Evaluation and management of measurable residual disease in acute lymphoblastic leukemia

Abstract

With standard chemotherapy regimens for adults with acute lymphoblastic leukemia, approximately 90% of patients achieve complete remission. However, up to half of patients have persistent minimal/measurable residual disease (MRD) not recognized by routine microscopy, which constitutes the leading determinant of relapse. Many studies in pediatric and adult populations have demonstrated that achievement of MRD negativity after induction chemotherapy or during consolidation is associated with significantly better long-term outcomes, and MRD status constitutes an independently prognostic marker, often superseding other conventional risk factors. Persistence of MRD after intensive chemotherapy is indicative of treatment refractoriness and warrants alternative therapeutic approaches including allogeneic stem cell transplantation, blinatumomab, or investigational therapies such as inotuzumab ozogamicin or chimeric antigen receptor T cells. Furthermore, the incorporation of novel monoclonal antibodies or potent BCR-ABL1 tyrosine kinase inhibitors, such as ponatinib into frontline treatment may have the advantage of achieving higher rates of MRD negativity while minimizing chemotherapy-related toxicities. Many studies are therefore ongoing to determine whether this strategy can improve cure rates without the need for allogeneic stem cell transplantation.

Introduction

The outcomes of adults with acute lymphoblastic leukemia (ALL) have dramatically improved with the use of multiagent chemotherapy. After standard multiagent chemotherapy, approximately 90% of patients achieve complete remission (CR), which is defined as the presence of less than 5% blasts on routine microscopy and adequate peripheral blood count recovery.^1,2 Despite these high remission rates, relapses still commonly occur. These relapses, which constitute the major cause of death in adults with ALL, are due to the persistence of leukemic blasts that generally exhibit resistance to cytotoxic chemotherapy and are present at low levels, making them undetectable by conventional pathologic assessment. With the use of sensitive technologies such as multiparameter flow cytometry (MFC) and polymerase chain reaction (PCR), persistent leukemia cells can be detected in approximately 30–50% of patients who achieve CR.^3,4 These persistent leukemia cells in the setting of CR are referred to as measurable (also called ‘minimal’) residual disease (MRD), which reflects the remaining disease burden after initial therapy, thus informing about chemosensitivity and treatment efficacy. MRD is highly prognostic in all ALL subtypes, including B- and T-cell lineages and in Philadelphia chromosome (Ph)-negative and Ph-positive disease. Across ALL subtypes, methods of MRD assessment, treatment regimens and other contexts, the detection of MRD after initial treatment nearly universally correlates with poorer relapse-free survival (RFS) and overall survival (OS).^5,6 Besides its prognostic importance, knowledge of MRD status can influence treatment strategies, such as informing the switch to blinatumomab, with or without subsequent consolidative allogeneic stem cell transplantation (alloSCT).⁷ Herein, we review the evaluation, prognostic impact, and management of MRD in adult patients with ALL.

Evaluation of MRD

Several different laboratory methods exist that are capable of detecting and quantifying MRD in ALL.^8,9 The most commonly used techniques are MFC by detection of immunophenotypic aberrancy on leukemia blasts and real-time quantitative polymerase chain reaction (RQ-PCR) by the analysis of rearranged immunoglobulin (IG) or T-cell receptor (TCR) genes or of recurrent gene fusions (e.g. BCR-ABL1).¹⁰ Even more sensitive and accurate techniques have also been recently introduced, including 10-color flow cytometry, droplet digital PCR, and high-throughput next-generation sequencing (NGS).¹¹ The main advantages and disadvantages of the commonly used methods of MRD assessment are summarized in Table 1.

Table 1.

Methods of measurable residual disease assessment.

