Combinations or sequences of targeted agents in CLL: is the whole greater than the sum of its parts (Aristotle, 360 BC)?

And changing it is. Chemoimmunotherapy (CIT) is being rapidly eschewed in favor of targeted approaches in the treatment of chronic lymphocytic leukemia (CLL). Derivatives of alkylating agents and nucleoside analogs with their associated cytopenias, infections, and notably, secondary malignancies are becoming mere relics along the side of the road toward a chemo-free world.

The era of targeted agents (TAs) has come rapidly upon us. These small molecules inhibit a variety of intracellular pathways, including Bruton tyrosine kinase (BTK), phosphatidylinositol 3-kinase (PI3K), B-cell lymphoma-2 (BCL-2), and others. Their impressive activity and ease of administration with relatively favorable toxicity profiles in CLL make them attractive therapeutic options.^1-5  These drugs markedly improve outcomes of patients with the dreaded 17p deletion (17p-) and other poor risk features, while comparing quite favorably (both from an efficacy and safety perspective) with standard CIT regimens in the majority of patients.

Nevertheless, progress in medicine must always remain in our rearview mirror, leaving behind what has been accomplished and searching for the next advances of the future. Despite the efficacy of current TAs, numerous challenges remain regarding their optimal use. Perhaps the most compelling question is whether these drugs are best used as sequential monotherapies or in combinations. Was the Greek from Stagira prescient when it came to TAs?

The availability of ibrutinib first raised the potential of a chemo-free future for CLL patients. In relapsed or refractory (R/R) CLL, this oral, irreversible BTK inhibitor demonstrated an impressive overall response rate (ORR) of 89%, leading to its first of several US Food and Drug Administration (FDA) approvals.¹  In the confirmatory RESONATE trial, patients with R/R CLL were randomized to ibrutinib vs ofatumumab. Ibrutinib was clearly superior with respect to overall survival (OS) (hazard ratio [HR], 0.43; P = .005) and progression-free survival (PFS) (not reached vs 8.1 months).^2,6  Recently, O’Brien and coworkers updated the ibrutinib data in both R/R as well as treatment naïve (TN) patients demonstrating an impressive 5-year PFS rate of 92% in 31 TN patients and 44% in 101 R/R patients.⁶  Overall, TN patients received ibrutinib for a median of 65 months. However, 45% discontinued treatment, mainly because of intolerance (19%) or disease progression (6%). The median treatment duration for R/R patients was shorter, 39 months, and 39% continued ibrutinib for >4 years. Seventy-two percent discontinued treatment, mostly because of disease progression (33%), followed by intolerance (21%), in contrast to TN patients.⁶  Additional data on prolonged therapy with ibrutinib confirmed its activity in patients with TP53 aberrations (17p- and/or TP53 mutation), and those with a poor outcome with CIT, with ORR 95.8% and an estimated 5-year PFS 58.2% in 1 study and an ORR 83% with a 24-month PFS of 63% in another.^6-8 

Sometimes when you ask the wrong question, the answer is more interesting. In the recently updated HELIOS trial, ibrutinib plus bendamustine rituximab (BR) was superior to BR alone (36 month PFS 68% vs 13.9%); however, an indirect comparison between the contour of the BR-ibrutinib PFS curve of HELIOS (from the original publication) with that from the RESONATE trial suggested that the benefit from BR-ibrutinib largely reflected the effect of ibrutinib.^2,9-11  Unfortunately, the proper study of ibrutinib vs BR-Ibrutinib has not been conducted.

