Jennifer R. Brown, MD, PhD, President of the Society of Hematologic Oncology, Director of the Chronic Lymphocytic Leukemia (CLL) Center at the Dana-Farber Cancer Institute, and Worthington and Margaret Collette Professor of Medicine in the Field of Hematologic Oncology at Harvard Medical School, and Anthony Mato, MD, MSCE, Director of the CLL Program at the Memorial Sloan Kettering Cancer Center, debate whether combination small molecules are better than sequencing small molecules in CLL treatment.
Combination Small Molecules Are Better Than Sequencing
By Dr. Brown
Why combinations? In general, we know that single-agent therapy selects for resistance. From a purely mathematical standpoint, if you have one drug, it’s much easier to have some cells present that are resistant to that single drug, while if you have two drugs that are non-cross resistant with different mechanisms of resistance, it becomes orders of magnitude harder to have cells that become resistant to both drugs. I would argue the fact that responses with single-agent therapy are shorter in patients with a TP53 aberration or complex karyotype also supports this. Furthermore, we know that combinations allow patients to achieve deep remission, often with undetectable minimal residual disease (MRD), which is a first step toward cure and is also associated with improved overall survival (OS) in some cases. It also allows discontinuation of therapy, which mitigates several problems, not only the selection pressure I just mentioned for resistance mutations, but also cumulative adverse events—which we know can be quite severe, particularly with ibrutinib—loss of adherence, and cost. I would also note that most therapeutic advances in cancer have often relied on combinations of highly effective therapies.
Just to review some of the single-agent data and some of the issues with it: We have the Ohio State University long-term experience with ibrutinib where we have early Richter transformation that plateaus, but we see continuous CLL progression over time on therapy, rising to quite a high level, with risk factors including complex karyotype, (del)17p, and younger age. If we look also at the ALLIANCE trial, we see that increased karyotypic abnormality is associated with worse outcomes in the ibrutinib groups as well as patients with unmutated immunoglobulin heavy chain variant (IGHV). Furthermore, Dr. Mato has published that 17p deletion leads to worse OS with ibrutinib in a real-world setting. Dr. Inhye Ahn has also published a model in which 17p deletion is implicated in significantly worse outcomes with single-agent ibrutinib.
What about venetoclax? In the long-term follow-up data from the phase II study of continuous venetoclax in patients with (del)17p, you can see a very similar pattern, early Richter transformation and then continuous progression. Furthermore, with the CLL14 trial, we again see the higher-risk patients are relapsing earlier, reflecting this steady development of resistance in the higher-risk groups first, but that we’ll see later in all patients. The median progression-free survival (PFS) is approximately 48 months for the TP53-aberrant patients, and a little over five years for the patients with unmutated IGHV. In data from the Mayo Clinic looking retrospectively at patients who were previously exposed to ibrutinib and venetoclax, then given combination therapy, you can see these patients have a very poor outcome with a combination of ibrutinib plus venetoclax. They had a 14-month median time to next therapy and only a two-year OS. That’s the outcome of our sequential single-agent approach over time.
How can we do better? Combinations lead to deep remissions, which is, again, a first step toward a cure. But even if it’s not a cure, depth of remission is associated with later and later relapse. In a large study from the United Kingdom looking at all-comer CLL patients, regardless of therapy or line of therapy, undetectable MRD was associated with better PFS and OS. In fact, outcomes were better when undetectable MRD was achieved after first-line therapy. Of course, we saw this with the chemoimmunotherapy data, that undetectable MRD was associated with prolonged PFS and a very long-term favorable outcome in patients with mutated IGHV. We now have data with venetoclax, similarly from CLL14, where depth of remission is associated with PFS, and in fact, undetectable MRD is associated with improved OS now in this study.