Method	Advantages	Limitations	Sensitivity
Multiparameter flow cytometry • Comparison of baseline and remission LAIPs • DfN method	- Rapid - Sensitive - Relatively inexpensive - Ability to quantify antigen expression for targeted agents - Does not require access to pretreatment specimen (DfN method only)	- Lack of standardization - Need for significant technical expertise - Requires fresh cells - Risk of immunophenotypic shifts and false-negative results - Difficulty to differentiate malignant lymphoblasts from hematogones	10–4 (0.01%)
Quantitative PCR for IG/TCR rearrangements	- Sensitive - Standard guidelines for application and interpretation (Euro-MRD)	- Time-consuming - Labor intensive - Requires pretreatment sample - Expensive - May not be accurate for early T-cell precursor ALL	10–4 to 10–5 (0.01–0.001%)
Quantitative PCR for gene fusions (e.g. BCR-ABL1)	- Sensitive - Simple (uses same standard primers as used for diagnosis)	- Applicable to < 50% of ALL cases	10–4 to 10–5 (0.01–0.001%)
NGS	- Ultrasensitive - Fast - Can detect multiple clones and track clonal evolution - Only US FDA-approved assay (ClonoSEQ)	- Lack of standardization - Requires pretreatment sample - Expensive - Minimal clinical validation	10–6

ALL, acute lymphoblastic leukemia; DfN, different from normal; FDA, Food and Drug Administration; IG, immunoglobulin; LAIP, leukemia-associated immunophenotypes; NGS, next-generation sequencing; PCR, polymerase chain reaction; TCR, T-cell receptor.

Multiparametric flow cytometry

MFC is a fast and relatively inexpensive procedure that is broadly applicable to nearly all ALL cases. Standard MFC assays can detect residual leukemia cells in 1 out of 10,000 cells (10^–4 or 0.01%) by relying on assessment of aberrant leukemia-associated immunophenotypes (LAIPs), which constitute either aberrant expression of myeloid antigens or increased or decreased density of antigens normally expressed on benign B-cell precursors. A higher sensitivity of 10^–5 can be reached with the use of some ⩾8-color flow cytometry assays, which can detect abnormal population at very low levels (⩽0.0002%) when a sufficient number of cells are acquired (2 to 5 million nucleated cells).^12–14 The precision in detecting LAIPs can be improved by using different combinations of fluorochromes that can simultaneously recognize multiple antigens for better sensitivity. A better combination of markers may contribute to improved flow cytometric MRD detection in patients with B-cell ALL, particularly by differentiating leukemia cells from regenerative blasts. For example, one group has shown that a single eight-color tube consisting of CD9, CD10, CD19, CD20, CD34, CD38, CD45, and CD58 could provide as much diagnostic utility as compared with a previously used three-tube panel with 12 markers.¹⁵

By one MFC-based method, all LAIPs detected on leukemic blasts at time of diagnosis are assessed over the course of therapy and constitute MRD when still detectable in the remission sample. Another approach called the ‘different-from-normal’ (DfN) method relies on the difference in immunophenotypes present in the remission sample as compared with a highly stereotypical normal immunophenotype distribution.¹⁶ This method has the advantage of not necessarily requiring an initial diagnostic sample and can also assess MRD regardless of any immunophenotypic changes that occur over the course of therapy.^10,16 However, there may be less certainty with this approach if information about diagnostic LAIPs is not available, and therefore even in laboratories were the DfN approach is used, baseline LAIPs are generally also used for comparison (when available). A combined ‘LAIP-based DfN’ approach has been advocated by some groups to evaluate MRD in acute myeloid leukemia, and this approach is also used in some laboratories when assessing ALL MRD.¹⁷ Regardless of the specific method used, a major downside of MFC for MRD assessment is the challenge of standardization across laboratories and pathologists. Furthermore, significant expertise and knowledge of antigen expression patterns seen during differentiation and maturation of normal hematopoietic progenitors in both resting and regeneration states is needed to properly analyze the resultant data, particularly when the DfN method is used. Conversely, when MFC is used to compare LAIPs between diagnostic and remission samples, immunophenotypic shift that may occur as the result of therapy can also decrease the accuracy of this approach, potentially leading to false-negative results by MFC.

