CE / CME
Pharmacists: 0.75 contact hour (0.075 CEUs)
Physicians: Maximum of 0.75 AMA PRA Category 1 Credit™
Nurses: 0.75 Nursing contact hour
Released: August 30, 2022
Expiration: August 29, 2023
Jennifer Woyach, MD:
In CLL, MRD is generally defined as <1 CLL cell in 10,000 leukocytes.1 You also might see this referred to as <0.01% of the cells or <10‑4.
With many fixed-duration therapies, undetectable MRD is associated with longer progression‑free survival. The standard methods of detection include multicolor flow cytometry, next-generation sequencing (NGS), and, especially in clinical trials, allele-specific oligonucleotide polymerase chain reaction (PCR), or allele-specific oligonucleotide PCR (ASO‑PCR). We will talk about each of these methods of detection and why you might choose one over another.
Jennifer Woyach, MD:
Until the FDA authorization of the NGS method clonoSEQ, flow cytometry was the most widely used method of MRD detection in CLL. The European Research Initiative on CLL (ERIC) has harmonized methods of flow cytometry–based MRD detection in CLL, and their standardized panels result in reliable MRD detection at levels of 10‑4 to 10‑5.2 Flow cytometry is widely used, relatively inexpensive, and does not require any pretreatment sample. However, this method does require a flow laboratory and is not approved by the FDA at this time.
Another standard method of MRD detection, which has been used primarily in clinical trials, is real‑time PCR using allele‑specific oligonucleotides of the immunoglobulin heavy chain (ASO-PCR). This method is very sensitive; it is comparable with flow cytometry at levels of 10‑4 and can reliably detect to 10‑5.3 However, ASO-PCR is not widely used because it is very time consuming and expensive.
In 2018, Adaptive Biotechnologies received FDA authorization for the clonoSEQ NGS-based method of MRD detection in CLL. This method sequences the clone‑specific CDR3 region of the heavy chain in a baseline sample with disease, then looks for that specific sequence in a post‑therapy sample.4 The advantage of this method is that it is extremely sensitive, with reliable detection to a level of 10‑6, or 1 cell in 1,000,000 leukocytes. No local facilities are needed because the samples are sent to run centrally at Adaptive Biotechnologies. One disadvantage of this method is that it is relatively expensive, although insurance often covers the cost because it is an FDA-authorized method. clonoSEQ also requires a pretherapy sample because the test is patient specific. The CDR3 region is not known to change in CLL, so the baseline sample can come from any point when the disease was present, even a banked sample from years earlier.
Jennifer Woyach, MD:
With these options for evaluation of MRD, we next have to think about when this type of assay might be helpful. With the currently available evidence, it does not appear to be helpful to evaluate MRD in patients taking a continuous therapy, such as a BTK inhibitor. These therapies rarely lead to undetectable MRD, and even in cases when they do, it is unclear if achieving undetectable MRD, especially in the blood, improves outcomes for the patient.
Anthony Perissinotti, PharmD, BCOP:
It is notable that currently approved BTK inhibitors are highly efficacious, but that most patients receiving these agents will not achieve undetectable MRD. With BTK inhibitor monotherapy or combination therapy with an anti-CD20 antibody, we don’t assess MRD because patients are not imminently relapsing despite being MRD positive throughout their entire therapy.
Jennifer Woyach, MD:
With some fixed-duration regimens, however, there are data suggesting that MRD assessment following the completion of therapy can help determine prognosis and may, in the future, help us decide how to treat patients.
Jennifer Woyach, MD:
We’ve known for some time that patients who achieve undetectable MRD after fludarabine, cyclophosphamide, and rituximab (FCR) chemoimmunotherapy have the potential to be cured with this regimen.5,6 Data from the CLL8 study and an MD Anderson cohort study showed that patients who have undetectable MRD at the end of FCR chemoimmunotherapy have a longer PFS than those with detectable MRD. In the long‑term data from the MD Anderson group, approximately 80% of patients with IGHV-mutated MRD‑negative disease went on to have very long‑term PFS and may have been effectively cured with this regimen.
An MRD assessment at the end of treatment can also predict long‑term outcomes after a venetoclax‑based therapy given in combination with an anti‑CD20 antibody. In the phase III MURANO study in patients with relapsed/refractory CLL, the majority of patients who were treated with venetoclax plus rituximab achieved undetectable MRD with this regimen at some point during treatment.7,8 A higher proportion of patients who achieved undetectable MRD had 18‑month PFS (90.3%) than those with low but detectable MRD (64.4%) or high MRD (8.33%) at the end of treatment.
Similarly, in the CLL14 study of frontline obinutuzumab with either venetoclax or chlorambucil, patients who received venetoclax plus obinutuzumab achieved high rates of undetectable MRD over time during the course of therapy.9 In addition, patients with undetectable MRD at the end of treatment, especially those with IGHV-mutated disease, went on to have long PFS off treatment.