You may say, how can this be true of Bruton’s tyrosine kinase (BTK) inhibitors? We know that if you stay on the drug, it doesn’t matter if you have persistent disease. There are some emerging datasets suggesting that depth of remission may matter also, even with BTK inhibitors. Data from the MD Anderson Cancer Center show that achieving complete remission with ibrutinib was associated with better PFS compared with partial remission. In a five-year follow-up of the phase II MD Anderson study of ibrutinib plus or minus rituximab in high-risk patients, depth of remission was associated with a much better PFS compared with those with only partial remission.
We also have follow-up data from the HELIOS study, which combined ibrutinib with bendamustine plus rituximab. Ibrutinib was given continuously. Undetectable MRD correlates with outcomes in the bendamustine plus rituximab arm. But in the ibrutinib plus bendamustine and rituximab arm, even though patients are on continuous ibrutinib, we’re still seeing decreases in PFS in patients with MRD at 1% or higher. These data suggest that perhaps depth of remission can improve PFS, potentially eventually OS, even with BTK inhibitors.
How do we achieve that? We know that combination with venetoclax or a B-cell lymphoma (BCL)-2 inhibitor is highly effective in vitro, certainly in combination with a BTK inhibitor, and clinically. In the phase II CLARITY study, patients had a median of one prior chemoimmunotherapy regimen and a median age in the mid-60s. They were given ibrutinib plus venetoclax for two years initially, then potentially extending to three years. There was a steady improvement in undetectable MRD up to the two-year time point in both blood and marrow. The rapidity with which disease was depleted was strongly associated with undetectable MRD and outcome. What’s particularly remarkable about this study is the three-year PFS in the relapsed setting of 96%, which is really a very good number. As I mentioned, these are patients with one prior chemoimmunotherapy regimen and they’re somewhat similar to the patients treated in the ASCEND trial, also with a median patient age in the mid-60s and one prior regimen. If you look at the three-year PFS in the acalabrutinib arm from the ASCEND trial, it’s 63%.
This is an unholy cross-trial comparison, but at the moment we don’t yet have any direct randomized trials, and this single-agent PFS is substantially lower than what we saw for the combination in CLARITY. The HOVON trial, also in relapsed patients, combined ibrutinib with venetoclax. MRD-positive patients continued on ibrutinib. Those who had undetectable MRD were randomized between ibrutinib or observation. At one year, 37% of patients had undetectable MRD. The study met its primary endpoint, which had a similar PFS in the observation and ibrutinib arms in patients with undetectable MRD.
What about frontline? MD Anderson data with ibrutinib plus venetoclax in high-risk patients showed a 93% three-year PFS rate, with no difference based on IGHV or (del)17p status. In the CAPTIVATE trial, the MRD arm first received three cycles of ibrutinib lead-in, then 12 cycles of ibrutinib plus venetoclax. Patients with undetectable confirmed MRD were randomized between ibrutinib and placebo. Those in whom MRD was not confirmed were randomized between ibrutinib or ibrutinib plus venetoclax. In patients with confirmed undetectable MRD, there was a two-year disease-free survival rate of 95% with placebo and 100% with ibrutinib. That was an excellent outcome. In the patients who did not have confirmed undetectable MRD and who were randomized to continue ibrutinib plus venetoclax, after the first year of therapy, there was a further improvement in the rate of undetectable MRD in bone marrow of over 30%.
What about the fixed-duration arm of CAPTIVATE? The median patient time on study is 38 months. These are patients with high-risk markers, although they are young and fit. They received three months of ibrutinib lead-in and 12 cycles of combination ibrutinib plus venetoclax. Undetectable MRD rates are about 60%. The estimated three-year PFS in this study is 88% and similar for (del)17p, TP53 aberrant, or unmutated IGHV patients. This is an extremely high PFS.
What about GLOW in older patients? These data are interesting, because they suggest that PFS is still well maintained, even for the patients who are MRD-positive. It has been suggested that something about the combination therapy may be altering the biology of the cells in terms of the rate of regrowth or recurrence, similar to what has been described in the CLL14 trial in the venetoclax arm.