Impressive results with ibrutinib in R/R patients stimulated frontline clinical trials. In RESONATE-2, 269 TN patients, age ≥65 years without 17p-, were randomized to ibrutinib vs chlorambucil. With a median follow-up of 18.4 months, ibrutinib achieved a longer PFS (18.9 months vs not reached) confirming an 84% reduction in the risk of progression or death (HR, 0.16; P < .001), with a prolongation of OS (98% at 24 months with ibrutinib vs 85% with chlorambucil [HR, 0.16; P = .001]).¹²  Moreover, ibrutinib was safe and effective even in patients older than 71 years (5-year PFS of 81.2%).^8,13 

To underscore the impressive activity of ibrutinib monotherapy in the frontline setting, Robak et al compared data with ibrutinib from RESONATE-2 with fludarabine, cyclophosphamide, rituximab (FCR) from the CLL8 study; FCR/BR from CLL10; chlorambucil/obinutuzumab or rituximab from CLL11; and chlorambucil/ofatumumab from COMPLEMENT-1.¹⁴  This comparison suggested a superior PFS with ibrutinib, supporting the fading role of CIT even in frontline. Conversely, in the relapse setting, Cuneo et al compared ibrutinib monotherapy to BR as first salvage in a matched-adjusted, indirect, retrospective analysis of CLL patients, in patients with intact 17p and no difference in OS was detected (63% and 74.4% alive at 36 months, respectively). One of the possible explanations could be a greater proportion of patients with high-risk CLL in the ibrutinib and BR group, respectively (17p- 36.1% vs 14.8%). PFS was most impressive using BR as first salvage if IGHV was mutated, 17p- was not present, and Rai stage ≤2 in multivariate analyses, suggesting that in first relapse, favorable risk patients CIT could be considered.¹⁵  Three large randomized, fully accrued, phase 3 trials compared ibrutinib-based treatment with CIT as initial therapy: in an ALLIANCE-led study, patients older than 65 were randomized to BR, ibrutinib/rituximab, or ibrutinib alone (A041202 NCT01886872). In an Eastern Cooperative Oncology Group–led study, patients received ibrutinib/rituximab or FCR (E1912 NCT02048813). In the UK FLAIR study (2013-001944-76), untreated patients also received ibrutinib/rituximab or FCR. A recent amendment added ibrutinib monotherapy and ibrutinib/venetoclax arms to the study. These results should be informative for frontline choices, but will not inform on the role for CIT at first relapse vs ibrutinib.

Alternate BTK inhibitors have been developed to improve efficacy and reduce toxicity compared with ibrutinib. Agents such as acalabrutinib and zanubrutinib are more selective BTK inhibitors that, although results from studies directly comparing them with ibrutinib are not yet available, induce at least similar responses in R/R CLL (ORR 95% [100% in deletion (17p)] and ORR 92%, respectively) with perhaps more favorable tolerability.^16,17  Noncovalently binding BTK inhibitors (SNS 062, LOXO 305, ARQ 531) are in early stages of clinical development as monotherapies. Preclinical data suggest that these agents may overcome ibrutinib resistance mechanisms, particularly in the setting of BTK Cys481mutations.^18-20 

Idelalisib, an oral PI3 δ-specific kinase inhibitor combined with rituximab and compared with rituximab plus placebo in a phase 3 randomized study in medically unfit patients with R/R CLL, not suitable for cytotoxic therapies, achieved a significantly longer PFS compared with rituximab plus placebo (not reached vs 5.5 months), with a 12-month OS benefit (92% vs 80%, P = .02).³  Serious adverse events occurred in 40% of patients with idelalisib including transaminitis, pneumonitis, and colitis. Umbralisib, a next-generation δ-isoform PI3K inhibitor, demonstrated an ORR 85% (75% in high-risk patents) in 90 patients with R/R lymphomas and CLL (n = 20) and appears to have a lower risk of severe colitis, pneumonitis, or transaminitis.²¹  Additional promising PI3K inhibitors are in development.^22,23 

CLL is a disorder primarily of lymphoaccumulation.²⁴  Thus, facilitating apoptosis is a promising strategy. Venetoclax, an oral, bcl-2 inhibitor has impressive efficacy in R/R CLL. In a phase 1 dose-escalation study in 116 patients with heavily pretreated and high-risk R/R CLL, 79%, including those with 17p-, achieved a response (complete response [CR] 20%).⁴  Venetoclax demonstrated similar impressive activity in a phase 2 study in patients with 17p- (ORR 79.7%, CR 8%).⁵  At a median follow-up of 12.1 months, median PFS and OS were not reached.