Real-time quantitative polymerase chain reaction

RQ-PCR is another standard method that is used for MRD detection and quantification. MRD targets for PCR involve rearranged immunoglobulin IG or TCR genes in Ph-negative B-cell ALL and T-cell ALL, while in Ph-positive ALL, BCR-ABL1 mRNA transcripts are the preferred MRD marker. Other gene fusions involving MLL or CRLF2 may also be used as targets in other subtypes of ALL, although there are few clinical data currently available to support their utility as reliable MRD markers. For patients with Ph-negative B-cell ALL or T-cell ALL, several studies have demonstrated a high concordance rate between MFC and PCR-based assays.^18–20 The choice between these two methods therefore largely depends on the level of expertise and availability in different laboratories.^18,19,21 MFC is widely used in hospitals and centers in the United States (US), as standardized allele specific oligonucleotide (ASO) PCR is generally not available. In contrast, there have been intense efforts to standardize ASO-based RQ-PCR in European countries, where the MRD assay is commonly used.⁸

In Ph-negative B-cell ALL and T-cell ALL, RQ-PCR analyzes unique sequences of the junctional regions of rearranged IG or TCR genes for which ASOs are specifically designed for each patient. Primers identified at diagnosis are then applied to subsequent post-therapy samples in order to quantify MRD.²² This approach can be applied to 90–95% of patients with ALL.⁸ In Europe, this process is standardized by international collaboration by the Euro-MRD group; however, there is no such standardization in the US, and therefore ASO-PCR is not used in clinical practice. Despite higher sensitivity compared with MFC (down to 10^–5), ASO-PCR is a time-consuming procedure, costly, and highly complex, requiring extensive knowledge and experience. Moreover, in early precursor T-ALL, it is difficult to monitor MRD by ASO-PCR, because the lymphoblasts are immature and often have not undergone TCR rearrangement.²³

In Ph-positive ALL, the BCR-ABL1 gene translocation is a reliable PCR target. Using reverse transcriptase PCR (RT–PCR), MRD is followed by quantification of BCR-ABL1 mRNA transcripts with the same standard probes used for diagnostic purposes in Ph-positive leukemia.²⁴ This technique is simple, rapid, and broadly applicable. Droplet digital PCR is a relatively new technique that may have utility in Ph-positive ALL, with some early studies suggesting that it may be more sensitive than standard RQ-PCR.^25,26

Next-generation sequencing

High-throughput NGS is a novel method in MRD detection in ALL that can overcome some of the limitations of standard methods. The targets are the same leukemia-specific rearranged IG and TCR genes analyzed by ASO-PCR. However, NGS has the capability of simultaneously amplifying multiple combinations of rearranged IG and TCR genes by multiplex PCR without the need of patient-specific probes. It can therefore identify and quantify multiple clones and subclones that can be tracked over the course of therapy, although the clinical utility of this theoretical advantage has yet to be robustly proven.^27,28 Another advantage of NGS is the achievement of very high levels of sensitivity based on dilution experiments, detecting as few as 1 leukemic cell in 1,000,000 nucleated cells (i.e. sensitivity of 10^–6), although only a few patients actually had MRD detectable at the 10^–6 level in these studies.²⁹ NGS is relatively rapid (around 1 week for one sample) and reliable, with high concordance with standard MFC or PCR techniques.^29–31 Despite the higher sensitivity of NGS, the prognostic significance of MRD at very low levels is unclear. Whether these very low levels of MRD should prompt any changes in therapeutic decision is largely unknown, and to date, only a few relatively small clinical studies of NGS-based MRD in ALL have been published.^30,32,33 However, given the high sensitivity of this approach, the clonoSEQ NGS technology (Adaptive Biotechnologies, Seattle, WA, USA) was recently the first MRD assay to be approved by the US Food and Drug Administration (FDA).³⁴

Prognostic impact of MRD

While historically ALL was risk-stratified using baseline characteristics such as white blood cell count, immunophenotype, and cytogenetics, MRD information outweighs many of these traditional prognostic factors, and is often the strongest independent predictor of outcomes.^4,35–42 A meta-analysis involving 13,637 children and adults demonstrated the benefit of MRD negativity across disease subtypes (e.g. Ph-negative and Ph-positive, B-lineage and T-lineage), therapies, methods, timing of MRD assessment, and MRD cut-offs. In adults, the 10-year event-free survival (EFS) for patients who achieved MRD negativity was 64% compared with 21% for those with detectable MRD [hazard ratio (HR), 0.28; 95% confidence interval (CI): 0.24–0.33]. A significant OS benefit to achieving MRD negativity was also observed in children (HR, 0.28; 95% CI: 0.19–0.41) and adults (HR, 0.28; 95% CI: 0.20–0.39).⁶ A subsequent meta-analysis of 23 published articles reporting on MRD in adults with B-cell ALL confirmed an overall improvement in both RFS and OS with random effects HRs of 2.44 (95% CI: 1.91–2.86) and 2.19 (95% CI: 1.63–2.94), respectively, for patients achieving MRD negativity.⁵