What about combinations? I’m also a fan of not forgetting the CD20 antibodies. The Ohio State University study combining ibrutinib, venetoclax, and obinutuzumab was the only one of the triple-drug trials that has PFS data of at least three years: 95% in the treatment-naïve cohort and 95% in the relapsed/refractory cohort. This is very impressive PFS data. It was a 14-month fixed-duration regimen that patients discontinued after that period. In terms of the next-generation BTK inhibitors, we have looked at acalabrutinib, venetoclax, and obinutuzumab where we find very high rates (77%) of undetectable MRD in bone marrow. It is still very immature for PFS. With zanubrutinib plus obinutuzumab and venetoclax, almost 90% of patients achieved undetectable MRD in blood and marrow and were able to stop therapy after a median of 10 months.
I would argue that combination small molecules are the future. They achieve deep remissions leading to previously unprecedented two- to three-year PFS, even in the relapsed setting. Outcomes are similar in patients with high-risk markers. There’s a suggestion that PFS may be better sustained off therapy even in patients without undetectable MRD in the GLOW trial. The combinations certainly result in less toxicity and cost.
What about resistance and retreatment? Data here are limited. In the MURANO trial—which is assessing single-agent venetoclax and subsequent retreatment—patients at the end of their first round of venetoclax plus rituximab had undetectable MRD. Then on retreatment, less than half of these patients achieved undetectable MRD. This was associated with an increase in their clonal fraction of 17p deletion and increased genomic complexity, which argues that we are still seeing clonal evolution after even this time-limited venetoclax regimen.
In terms of retreatment with combinations, data are extremely limited. We have limited data from the HOVON trial where they were able to reinduce undetectable MRD in seven patients with ibrutinib plus venetoclax. We have the CAPTIVATE retreatment data, in which patients have been retreated with single-agent ibrutinib with response. In terms of acquisition of mutations in these patients relapsing, in CLL14, we do see a little bit of clonal evolution with TP53 and BIRC3, high-risk mutations, in one patient each on the venetoclax plus obinutuzumab arm. This was better than in patients who received obinutuzumab plus chlorambucil, though. Thus far in CAPTIVATE, after ibrutinib plus venetoclax, we are not seeing clonal evolution mutations in the 13 patients.
Combination small molecules are highly effective in achieving deep remission with undetectable MRD, which is associated with improved OS in multiple studies, including a venetoclax-based study. Tolerability is much improved with second-generation BTK inhibitors. We can discontinue therapy for extended periods, avoiding toxicity, cost, and development of resistance. The early data show few resistance mutations at relapse, suggesting the possibility of retreatment.
I do have a provocative and somewhat speculative question: What is wrong with sequential single agents? We obviously need randomized data comparing combinations with single agents, and we will get these data with the German CLL study group’s CLL17 trial. In the relapsed setting, in a cross-trial comparison, initial PFS with a single agent appeared shorter compared with combination therapy. In the data we have, we know PFS to subsequent lines of therapy is shorter. If you use venetoclax after a patient has progressed on a BTK inhibitor, the median PFS is two years. If you use venetoclax like in the MURANO trial, it’s five years. If you look at patients who have received both a BTK inhibitor and venetoclax—treated now with pirtobrutinib, for example—the median PFS is 18 months in patients who have received both previously, versus not reached, with 74% of patients still on drug, if they’re only pretreated with a BTK inhibitor. If we look at ibrutinib plus venetoclax, we saw the frontline three-year PFS is 80% to 88% in the treatment-naïve group, but if we look at patients receiving the combination after both sequential ibrutinib or venetoclax, the median for this combination was only 14 months.
Data for retreatment after combinations are extraordinarily immature, but both the rationale and the early data suggest that we may be able to do better than these outcomes thus far.