Although data for TAs are impressive, most patients will likely require a sequence of agents over time. Reasons for discontinuation and the preferred subsequent sequence of agents become critical issues.²⁵  In the first large series to address sequencing of TA, Mato et al investigated the outcome of 178 patients following discontinuation of ibrutinib or idelalisib.²⁶  Patients were distinguished into 3 groups: (1) kinase inhibitor (KI) naïve, (2) KI intolerant, and (3) KI progression. KI-naïve patients were further characterized into 2 subgroups: TN and prior CIT. Most patients had high-risk disease (median of 3 prior therapies [range 0-11]; 17p- [34%], p53 mutation [27%], del11q [33%], and complex karyotype [29%]). The most common reasons for discontinuation were toxicity (51%), CLL progression (29%), and Richter transformation (9%). After discontinuation, more than half the patients were treated with an alternate KI. KI-intolerant patients could be successfully treated with an alternate KI. With a median follow-up of 14 months, the median PFS after the second KI was not reached compared with 7 months in patients who had CLL progression as the reason for discontinuation of the first KI. Prospective studies are attempting to confirm that KI-intolerant patients could be switched to another KI, such as umbralisib or acalabrutinib, and maintain response.^27,28 

To identify the optimal sequencing of TA, Mato et al analyzed 683 patients (86% R/R CLL) treated most recently with KI or venetoclax.²⁵  Those treated with ibrutinib had a longer PFS than idelalisib in frontline (HR 2.8, P = .01), R/R (HR 2.8, P < .001), 17p- (P = .008), and complex karyotype (P = .02).²⁵  Furthermore, at the time of KI failure, venetoclax or an alternate KI prolonged PFS compared with CIT (median PFS not reached with alternate KI or venetoclax vs 5.1 months with CIT, P < .001). Moreover, ORR and CR rates were higher in patients who received venetoclax after KI failure. The ORRs (CR) to CIT, alternative KI and venetoclax were 49.4% (2.1%), 58.5% (4.1%), and 73.6% (31.5%), respectively. Thus, venetoclax was preferred after ibrutinib discontinuation, particularly following CLL progression resulting in ibrutinib or idelalisib discontinuation.²⁵  Notably, CIT resulted in poor outcomes when used as salvage after KI failure as most patients in this series had likely seen prior CIT.

Jones and colleagues investigated 127 patients with CLL who received venetoclax following ibrutinib discontinuation and reported an ORR 65% and median PFS 24.7 months.²⁹  They also noted promising activity and favorable tolerability in 36 patients with CLL receiving venetoclax after idelalisib discontinuation.³⁰  With a median follow up of 14 months, the median PFS was not reached, and estimated 12-month PFS was 79%.

The most obvious knowledge gap regarding sequencing of TAs is the optimal treatment following venetoclax discontinuation. A multicenter collaboration studied outcomes of 141 patients previously treated with venetoclax in a real-world setting, including 24 who received subsequent therapy. No clearly effective sequencing pattern following venetoclax was identified (KI, n = 7; rituximab monotherapy, n = 3; anthracycline-based therapy, n = 3; cellular therapy, n = 5).31 Prospective data to support a sequence where venetoclax supersedes ibrutinib are needed.

So, if 1 TA is good, is a combination of them any better? Can they overcome residual disease, acquired resistance, suboptimal durability of response with high-risk disease, cost of indefinite treatment duration, and decreased compliance over time? To address these challenges, an increasing number of studies are evaluating different TA combinations (Tables 1 and 2).