Jennifer Woyach, MD:
These data make it clear that undetectable MRD predicts PFS for patients treated with chemoimmunotherapy as well as for those treated with venetoclax plus anti‑CD20 antibodies. We can also see that serial measurements of MRD following therapy can reliably predict progression events, although there is obviously a delay between when MRD first becomes detectable in the blood and when the patient eventually relapses. In a follow-up of the MURANO study, there was approximately 2 years between first MRD detection and clinical progression.8
Anthony Perissinotti, PharmD, BCOP:
MRD is useful to determine a patient’s prognosis, again, depending on their treatment regimen. It is interesting that in trials like CLL14, patients who achieved MRD negativity with venetoclax plus obinutuzumab appeared to have improved rates of PFS compared with patients who were MRD negative with chlorambucil plus obinutuzumab. This opens up further lines of questioning to which we do not yet have answers: Do patients achieve “deeper” MRD negativity with venetoclax plus obinutuzumab than that achieved with chemoimmunotherapy? Is there some other quality to these responses? If you can achieve MRD negativity at a very deep level, are you curing patients?
Nichole Fisher, RN, BSN:
From a nursing perspective, MRD assessment is not performed in many cases outside of a clinical trial, so it is not something that we often think about in clinical practice. In the context of a clinical trial, it can be helpful to explain to patients how they will be tested for MRD and discuss how undetectable MRD findings can reinforce that the drug they are taking is working and that they are doing well.
Jennifer Woyach, MD:
Venetoclax is currently being investigated as a combination therapy with BTK inhibitors in some exciting new regimens, so next we’ll look at how MRD might be useful in these scenarios. An important caveat here, though, is that no study of venetoclax plus a BTK inhibitor has sufficient follow-up to determine whether MRD assessment is important at the end of treatment.
The first study I want to discuss is a phase II study of venetoclax plus obinutuzumab and ibrutinib in CLL.10 This triplet was given to 25 patients with treatment-naive CLL and 25 patients with relapsed/refractory disease for a fixed duration of 14 cycles, which was just longer than 1 year. After the end of therapy, some patients switched from MRD negative to MRD positive over time; however, this has not yet translated into a PFS difference. In this study, approximately 66% of patients had undetectable MRD at the end of treatment, but because almost all patients who receive treatment remain in remission, we don’t know whether MRD status at the end of treatment predicts PFS outcomes.
In a phase II study of 2 years of ibrutinib plus venetoclax in 80 previously untreated patients with CLL, the rate of undetectable MRD increased over time from 56% after 12 cycles to 66% after 24 cycles.11 The 3‑year PFS rate for this group was 93% and, at least at this point, we cannot see a survival difference between patients who were MRD positive or MRD negative at the end of treatment.
The CAPTIVATE study is a large phase II trial that evaluated ibrutinib plus venetoclax in 2 cohorts of treatment-naive patients with CLL. In the fixed-duration cohort, patients received therapy for 1 year, then stopped treatment. In the second cohort, the treatment duration was guided by MRD status.
The fixed-duration cohort enrolled 159 patients. After 1 year of treatment and 1 year of follow-up, only 5% of patients had progressed. At this point, the attainment of undetectable MRD has not predicted PFS.12,13
The phase III GLOW trial used a similar design. Previously untreated patients with CLL received 1 year of treatment with ibrutinib plus venetoclax, then stopped treatment.14,15 Again, we haven’t yet seen a difference in progression‑free survival between patients who achieved undetectable vs detectable MRD at the end of treatment.
Jennifer Woyach, MD:
Altogether, these data suggest that, at least in the short term, MRD status does not help inform prognosis for BTK inhibitor plus venetoclax regimens because almost all patients remain in remission. Longer‑term data will be helpful to determine whether we can use MRD as a prognostic factor with this type of regimen.
Jennifer Woyach, MD:
The next question is whether we can use MRD status to guide therapy. As discussed earlier, the MD Anderson study of frontline ibrutinib plus venetoclax found that the undetectable MRD rate increased between 1 and 2 years, suggesting that outcomes may be improved by continuing the duration of therapy.11
In the MRD cohort of the CAPTIVATE study, patients with undetectable MRD after 1 year of ibrutinib plus venetoclax were randomized to either placebo or ibrutinib maintenance.16 Patients with detectable MRD at the end of the 1‑year treatment were randomized to either ibrutinib or ibrutinib plus venetoclax. At 36 months of follow-up—1 year of the initial therapy and 2 years of maintenance—there was no difference in PFS between patients receiving placebo or ibrutinib, or those receiving ibrutinib or ibrutinib plus venetoclax maintenance.