Sequencing Small Molecules Is the Way to Go
By Dr. Mato
I don’t think anyone woke up one day and said, “Let’s do a triplet.” Those were developed. For example, the fludarabine, cyclophosphamide, and rituximab regimen was done in the setting of understanding the activity of monotherapy as compared with doublets and triplets. Then maybe quadruplets were tested, and they said, “Nah, a little bit too toxic. Let’s go back to triplets.” Oncologists who are advocating for combinations in the modern era have completely forgotten the lessons of the past. I’ll ask the question, “Is the whole greater than the sum of the parts?” as my argument for why we should stick with monotherapy until somebody does the appropriate study to convince me we should be doing a combination.
I would argue combining agents in the modern era has not been well studied in CLL despite the many studies Dr. Brown offered. There are a few prospective studies comparing novel agents with relevant controls. Contribution of effect is rarely assessed, and I think we should demand that. Follow-up is lacking when subjects are censored on clinical trials. There is absolutely very little data on retreatment. When you talk about sequencing of novel agents, they’re mostly from retrospective trials, and in the setting of combinations, which aren’t even approved in the United States, we have no real-world data on this particular topic. I would say we have to demand certain things to move forward in the era of combinations.
Certainly, we need data for time-limited approaches, and that exists. But what else is needed? We need decisions based on MRD status and depth of remission. We need combinations that don’t include venetoclax as the backbone. Certainly, we’re going to run out of combinations very quickly if it’s venetoclax, venetoclax, venetoclax, plus something else. We need sequencing data upon progression. Dr. Brown already admitted that was a major weakness from the data noted. We really need to understand the mechanisms of resistance for novel-novel combinations, not, “This is a BTK inhibitor, this is venetoclax, and so combined, we have no idea.” We need studies to assess the contribution of effect. It’s not okay to say, “I want to prescribe ibrutinib plus venetoclax or ibrutinib plus venetoclax plus obinutuzumab,” without really understanding if a doublet is better than a triplet or if a doublet is better than a monotherapy.
We need clinically relevant monotherapy controls, and then we need to understand comparisons of combinations to decide what’s most appropriate. I would argue, in terms of standard of care or for controls for clinical trials, that the gold standards have to be monotherapy with a BTK inhibitor or venetoclax with obinutuzumab. I’ll consider that a monotherapy as well since no one ever assessed for contribution of effect for obinutuzumab. The major phase III trials that support the use of continuous BTK inhibitors in the frontline setting—SEQUOIA, ELEVATE-TN, ECOG 1912, FLAIR, ALLIANCE, ILLUMINATE, and RESONATE-3—are the controls to beat, if we want to make the argument that we should use a combination. I won’t go through these trials in detail, but for ibrutinib, there are five positive clinical trials, some of which have an OS benefit. In the impressive ELEVATE-TN trial of acalabrutinib, not only do we have a positive result, but we also assessed for contribution of effect.
Then, zanubrutinib is compared with bendamustine plus rituximab. I’ll start by highlighting the longest-term follow-up data we have for a BTK inhibitor in the frontline setting, which is from ibrutinib in the RESONATE-2 trial. The comparison is against chlorambucil. The eight-year, and even now nine-year, follow-up data are quite impressive for ibrutinib monotherapy. More than 60% of patients are still progression-free when we have the longest-term follow-up for this molecule. The sustained PFS benefit with ibrutinib versus chlorambucil was also similar in mutated versus unmutated IGHV patients.
Why aren’t novel agents the controls in future novel agent combination studies? We continue to rely on controls that provide very limited knowledge. While we have a PFS benefit for BTK inhibitors and venetoclax, we won’t be able to highlight a combination trial that demonstrates a benefit against any clinically relevant monotherapy or combination with venetoclax control.
One could say, “Maybe ibrutinib is too toxic.” I’m not saying that, but I am saying that there are certain adverse events that are associated with continuous therapies. Atrial fibrillation rates range between 6% and 17%, and hypertension rates range between 14% and 34%. Even if you want to say that ibrutinib monotherapy may not be the go-forward or the standard of care now—which I still argue—there’s a point to be made for using that molecule.