Different levels of detectable MRD may also have prognostic value, and risk of relapse is proportional to the quantity of MRD in several studies. For example, in one study, patients with lower detectable MRD (between 10⁻⁴ and 10⁻³) by either ASO-PCR or MFC had significantly longer duration of remission, RFS and OS than those with very high MRD (⩾10⁻¹).⁴³ While the consensus for what constitutes clinically relevant MRD response is generally defined as the achievement of a level below 10^–4, patients with detectable MRD have variable outcomes based on the quantity of MRD, with the best outcomes seen with early achievement of absence of any detectable residual disease.⁷

More recent reports have also suggested that coupling MRD information with different ALL molecular subtypes may improve prediction of relapse. In patients with Ph-negative ALL, high-risk genetics are independently associated with poorer outcomes and a high rate of MRD persistence after initial therapy.⁴⁴ In one study, the presence of IZKF1 gene deletion or MLL gene rearrangement was associated with increased risk of relapse in B-cell ALL and a genetic profile defined as the absence of NOTCH1/FBXW7 mutation or the presence of NRAS/KRAS mutation or PTEN alteration was also associated with worse outcomes in T-cell ALL.³⁹ Both molecular–genetic features and MRD status were independently associated with relapse and survival, suggesting that both should be incorporated into risk stratification. Other studies have similarly showed that cytogenetic features, such as complex karyotype (defined as ⩾5 chromosomal abnormalities) or low hypodiploidy/near triploidy, are associated with poor outcomes regardless of the MRD status.⁴⁵ Thus, while MRD negativity is desirable in all cases, it does not appear to override the negative prognostic impact of these adverse-risk genomic alterations. Future prospective studies are needed to determine how to fully incorporate genetic profiling, cytogenetics, and MRD status into risk stratification schemes that can inform therapeutic decision-making.

Ph-positive ALL

MRD is also highly prognostic in patients with Ph-positive ALL. In adults with Ph-positive ALL, detection of MRD measured by RT–PCR of BCR-ABL1 transcripts is associated with worse outcomes.^46–49 In one study, patients who received chemotherapy and tyrosine kinase inhibitor (TKI) and achieved a complete molecular response (CMR; defined as absence of a quantifiable BCR-ABL1 transcript by RT–PCR) after approximately 3 months of treatment had excellent long-term OS of 66% at 4 years in the absence of alloSCT.⁵⁰ Achievement of CMR was the only factor independently prognostic for OS. These data raise questions as to whether assessment of MRD can identify patients with Ph-positive ALL who do not require alloSCT in first remission. Integration of genomic features (e.g. IKZF1 or CDKN2A/B deletions) into this assessment may further improve our prognostication.^49,51

Pre- and post-transplant MRD

In patients undergoing alloSCT, both pre- and post-transplant MRD predict for higher risk of post-transplant relapse.^52–58 In one study, children achieving MRD negativity before alloSCT had better disease-free survival (DFS) and OS than those with persistent MRD (DFS of 83% versus 41%; OS of 92% versus 64%, respectively, p < 0.0001 for both).⁵³ In another prospective study of children with relapsed ALL, pre-alloSCT MRD was also prognostic in this context.⁵² Similarly, in adults, pre-transplant MRD at a level ⩾ 10^–4 as measured by NGS was predictive of post-transplant relapse (HR 7.7, 95% CI: 2.0–30, p = 0.003).⁵⁵ Conversely, MRD reappearance after initial chemotherapy or alloSCT is also a sign of impending leukemia relapse.^{54,55,59–61} After chemotherapy, 60–80% of patients with MRD recurrence experience hematologic relapse after a median of 3 months.^59,62 Among patients who received alloSCT in the NILG study, those with detectable MRD post-transplant (day +100) had a relapse risk of 80% compared with only 7% to those with undetectable MRD (p = 0.0006).⁵⁴