In the phase 3, randomized GENUINE trial, 126 patients with R/R high-risk CLL (17p-, 11q deletion, or p53 mutation) received ibrutinib alone or with ublituximab.³²  At a median follow-up of 12 months, this combination achieved a superior ORR compared with ibrutinib (ORR 80% vs 47%, P < .001). In a preliminary analysis, significant prolongation of PFS was not observed. Woyach et al evaluated acalabrutinib in combination with obinutuzumab in R/R (n = 26) and TN (n = 19) patients with CLL (including 17p- and mutated immunoglobulin heavy-chain variable region gene [IGHV]) in a single arm phase1b/2 study.³³  Acalabrutinib was administered daily until progressive disease, with obinutuzumab on days 1, 2, 8, and 15 of cycle 2 and day 1 of cycles 3 to 7. ORRs (CR) were 95% (16%) and 92% (8%), respectively. Median PFS was not reached in either cohort at a median follow-up of 17.8 months in TN and 21.2 months in R/R patients. Results from this combination appear similar to acalabrutinib monotherapy (ORR 85%, 18-month PFS 88%), again raising questions regarding the additional value of an anti-CD20 antibody with a BTK inhibitor.³⁴  The most sobering example was presented by Burger and colleagues in a randomized phase 2 trial of ibrutinib vs ibrutinib plus rituximab in 206 patients with mostly high-risk CLL.^35,36  Ibrutinib was given continuously until progression or intolerance in both arms. Surprisingly, there was no difference in PFS between the cohorts treated with ibrutinib or ibrutinib plus rituximab during the 25.2 and 22.7 months observation time (91.2% vs 90.4%, P = .788). Furthermore, ORR and CR rates were similar (ORR 98% vs 100%, CR 21% vs 28% P = .309). More patients on the combination arm discontinued treatment because of toxicities. These examples support the conclusion that there is no benefit to combining rituximab or ublituximab with a BTK inhibitor as compared with BTK inhibitor monotherapy. The iLLUMINATE trial compared ibrutinib plus obinutuzumab to obinutuzumab plus chlorambucil and suggested a superior PFS in the frontline setting.³⁷  Unfortunately, ibrutinib monotherapy was not included as a control arm, which is a major limitation of the study.

Despite the high ORR of venetoclax monotherapy in CLL, CR rates are not impressive (although higher than reported for KI monotherapies), mostly because of residual lymphadenopathy. Furthermore, venetoclax is administered indefinitely. Therefore, combination strategies have been evaluated to improve depth of response and, thus, patient outcomes, utilizing time-limited designs. Flinn et al combined venetoclax with obinutuzumab in a response-adapted phase 1b study in 32 TN CLL patients (including 17p-).³⁸  TN patients received 6 cycles of the combination, followed by an additional 6 cycles of venetoclax monotherapy. Venetoclax could be extended after 1 year of treatment depending on disease status. The ORR was 100% (CR/CR with incomplete marrow recovery 56.3%, partial remission [PR] 43.8%). Median time on study was 11.3 months, and PFS was 100% at 1 year. Undetectable peripheral blood MRD was achieved in all patients, and in 62.5% of available bone marrow samples. These encouraging data led the German CLL Study Group to compare the combination of venetoclax and obinutuzumab to obinutuzumab plus chlorambucil in the CLL14 study (NCT02242942).^39,40  Venetoclax plus obinutuzumab following bendamustine induction resulted in low rates of tumor lysis syndrome and infusion reactions, with an ORR of 100% in frontline and 94% in R/R patients.⁴¹ 

In the MURANO trial, 389 patients with R/R CLL were randomized to rituximab plus either venetoclax (VR, n = 194) for 2 years or until disease progression or intolerance, or bendamustine (n = 195).⁴²  With a 23.8 month-median follow-up, median PFS was not reached in the venetoclax arm compared with 17 months with BR (P < .001). VR achieved an ORR 92.3% (CR 26.8%), higher than previously reported with venetoclax monotherapy (ORR 65% to 79%, CR 9% to 20% and median PFS 24.7-27.2 months).^4,29,43 Additionally, a higher rate of undetectable MRD was achieved with venetoclax and rituximab (84% vs 23%) in peripheral blood. Based on these findings, venetoclax and rituximab was granted FDA approval for patients with CLL following at least 1 prior therapy. Unfortunately this trial did not include a venetoclax monotherapy arm; therefore, the relative contributions of each agent in VR cannot be delineated.