We can think about next-generation drugs. The acalabrutinib long-term follow-up data in the frontline setting show in 99 patients quite excellent PFS. At 48 months, it’s 95.7% for patients overall; it’s 82% for patients with a TP53 aberration and 91% for patients with a complex karyotype. In the ELEVATE-TN trial, we finally get an assessment of contribution of effect. We have a comparison of acalabrutinib plus obinutuzumab versus acalabrutinib, and we have a significant advantage in PFS for acalabrutinib plus obinutuzumab versus acalabrutinib monotherapy. There is maybe an OS advantage, and I would argue that is a combination that has been tested appropriately.
The combinations noted by Dr. Brown have not had an appropriate assessment for contribution of effect. We have maybe a snippet of data suggesting lower adverse event rates associated with the next-generation BTK inhibitor acalabrutinib. I think the case is made for BTK inhibitors as a monotherapy. That is an excellent choice for treatment-naïve CLL. The next-generation BTK inhibitors seem potentially equally active and better tolerated. We now have evolving data for acalabrutinib plus obinutuzumab as a standard of care. That is a combination that has been appropriately developed. Dr. Brown also mentioned poor-risk disease, and we always like to say, “With (del)17p, we need to have combinations in order to overcome the poor-risk features associated with CLL.” But how do the BTK inhibitors perform in that situation?
I’ll highlight the longest-term follow-up data we have for ibrutinib in (del)17p disease, which are excellent data. The PFS and OS estimates at four years were 79% and 88%. Dr. Matt Davids did a summary of the acalabrutinib experience, and there was a quite high overall response rate, with a 48-month PFS of 89%. In the largest prospective study of a BTK inhibitor in the frontline setting in (del)17p for zanubrutinib, at 24 months, the PFS is excellent at 88.9%. I don’t think the ibrutinib plus venetoclax data or any of the triplet data look any worse or any better than this. The data Dr. Brown noted suggested that maybe there’s still room to go, and yes, I agree with that. But have the combinations demonstrated a benefit?
The gold standards for future comparison and standard of care should be BTK inhibitor-based therapy and venetoclax plus obinutuzumab. In the CLL14 study, the five-year PFS for venetoclax plus obinutuzumab is 62.6%, and even for (del)17p, it’s 49%. So, if the gold standards for future comparison—and standard of care—are BTK inhibitors and venetoclax plus obinutuzumab, are there adequate data to support current doublet combinations in clinical practice today? I argue no.
There’s an argument people make that you need a combination to overcome the poor-risk features of CLL. I don’t think that is held up, even with this snippet of data we have from CAPTIVATE, which shows patients with (del)17p are falling off the curve compared with all the other patients. In the data for GLOW, if we’re making the argument that ibrutinib plus venetoclax is a standard of care moving forward, then we need to compare it with either ibrutinib or venetoclax or venetoclax plus obinutuzumab. Certainly, ibrutinib has already beat the combination of obinutuzumab plus chlorambucil, as has venetoclax-based therapy. So why wouldn’t the combination of ibrutinib plus venetoclax beat a control that is essentially a straw man?
There is certainly a price to be paid for putting targeted therapies together in terms of what the adverse event profile looks like. Data from the Ohio State University that look at the combination of ibrutinib, venetoclax, and obinutuzumab are quite impressive in terms of PFS and OS, but I can’t tell you ibrutinib, venetoclax, and obinutuzumab is better than ibrutinib plus venetoclax, and I can’t tell you ibrutinib plus venetoclax is better than ibrutinib. However, we’ve already decided in the Eastern Cooperative Oncology Group and ALLIANCE studies that the triplet is already the experimental arm to move forward in the future, even though I haven’t even made the case for the doublet yet.