A similar study is the phase II BOVen study, in which treatment‑naive patients with CLL were treated with zanubrutinib plus venetoclax plus obinutuzumab, and the length of therapy was determined by how long it took to achieve an undetectable MRD state.17 MRD was assessed beginning with cycle 7 and then again every 2 cycles. Treatment was discontinued after a patient had undetectable MRD in blood and bone marrow. The median time to undetectable MRD in bone marrow in this study, so far, was only 6 months, and 77% of patients discontinued therapy at a median of 10 months. Although the follow-up is still very short at 14 months, no recurrent MRD or progression has been observed. These early results suggest that it may be feasible to adjust the duration of therapy and even shorten it based on MRD assessment, although we don’t yet know how successful this approach will be in providing the longest possible remission duration with the lowest amount of toxicity.
Jennifer Woyach, MD:
Overall, the results of these trials indicate that we can use MRD status to guide therapy, but it remains unclear how helpful this will be. I would not recommend this outside of a clinical trial right now.
With this in mind, an important question is when we should order MRD testing. I usually order MRD testing following venetoclax‑based therapies, but I don’t change management based on the results at this time. For patients receiving venetoclax after progressing on ibrutinib, I usually order an MRD assessment after 2 years of venetoclax, then consider continuation of treatment based on the results. I would order either flow-based or NGS MRD assessment, depending on the situation and what’s available.
Anthony Perissinotti, PharmD, BCOP:
Although MRD can be prognostic with some treatment regimens, should we use MRD results to guide treatment? I, personally, don’t think MRD results are actionable in current clinical practice. Considering the data from CLL14, many patients who were MRD positive with venetoclax plus obinutuzumab still achieved remarkable outcomes, and many patients won’t relapse for several years despite having detectable MRD. Even for patients who were MRD negative at one time before becoming MRD positive, there will still be a long window of rising MRD before the patient will overtly relapse. My concern at present is that if we are checking MRD in clinical practice, we might feel compelled to unnecessarily continue therapy on a patient who is MRD positive. Similarly, if a patient who previously had undetectable MRD is now MRD positive, we may feel compelled to restart treatment. I think we need to take great care with how we interpret and use MRD results outside of a clinical trial, so that we are not overtreating patients. We do not currently know whether or not we are improving outcomes for patients in the aforementioned 2 scenarios, but it is possible that we are subjecting patients to unnecessary treatment and associated side effects. As CLL is typically a slow-growing disease, we can observe patients for symptoms before deciding on treatment. And if we are not using MRD to make decisions, it begs the question of why we would conduct a test for which we have no plan for the results. Clinical trials will hopefully provide us with answers to the current questions surrounding the use of MRD.
With data from the venetoclax plus BTK inhibitor trials showing incredible PFS, it begs the question: When we have such effective therapies, do we really need to be chasing MRD? If outcomes are this good, giving more therapy to an MRD‑positive patient may mean that we are overtreating patients. Providing additional therapy to an MRD-positive patient would assume that their CLL remains sensitive to therapy; it may make sense to give these patients a treatment break because their cells should still be sensitive when they have morphologically relapsed.
Jennifer Woyach, MD:
Another important question when deciding to assess MRD is whether you can use blood vs marrow to evaluate. Several studies have looked at the concordance of peripheral blood and marrow in terms of MRD, and most studies find between 75% and 90% concordance. Although this is a reasonably high level of concordance, it also means that there are still a significant number of patients who would have a negative blood MRD result but a positive marrow MRD result. My approach is to order a blood MRD test first, and if it’s negative, discuss bone marrow testing with the patient.
Jennifer Woyach, MD:
In conclusion, MRD is a powerful prognostic tool in CLL, especially for chemoimmunotherapy‑based regimens and venetoclax/antibody therapies. It may also be applicable to venetoclax plus BTK inhibitor–based therapies, but there is no evidence of a survival benefit with undetectable vs detectable MRD at this time; we will need to wait for longer follow-up of the trials assessing these combinations.
For MRD assessment, both clonoSEQ and flow cytometry provide excellent sensitivity in detection, and ongoing and future studies will help determine whether modifications of therapy based on MRD are beneficial to our patients.
Anthony Perissinotti, PharmD, BCOP:
Currently, although MRD status can aid in prognosis for some therapies, I don’t recommend the use of MRD to guide treatment outside of a clinical trial. I do think the use of MRD data is a very important research question, and I think data from clinical trials will shed more light on ongoing questions in using MRD assessment in clinical practice. In the future, MRD may be a tool that allows us to identify patients who can stop therapy and who should continue or restart therapy. For now, we will continue to wait for clinical trial data with MRD to potentially guide our approach.
Let’s return to our question from earlier in the program.
Back