In the CLARITY study, venetoclax plus ibrutinib was administered to 50 R/R CLL patients (including 17p-).⁴⁴  This combination was supported by preclinical studies showing that BTK inhibitors increase BCL-2 dependence and sensitize CLL cells to venetoclax.^45-47  Venetoclax was added after 8 weeks of ibrutinib monotherapy. Dosing of ibrutinib and venetoclax was based on time to achieve undetectable MRD. Patients with undetectable MRD at 6 or 12 months of the combination received an additional 6 or 12 months of therapy. All 25 patients who completed 8 months of therapy (6 months of combination) responded (ORR 100%, CR 60%) and 28% achieved undetectable MRD. ORR and CR rates appeared higher than expected with monotherapy with venetoclax or ibrutinib, although eradication of MRD was not clearly superior.

The CAPTIVATE trial is exploring the combination of venetoclax and ibrutinib in untreated CLL. Randomization is based on MRD status after combination therapy, but with a different schedule compared with the CLARITY trial.⁴⁸  Patients with confirmed, undetectable MRD after 12 cycles of combination therapy are randomized to placebo or ibrutinib. If undetectable MRD is not confirmed, patients continue on the venetoclax-ibrutinib combination or are randomized to ibrutinib alone. All 14 patients enrolled for safety run-in who had completed 6 cycles of the combination responded (confirmed CR in 1; 13 PR). Undetectable MRD was reported in 82% (9/11) of the assessed patients. Although we have data that MRD status influences the outcome of patients, at this time, the importance of undetectable MRD in patients on TAs who have achieved a PR is not known. Similar to when BCL-2 inhibitors are administered as a monotherapy, the combination of ibrutinib and venetoclax may remove the more sensitive CLL subclones more rapidly in the periphery and bone marrow, while the venetoclax-resistant subclones persist in the lymphoid tissues. Therefore, the patient achieves a MRD-undetectable PR rather than a CR. Whether patients who achieve MRD-undetectable PRs with TA combinations have clinical outcomes similar to those with MRD-undetectable CR remains unknown. This combination was well tolerated, yet longer follow-up is critical to show that durability of response is superior to what might be expected with the 2 drugs given in sequence.

Considering the impressive results of doublet TA combinations, Rogers and colleagues investigated venetoclax, ibrutinib, and obinutuzumab in a phase 1b study in 25 TN CLL patients (including 17p-).⁴⁹  Treatment was planned for 14 cycles and the drugs were started sequentially over the first 3 cycles. At the interim analysis, 23 patients had reached the first (post–cycle 8) response assessment and all had an objective response with CR/CR with incomplete marrow recovery 50%. Undetectable MRD was achieved in blood and bone marrow in 58% of patients. Based on the current randomized data available with rituximab or ublituximab plus ibrutinib and the depth of response already observed with novel-novel doublets, the addition of an anti-CD20 mAb to a doublet such as venetoclax and ibrutinib may not add significant benefit.³²  It may improve the rapidity of response, but without affecting overall outcome. Much like prior studies examining chemotherapy doublets vs CIT triplets, randomized clinical trials with appropriate controls are critically needed to further investigate this hypothesis. This combination was studied in 12 patients with R/R CLL with an ORR and CR rate of 92% and 42%, respectively.⁵⁰  Furthermore, 50% of patients achieved undetectable MRD in blood and marrow.