Finally, I’ll highlight important lessons of CLL13. First, there’s an important assessment of contribution of effect rituximab versus obinutuzumab. There’s an important assessment of doublet versus triplet, and there are relevant comparisons and endpoints. I’ll argue that the MRD data are very similar between venetoclax plus obinutuzumab and the triplet, which is venetoclax plus obinutuzumab and ibrutinib. There are no data for ibrutinib plus venetoclax because it wasn’t included. I’ll also argue that when you think about the PFS for this trial, they’re very similar between venetoclax plus obinutuzumab and the triplet. We haven’t made a strong case from the perspective of MRD with the limited follow-up. We haven’t made a case from the perspective of PFS, but I can make the case that toxicities are higher with a triplet compared with a doublet.
The rates of febrile neutropenia are 7.8% for venetoclax, ibrutinib, obinutuzumab versus 3.1% for venetoclax and obinutuzumab. So, what do we need to move forward? We have the data for time-limited approaches. We need data on MRD-driven decision-making. We need combinations that don’t include venetoclax. We need sequencing data upon progression. We need to understand mechanisms of resistance and contribution of effect. We need to have clinically relevant monotherapy controls to make the case and comparison of combinations.
What about relapsed/refractory disease? I made the case for what I would do in the frontline if I start with ibrutinib monotherapy or acalabrutinib, but what do I do next? I think there is adequate therapy to manage patients in the relapsed/refractory setting if they’ve started with one targeted therapy. An example of data from Dr. Kerry Rogers shows PFS looks quite good when sequencing from ibrutinib to acalabrutinib. This is in the setting of intolerance, of course. There were very limited adverse events.
You can think of a sequence where you can go from BTK inhibitor to BTK inhibitor in the setting of intolerance. In the data for venetoclax 24-month median PFS—Dr. Brown already pointed this out—I think it’s unfair to make a comparison between this study and the MURANO trial. Certainly, MURANO was a BTK inhibitor-naïve population; however, patients are much less heavily pretreated in MURANO. I would argue it’s probably somewhere in between where the median PFS for venetoclax plus rituximab lands in a patient population that previously received a BTK inhibitor.
Even the National Comprehensive Cancer Network guidelines include venetoclax retreatment, so it must be okay. You can think about going from venetoclax—if you’re BTK inhibitor-naïve—to a BTK inhibitor. You can go to ibrutinib. The median PFS is 32 months, and the response rate is 84%. These aren’t just data we generated; these data were also generated in Australia and are nearly identical with a median PFS of 34 months. You could say, “You accumulate this resistance along the way if you start with a BTK inhibitor and then you go to venetoclax or vice versa in these double-exposed or double-refractory patients.” However, there are drugs like the non-covalent BTK inhibitors, like pirtobrutinib, that have been developed. Even in the double-exposed patient population, there is a median PFS after five prior lines of therapy of 18 months.
The conundrum that remains is, should we use all our best agents together or save some of the ammunition for the future? This is just my own math, probably not completely accurate, but I think it makes the point: If you start with ibrutinib monotherapy as first-line therapy, I would guess the medium PFS conservatively is going to be about 100 months; it hasn’t been reached yet to my knowledge. In the second-line setting, you could think about venetoclax monotherapy PFS being about 24 months, or even longer with venetoclax plus rituximab. I’m going to discount the MURANO data and say it’s not 53 months, it’s 40 months. In third-line treatment, you might get 18 to 20 months with pirtobrutinib. Ibrutinib plus venetoclax has to have a median PFS of 130 or 140 months in order to just meet the data that we would estimate from sequencing monotherapies across three lines of therapy.
There’s almost no data on retreatment or mechanisms of resistance. Without long-term data, we should not dangle the word cure for patients with any novel agent combination. Certainly, if a cure is not there, if a contribution of effect is not there, and if sequencing data is not there, then just saying there is a high rate of undetectable MRD alone is not enough for me.
I’ll end with this quote, “More is not always better. Sometimes it’s just more.”
© 2025 Mashup Media, LLC, a Formedics Property. All Rights Reserved.