Relapsed or refractory ALL

While there is evidence that MRD is highly prognostic in newly diagnosed ALL, its impact and how this information should guide therapy is less clear in the relapse/refractory (R/R) setting. In studies evaluating single novel agents (e.g. blinatumomab and inotuzumab ozogamicin) in R/R ALL, achievement of MRD responses was associated with lower rates of relapse.^63,64 In a single-arm phase II study of 36 patients with R/R B-cell ALL treated with blinatumomab, 69% of patients achieved MRD response defined as MRD < 10^–4 by ASO-PCR, which was associated with a 67% reduction in relapse risk.⁶³ In the INO-VATE trial comparing inotuzumab ozogamicin versus combination chemotherapy for patients with R/R B-cell ALL, MRD negativity (defined as < 10^–4 by MFC) was achieved in 63% of patients in the inotuzumab ozogamicin arm, and MRD response was associated with prolongation of both progression-free survival and OS compared with MRD nonresponders (median progression-free survival: 8.6 versus 5.4 months, p < 0.0001; median OS: 14.1 versus 7.2 months, p = 0.009).⁶⁴ Some data also suggest that the significance of MRD negativity may be more pronounced in first salvage than in later salvages. In a study involving 130 patients with R/R ALL, it was demonstrated that MRD negativity by MFC at the time of best response was associated with significant better EFS for patients treated in first salvage (median 18 versus 7 months, p = 0.06), but not in second salvage and beyond.⁶⁵ Patients who achieved MRD negativity after their first salvage and subsequently underwent alloSCT had the best outcomes, with a 2-year OS rate of 65%.⁶⁵

Management of MRD

MRD status is increasingly used not only for risk classification and prediction, but also for post-remission treatment decision-making.⁷ By tailoring treatment strategies based on MRD status, patients at higher risk of relapse may receive risk-adapted, novel therapies such as the CD3-CD19 bispecific T-cell engager blinatumomab or the anti-CD22 antibody–drug conjugate inotuzumab ozogamicin, with or without subsequent alloSCT. Conversely, patients at lower risk of relapse (e.g. those without baseline adverse-risk genomic features who rapidly achieve MRD negativity with standard therapy) may benefit from treatment de-escalation and not undergoing alloSCT in first remission, thereby potentially sparing them from unnecessary treatment-related toxicities.^66,67

Allogeneic stem cell transplantation

Several studies have suggested that alloSCT in first CR is associated with lower risk of relapse and longer survival in patients with ALL who achieve a suboptimal MRD response.^68,69 In the German Multicenter ALL Study Group (GMALL 07/03 trial), patients with persistent MRD (⩾ 10^–4) measured by RQ-PCR of leukemia-specific IG and TCR gene rearrangements after first consolidation (week 16) were considered high risk for relapse and were offered alloSCT. Overall, 47% of patients with MRD persistence received alloSCT in first CR (alloSCT rates: 71% in the high-risk group and 39% in the standard-risk group). The 5-year continuous CR was significantly higher for patients who received alloSCT in first CR compared with those with chemotherapy alone (66% versus 12%, p < 0.0001), which also translated into better 5-year OS (54% versus 33%, respectively, p = 0.06).⁴ In contrast, patients who achieved MRD negativity at week 16 had 5-year continuous CR and 5-year OS rates of 74% and 81%, respectively, in the absence of alloSCT. In the PETHEMA ALL-AR-03 prospective trial, adolescents or adults with high-risk Ph-negative ALL based on at least one high-risk disease feature (i.e. age between 30 and 60 years, white blood cells > 30×10⁹/l, or t(4;11) or other MLL rearrangements) were assigned to post-remission therapies based on early cytologic response (<10% blasts in bone marrow at day 14 of induction) and MRD status. Patient with favorable cytologic and MRD response continued to receive chemotherapy alone (n = 108) and those with poor cytologic response or suboptimal MRD response (n = 71) were assigned to receive alloSCT. The 5-year DFS and OS were 32% and 37%, respectively, for patients assigned to alloSCT, and 55% and 59% for those assigned to chemotherapy.³⁵ Together, these studies suggest that MRD assessment after initial chemotherapy can be used to identify patients most likely to benefit from alloSCT in first remission, even among patients who appear otherwise high-risk based on pretreatment characteristics. They also highlight the relatively poor outcomes for patients with persistent MRD positivity, even when alloSCT is performed.