The majority of combination regimens appear to achieve a significantly higher likelihood of undetectable MRD compared with single TA and standard CIT combinations. Based on the prognostic value of MRD (mostly extrapolated from CIT studies), it is being used more frequently as a surrogate marker. Pooled data from the German CLL8 and CLL10 trials, including FCR and BR treated patients, demonstrated that MRD status outperformed traditional response assessments when predicting PFS.⁵¹  Despite enthusiasm for TA combinations, questions remain regarding the predictive value of MRD status, durability of response, fixed treatment durations, and, importantly, criteria for selecting the optimal combination for individual patients. To date, studies have been mostly small, single arm trials in heterogeneous patient populations using different methods for response and MRD assessment. Longer follow-up is critical to confirm the durability responses and impact on outcomes.

At first glance, recently reported TA combinations are impressive and suggest another impending paradigm shift. Response rates are nearly 100% in all studies, with high rates of undetectable MRD, with the promise of time-limited therapies. Treatment-free intervals are also desirable to reduce the financial burden, particularly in the United States with the highest health care costs in the world. Nonetheless, there are serious limitations in the design of the current iterations of combination studies, and, therefore, with the exception of the MURANO (VR) and iLLUMINATE regimens (ibrutinib+ obinutuzumab), combinations of TAs are not yet ready for routine clinical practice. Proponents of such approaches would surely highlight the impressive ORR, CR rates, and MRD data obtained with the doublets, triplets, or even quadruplets presented at recent national and international meetings. The fundamental problem is the limited randomized data against clinically relevant controls demonstrating that novel-novel combinations improve outcomes that are important to our patients. Therefore, the use of the word “cure” is not yet close to reality.

Whereas combinations appear to increase the likelihood of MRD eradication, a surrogate for PFS with CIT, whether this observation can be extrapolated to TA combinations has yet to be demonstrated, with a few exceptions.^29,42,52  The lesson learned from the CLL8 and CLL10 studies is that not all MRD is created equal.⁵¹  Outcomes of patients with IGHV unmutated, MRD undetectable CLL treated with FCR are poor, relative to IGHV mutated CLL, despite similar depths of response. How prognostic markers and MRD status interplay in patients treated with TAs remains unknown. The question arises: Do we have enough evidence to use MRD as the surrogate end point in clinical trials and eventually practice? The answer is no. Given the durability of responses to frontline ibrutinib despite failure to eradicate MRD, the importance of MRD status can be further questioned.⁶  In addition, identifying patients destined to have inferior outcomes with TAs is still in its infancy. For example, in the RESONATE study, the 18-month PFS with ibrutinib was similar regardless of baseline genetic factors including unmutated IGHV, del17p, del11q, or mutations including NOTCH1, SF3B1, ATM, and BIRC3.⁵³  With the exception of del17p and/or complex karyotype, biomarkers to identify which patients might benefit from TA combinations over monotherapies are largely unknown and untested with TA combinations. Based on durability of response to ibrutinib (despite the lack of CRs in most patients) in long-term follow-up, one could hypothesize that residual CLL cells in the bone marrow are biologically senescent, not contributing to clinical compromise, and do not necessarily need to be eradicated nor achieving MRD undetectable disease.

Another issue of importance is toxicity in the real-world setting. Large real-world evidence series have consistently demonstrated that, even as monotherapies, novel agents induce far greater toxicities than would have been predicted from the landmark clinical trials. For example, Mato et al reported 3 separate data sets in which more than half of the discontinuations of ibrutinib or idelalisib in clinical practice were because of adverse events.^25,26,31  Similar results were observed by Follows et al.⁵⁴  Toxicities of novel-novel combinations, particularly long- term, are not yet characterized but are expected to be greater than the single components, with increased discontinuations. Although combinations may limit the duration of treatment, they still last 1 to 2 years. Moreover discontinuations because of toxicity usually occur within the first 6 months of treatment.⁵⁵  It is likely that dose modification earlier in the course of treatment could reduce toxicity without compromising efficacy. In a pilot study of 11 patients with CLL, stepwise dose reduction of ibrutinib after the first cycle did not affect the biological activity or clinical outcomes.⁵⁶  A larger study with formal response assessment and pharmacokinetic and pharmacodynamic studies is being planned by the same group.

