female patients.
Anna Lambertson:
Another question that came through the Q&A is about the term ‘probability to time to next treatment’ or ‘probability to next treatment’. Could you quickly clarify what you mean by probability.
Dr. Naren Epperla:
This probability of not starting, that is essentially a statistical term of saying. I would strictly just look at time to next treatment in a very simplistic way. When I first started the treatment to when I received the next treatment, that is essentially it.
And then the question that may have come up is why did we choose time to next treatment over a progression-free survival, let's say, or failure-free survival, some other endpoints? One of the reasons being this is an absolute that has been captured in the data registry because it's a date. You start on one date, that is when you started your first treatment. You start on a second some other date if you get started is the second treatment. These are two concrete time points. But let's say if you start talking about progression-free survival, the event is defined as progression, death. So there are a lot of other things that if you don't have enough events or if you're not capturing properly, then the curves will get skewed.
And that's essentially why in this analysis, we chose time to next treatment over other statistical endpoints such as PFS, OS.
Anna Lambertson:
We know in the registry, we are capturing people's locations so we know what state or country they live in. You didn't include that in the study, but is that something that you think could be looked at and that also could be an important data point to include in further evaluation?
Dr. Naren Epperla:
I think that's definitely a very important point and thank you for bringing it up. So hopefully in the subsequent studies, we will make sure to include that as an additional variable so we can learn more about it.
Anna Lambertson:
Another individual is asking, "Okay, this is when I was diagnosed, but I believe that I probably had HCL for much longer." And I'm just curious if that is something you've thought about or if that's even something that we could know or document?
Dr. Leslie Andritsos:
So we made a decision at the very beginning that we were only going to include source data in our what we call discreet data elements, meaning things that we can look up and see the date of. And by source data, that means the day of the bone marrow biopsy or the date of a blood count or the date a CAT scan was performed that showed a huge spleen or something like that. And that's again, because as Dr. Epperla was alluding to, it becomes very difficult to discern what we're doing if we don't have time points.
And so many people will say, "Well, my neutrophil count was low for years and nobody knew why." And so in cases like that, we would be so happy to include that data if we can get the blood count reports. And I think it actually is really important to do that because there's probably tons of people who had hairy cell way before they knew it and were doing fine. And so right now, we would say that about 10% of people with hairy cell leukemia don't need treatment when they're first diagnosed. But usually people get diagnosed because there's a problem, because they had a big spleen or something was happening and they had a symptom. And so I suspect that there actually are a lot of asymptomatic people walking around doing fine. And that would be also very important to know, time to first abnormal lab to next to first treatment. That would be extremely interesting to look at.
Anna Lambertson:
Thank you. This is a snapshot of your initial findings in what I think is an extremely important and extremely interesting area of study. And we just want to caution those who have joined the webinar today or who may read your paper or look at these slides later. Each patient is unique and it'll be really important that you are in communication with your doctor about your own treatment, about your own health situation and when treatment should begin.
While this raises really interesting questions about the effect of hormones, I think you don't know for sure what is really causing these big differences yet. So I just want to caution those listening. We had somebody in the Q&A ask, "Well, if time to next treatment is longer under the age of 60, should I be having hormone replacement so that I can extend my remission?" And we don't want people to draw too many of these types of conclusions.
Dr. Leslie Andritsos:
Yes and we will try to find an answer. So please we know that there are side effects and potential toxicities from hormone replacement. So please don't start anything new that you don't need on the basis of this study.
Anna Lambertson:
I want to thank both of you for, one, your commitment to hairy cell leukemia research, for your commitment to being involved with the Hairy Cell Leukemia Foundation and being part of our Hairy Cell Leukemia Center of Excellence Network. Each of your institutions is really important to us and to ongoing research to advance understanding. And I want to thank everybody who, as Dr. Andritsos said in the very beginning, everyone who's a part of the patient data registry. There's no way that we can learn new things about this disease and patient outcomes without data. And the patient data registry is a critical tool in that.
Dr. Naren Epperla:
Again, thank you all for being able to attend. And second, what I want to really also emphasize, we need to be careful not to over-interpret the data like we have been discussing throughout. This is truly a snapshot and the beginning, and we will have a lot more to dive into and then understand why these differences exist and what is driving it.
Until then, I would definitely continue to follow with your physicians, whoever you are following with, and do the rightful thing rather than try to change based on what we discussed today. And last but not the least, big shout out to all the women. This is Women's Month, so it worked out very well. Even though being a male, I'm extremely proud to be a part of this project, and I look forward to many more important studies to come in the hairy cell.
Anna Lambertson:
Take care, everybody.
This transcript has been edited for clarity.