Nontransplant MRD-directed therapies

Novel monoclonal antibodies such as blinatumomab or inotuzumab ozogamicin are capable of inducing remissions in R/R B-cell ALL.^70–72 Blinatumomab is generally more effective in patients with lower burden of disease, making a particularly promising agent for the treatment of MRD.^73,74 In the open-label, single-arm phase II BLAST study, adult patients with B-cell ALL in CR but with persistent or recurrent MRD at a level of ⩾10^–3 after intensive chemotherapy received up to four cycles of blinatumomab. Among 116 patients, 78% achieved complete MRD response after the first cycle. Despite the inclusion of higher-risk patients (35% of patients in second or later remission and 47% with MRD levels ⩾10^–2), the 18-month RFS rate was 54% and the median OS was 36.5 months. Patients who achieved complete MRD response had significantly longer RFS (23.6 months versus 5.7 months, p = 0.002) and OS (38.9 months versus 12.5 months, p = 0.002) compared with MRD nonresponders.⁷⁵ Based on these results, blinatumomab was approved by the US FDA in March 2018 for the treatment of patients with B-cell ALL in CR but with detectable MRD at a level of ⩾ 0.1%.⁷⁶

Whether alloSCT should be routinely offered to patients who achieve MRD negativity after blinatumomab is an open question. Interestingly, in the BLAST study, 33% of patients who achieved MRD negativity did not receive any additional treatment after blinatumomab, and 25% of them remained in continuous CR after a median follow up of 24 months (range, 2.8–41.6 months), suggesting that a proportion of patients with MRD-positive disease who respond to MRD-directed therapy can achieve prolonged remission duration, or possibly cure, without the need of alloSCT.⁷⁵ In a post hoc analysis, there was no statistical difference in OS between transplanted and nontransplanted patients (odds ratio = 1.83; 95% CI: 0.69–4.9, p = 0.24), in part because 27% of transplanted patients died from transplant-related mortality. Overall, the balance of evidence suggests that proceeding with alloSCT after blinatumomab for MRD-positive disease is associated with decreased risk of relapse but increased risk of transplant-related mortality, with similar long-term survival whether or not subsequent alloSCT is performed.

Inotuzumab ozogamicin was also associated with higher MRD response rates and improved OS compared with conventional chemotherapy in a randomized phase III study of patients with R/R B-cell ALL.^64,72 For patients with morphological relapse, inotuzumab ozogamicin appears to be more effective than blinatumomab in inducing remissions (CR rates with or without full hematologic recovery: 81% versus 44% comparing across two randomized phase III trials).^70,71 However, the role of inotuzumab ozogamicin in eradicating MRD in patients with MRD-positive remission is not currently known. An ongoing clinical trial using inotuzumab ozogamicin for patients with B-cell ALL and persistent or recurrent MRD is currently enrolling (ClinicalTrials.gov identifier: NCT03441061).

Other immunotherapeutic strategies include the use of CD19-directed chimeric antigen receptor (CAR) T cells for MRD eradication. In a phase I trial, 53 adult patients with R/R B-cell ALL received autologous CD19 CAR T cells and achieved a CR rate of 83%. Patients with low disease burden (defined as < 5% bone marrow blasts) had significantly longer EFS and OS compared with patients with higher disease burden (defined as ⩾5% bone marrow blasts or presence of extramedullary disease; median EFS: 10.6 versus 5.3 months, p = 0.01; median OS: 20.1 versus 12.4 months, p = 0.02, respectively).⁷⁷ These findings suggest that CAR T cells may play a particularly important role in the management of MRD-positive disease, where such therapy may be curative for a subset of patients.