Then come questions regarding study design and generalizability. Combination studies have largely included single arms, conducted at highly specialized centers, with small samples sizes, notable differences in study entry criteria, and different MRD assessment methods with varying sensitivities. These limitations emphasize the flaws in making any cross-trial comparisons. With the notable exceptions of UNITY CLL (NCT02612311), ALLIANCE (A041202 NCT01886872), and UK FLAIR (2013-001944-76) studies, the major deficit in study design is the omission of respective TA monotherapies as controls (ie, venetoclax + rituximab vs venetoclax monotherapy) and control arms, which consider sequential monotherapies (ie, ibrutinib + venetoclax vs ibrutinib followed by venetoclax at disease progression). For example, median PFS for patients who receive a KI in the relapse setting was estimated at 35 months. Considering the median PFS of 24.7 months for patients who receive venetoclax after ibrutinib, the disease control duration could be extended to at least 5 years (PFS2); therefore, the 5-year bar for PFS needs to be met by combination therapies at a minimum. Whether retreatment with the same novel-novel combination after a treatment-free interval results in an even longer PFS2 is untested in clinical practice, with the exception of 1 limited report.⁵⁷  Considering the impressive PFS with ibrutinib in the frontline and venetoclax at first relapse, it becomes staggering to imagine how to design a novel combination to compete with these data, and the sample size and follow-up required to do so. Although treating until progression may not seem appealing at the surface, in reality time-limited durations included 15 months (CAPTIVATE) or 24 months (MURANO) of therapy, with actually very little practical differences between the treat to progression and current time-limited approaches as many patients will discontinue combinations prior to completing the planned duration of therapy.^42,48  Although only a minority of patients discontinued VR prior to 2 years on the MURANO trial, from prior real-world experiences of TA monotherapies the rates of discontinuation are likely to be magnified significantly as compared with reports from clinical trials. Whether this is the case for TA combinations in clinical practice remains unstudied.

Burger et al (ibrutinib plus rituximab vs ibrutinib), Sharman et al (ibrutinib plus ublituximab vs ibrutinib) and Hillmen et al (HELIOS PFS curve = RESONATE PFS curve) support a need for a single TA control arm in randomized studies.^10,32,36  For now, agents like ibrutinib and venetoclax may offer comparable outcomes when delivered sequentially. Given the finite number of effective options, the high cost of combinations, the limited information regarding the length of resulting treatment-free intervals, and the potential for increased toxicities, whether a sequence or combination is preferred is a critical question. We cannot yet adopt the promising combination regimens without addressing the outcomes from sequencing, the unknowns of clonal evolution, and resistance to combinations.

The countless potential combinations and permutations may expose patients to multiple classes of TA simultaneously. Although most combinations will likely achieve undetectable MRD, finding salvage therapies for patients who have exhausted multiple (or all) current novel drug options, and then progress would be challenging.⁴⁹  Should the goal be to achieve the best response early with multiple drugs possible, or is it preferable to forgo the initial maximum response for the sake of having further choices in the future? The conundrum remains: to use all our best agents at once or save ammunition for the future.

Some might argue that sequencing TA monotherapies are more appropriate for patients with so-called low-risk profile (mutated IGHV, negative TP53, lack of 17p-) or elderly patients with comorbidities. Interestingly, data from the National Cancer Institute suggest that of the small number of patients with undetectable MRD on single agent ibrutinib most frequently are IGHV unmutated.⁸  Defining risk status in the era of TA is still not well understood. Studies suggest that patients with 17p-, complex karyotype, BCL6 rearrangement, and those who are heavily pretreated are at highest risk to progress or transform on ibrutinib monotherapy.^58-60  Woyach and coworkers reported that among patients who experienced relapse on ibrutinib, acquired mutations of BTK or PLCG2 were found in 85%, and these mutations were detected an estimated median of 9.3 months before relapse.⁶⁰  Furthermore, genomic instability in patients with complex karyotype may increase the risk of developing ibrutinib resistance with new mutations such as the BTK C481S mutation and activating mutations in PLCγ2.^58,60  Markers of ibrutinib resistance may be useful in predicting the timing of future relapse and therefore identifying candidates for early intervention by switching to a new TA (eg, noncovalently binding BTK inhibitor) or adding another agent to overcome resistance. Predictors for venetoclax failure are less well understood; however, pretreatment complex karyotype or 17p- appear to predict inferior outcomes.^47,61  Recently developed assays may help analyze resistance mechanisms and predict clinical response to BCL-2 inhibitors.⁶²  Dynamic BH3 profiling assesses early proapoptotic signaling stimulated by a drug and correlates the findings with response to treatment in lymphomas and leukemias.^47,63,64  Recent data examining clonal dynamics from 8 CLL patients with progression on venetoclax identified mutations in BTG1, CDKN2A/B, BRAF, and amplification of CD274 as potential mechanisms of venetoclax resistance and as novel targets for combination studies in resistant patients.⁶⁵  Incorporating such novel techniques in treatment planning along with response-adapted approaches could help design combination studies for patients at highest risk for venetoclax failure. In an individualized approach, perhaps patients at highest risk for poor disease control on TA monotherapy, such as patients with TP53 aberrations or complex karyotype, are the best candidates for studies of novel combinations. Whether novel combinations (as compared with novel agent monotherapies or CIT) can overcome adverse biology in terms of improving disease control and preventing transformation and the occurrence of second malignancies through immune reconstitution is a question worth asking.

Several reports suggest an increased risk of opportunistic infections with TA monotherapies.^66-69  BTK is a critical part of the innate immune response against Aspergillus, and BTK activation increases levels of reactive oxygen species and nitric oxide, and proteases, which are all important in eliminating the fungal infections.^70,71  Notably, patients who receive ibrutinib are at a higher risk of developing invasive Aspergillus infection.⁶⁹  Whether the risk of opportunistic infections will increase with TA combinations (even with a limited duration of exposure) is unknown and of concern.

The revolution in CLL is still in its infancy. The next paradigm shift will likely result from combinations of TAs. Continuing progress requires 4 important steps: (1) randomized studies comparing the most promising combinations to single agent TAs and to one another, (2) demonstration that multi-TA combinations are better than sequencing of monotherapies, (3) long-term follow-up data of both efficacy and toxicity, and (4) a precision medicine approach based on patient-specific biology. With all the interest in combination regimens, it is important to remember: “More isn’t always better. Sometimes it’s just more” (Barbara Benedek).

Contribution: M.S.Y., A.R.M., and B.D.C. participated equally in drafting, writing, and revising the manuscript.

Conflict-of-interest disclosure: M.S.Y. received honoraria from Bayer as a speaker. A.R.M. holds a consultancy role for TG Therapeutics, Abbvie, Pharamacyclics, Johnson & Johnson, Regeneron, Astra Zeneca, and Celgene and has received research funding from TG Therapeutics, Abbvie, Pharamacyclics, Johnson & Johnson, Regeneron, Portola, DTRM, and Acerta. B.D.C. holds a consultancy role for Abbvie, Acerta, Bayer, Celgene, Gilead, Roche-Genentech, Sunesis, TG Therapeutics, and Astra Zeneca and has received research funding for the institution from Abbvie, Acerta, Roche-Genentech, and TG Therapeutics.

Correspondence: Maryam Sarraf Yazdy, Lombardi Comprehensive Cancer Center, Georgetown University Hospital, 3800 Reservoir Rd NW, Washington, DC 20007; e-mail: maryam.s.yazdy@gunet.georgetown.edu.