Frontline approaches to eradicating MRD

Ultimately, eradicating MRD with frontline therapy is likely to lead to the best outcomes. Incorporation of novel monoclonal agents into the frontline chemotherapy is a very exciting strategy under investigation, with the aim of increasing initial MRD responses and reducing chemotherapy-related toxicities and the need for alloSCT. This strategy was first evaluated in older patients where the use of intensive chemotherapy was historically associated with very poor outcomes, primarily due to treatment-related toxicities, including infections from myelosuppression.⁷⁸ In one large retrospective study of older patients (above 60 years) receiving hyper-CVAD chemotherapy without modification, the induction mortality rate was 10% and the death rate in CR was 34%, leading to a 5-year OS rate of only 20%.⁷⁹ In an attempt to reduce toxicity and improve survival, a phase II trial evaluated the combination of reduced intensity chemotherapy (the mini-hyper-CVD regimen) with inotuzumab ozogamicin in older patients with newly diagnosed Ph-negative B-cell ALL. The overall response rate was 98% and the rate of MRD negativity by MFC after one cycle of chemotherapy was 80%, with no early treatment-related deaths. The 3-year OS rate was 54%, which compares very favorably with historical outcomes with both intensive and low-intensity regimens.^80,81 Preliminary analysis of a phase II trial of older adults with ALL receiving frontline blinatumomab followed by POMP maintenance reported an overall response rate of 66%, with MRD negativity achieved in 92% of responders. The estimated 1-year OS was 65%, with no reported treatment-related mortality.⁸² In younger adults, an ongoing study is investigating the addition of blinatumomab after four cycles of intensive chemotherapy with hyper-CVAD in the frontline setting. Preliminary results showed very high response rate (100%) with MRD negativity in 93%, and a 1-year OS of 90% with acceptable toxicity profile.⁸³

Ph-positive ALL

In Ph-positive ALL, the achievement of CMR is highly predictive of longer survival.⁵⁰ When combined with intensive chemotherapy, the use of the third-generation pan-BCR-ABL TKI ponatinib achieves a CMR rate of 78%, translating into better OS compared with regimens using earlier-generation TKIs such as imatinib or dasatinib. In a recent update of a phase II trial using the hyper-CVAD regimen with ponatinib for adults with newly diagnosed Ph-positive ALL, the 3-year OS was 76%, with only three patients relapsing while on ponatinib.⁸⁴ Using a propensity score analysis comparing results from two phase II trials, hyper-CVAD plus ponatinib was associated with a superior MRD response rate, RFS and OS compared with hyper-CVAD plus dasatinib.⁸⁵ The benefit of ponatinib in this setting is likely driven by the higher rate of CMR (78%) compared with that achieved with other TKIs (30–50%), which has been shown to strongly correlate with survival.^86,87 In patients undergoing alloSCT for Ph-positive ALL, the use of post-transplant TKI has also been recommended to reduce risk of relapse post-transplant, particularly in patients with detectable MRD.⁸⁸ Blinatumomab is also effective in Ph-positive ALL in both the R/R setting and for MRD eradication.^75,89 Several ongoing studies are therefore evaluating chemotherapy-free frontline regimens with the combination of blinatumomab with TKIs (ponatinib, ClinicalTrials.gov identifier: NCT03263572; dasatinib, ClinicalTrials.gov identifiers: NCT02143414 and NCT02744768) with the goal of increasing MRD responses with acceptable toxicity, and decreasing the need for alloSCT in these patients. Re-emergence of MRD disease in Ph-positive ALL is indicative of impending relapse, mostly due to acquired ABL1 mutations, although these mutations may be difficult to detect a low levels of MRD using standard approaches.⁹⁰ The use of blinatumomab in addition to ponatinib in this setting may be reasonable in order to cover potential resistance mutations in the setting of MRD-only disease.

Conclusion

Assessment of MRD status is enormously important in the management of patients with ALL, not only in risk stratification, but also to inform subsequent treatment strategies. The development of more sensitive MRD assays, including NGS, may allow for even better risk stratification, although how very small amounts of residual leukemia detected at a level of <10^–4 should influence treatment (if at all) is largely unknown at the present time. With the availability of highly effective ALL therapies, particularly blinatumomab, inotuzumab ozogamicin and CD19 CAR T cells, these agents are likely to increasingly play a role in MRD eradication, both for patients with MRD-only disease and through their incorporation into frontline regimens in order to render patients MRD-negative early in their treatment. While long-term data are still eagerly awaited, this strategy holds promise in reducing the need for myelosuppressive chemotherapy and subsequent alloSCT for many patients. Chemotherapy-free regimens incorporating these active agents into the frontline setting may also be a possibility in the near future, particularly for older patients. Given its close association with better long-term outcomes, achievement of MRD negativity is already being used as a surrogate endpoint in several clinical trials, and this approach should ultimately allow for even more rapid approval of effective, novel regimens to patients with ALL.