Role of Chimeric Antigen Receptor T-Cell Therapy: Bridge to Transplantation or Stand-Alone Therapy in Pediatric Acute Lymphoblastic Leukemia

Abstract

Purpose of review:

To discuss the curative potential for chimeric antigen receptor T-cell (CAR-T) therapy, with or without consolidative hematopoietic stem cell transplantation (HCT) in the treatment of children and young adults with B lineage acute lymphoblastic leukemia (B-ALL).

Recent findings:

CAR-T targeting CD19 can induce durable remissions and prolong life in patients with relapsed/refractory B-ALL. Whether HCT is needed to consolidate remission and cure relapse/refractory B-ALL following a CD19 CAR-T induced remission remains controversial. Preliminary evidence suggests that consolidative HCT following CAR-T in HCT-naïve children improves leukemia-free survival. However, avoiding HCT-related late effects is a desirable goal, so identification of patients at high risk of relapse is needed to appropriately direct those patients to HCT when necessary, while avoiding HCT in others. High disease burden prior to CAR-T infusion, loss of B-cell aplasia and detection of measurable residual disease by flow cytometry or next-generation sequencing following CAR-T therapy associate with a higher relapse risk and may identify patients requiring consolidative HCT for relapse prevention.

Summary:

There is a pressing need to determine when CD19 CAR-T alone is likely to be curative and when a consolidative HCT will be required. We discuss the current state of knowledge and future directions.

Introduction:

Significant advances have been made in the management of B lineage acute lymphoblastic leukemia (B-ALL) and over 85–90% of children now survive without relapse following upfront chemotherapy.^1–3 The complete remission rates (CR) in adolescent and young adult (AYA) patients have also improved to 70–80% with introduction of pediatric-inspired regimens, although the 5-year overall survival (OS) remains only 61%.^2,4 Despite the successes of B-ALL therapy, relapse and refractory (r/r) disease remain major concerns.^5,6 In patients experiencing first relapse, the current 5-year OS is currently approximately 50%. ^3,7,8

Chimeric antigen receptor (CAR) T-cell therapy (CAR-T) targeting CD19 has impressive induction rates in r/r B-ALL. On the ELIANA registration study, which led to FDA approval of tisagenlecleucel for children and AYA with r/r B-ALL, 81% of patients achieved a CR without detectable measurable residual disease (MRD) within 3 months of infusion.⁹ Other experiences with CD19 CAR-T in children and AYA with r/r B-ALL have demonstrated similarly high CR rates.^9–14 Additionally, in many patients remissions are durable with CD19 CAR-T alone.⁹

While CD19 CAR-T can effectively induce remission and can lead to long-term durable remission in a fraction of pediatric patients with B-ALL, many other recipients will relapse. Although the ELIANA study demonstrated OS rates of 90% and 76% at 6 and 12 months, the corresponding event-free survival (EFS) at those time points was 73% and 50%.⁹ Other study results are consistent with the ELIANA trial and suggest that at least half of CD19 CAR-T recipients will ultimately relapse, most in the first year post-infusion.^12,15 Following post-CAR-T relapse, salvage options for these patients are limited.^16–19

HCT reduces relapse in high-risk B-ALL patients, but is associated with significant short-term and long-term risks, partially due to the total body irradiation (TBI) which also contributes to the HCT’s therapeutic effect in B-ALL.^20,21 Published, retrospective CAR-T studies have shown an inconsistent benefit of HCT as consolidation following CAR-T ^12,22–26 and may be impacted by a history of prior HCT. The balance of risk and potential benefit of HCT following CAR-T for a given patient depends on that individual’s relapse risk following CAR-T alone, determinants of which are now better understood.

At this time, it is unclear whether the CAR-T should generally be viewed as a definitive treatment or as a bridge to a consolidative HCT, or if the role of CAR-T will depend on individual patient characteristics. Currently, there are no surveillance plans that take into account dynamic changes in risk of relapse to implement preemptive relapse prevention strategies in patients who achieve a CR with CAR-T. Additionally, no studies have randomized patients to HCT versus observation following CAR-T induced remission.^27,28 In this review, we will discuss the current knowledge surrounding the necessity of post-CAR consolidative HCT.

Use of CAR-T as standalone therapy or ‘bridge-to-HCT’

The first pediatric patient to receive CD19 CAR-T for r/r B-ALL remains in an ongoing remission, now 9 years from her infusion without having received any additional therapy.²⁹ Her story, amongst other reports of prolonged CAR-T induced remissions provides clear proof-of-concept that CD19 CAR-T have curative potential. Durability of remission, however, is dependent on numerous factors—including functional persistence of CAR-T and the potential for leukemia immune escape due to loss or down-regulation of the target antigen.

CAR-T persistence is influenced by the properties of the CAR construct and the product composition, with inherent interpatient variability due to the need to generate personalized products from patients who have received prior therapy that may impact T-cell functionality and proliferative potential.³⁰ One key distinguishing feature between various CAR constructs is the co-stimulatory domain. In general, CD19/28ζ-based constructs, while equally effective in remission induction,^10,15,31 tend to be associated with shorter CAR-T persistence than the CD19/4-1BB CAR constructs.^11,32 As a result of reduced persistence, CD19/28ζ CAR-T in pediatric and young adult patients with B-ALL, are generally used for remission induction and as a bridge to HCT and in that context have shown a favorable OS and low relapse risk.^15,23 Conversely, experience with CD19/4-1BB CAR-T directly challenged the need for HCT, with substantial numbers of patients achieving long-term remission with CAR-T alone.^9,24,26,33

As antigen escape or downregulation represents one of the leading causes of relapse, the field is advancing immunotherapy that is not entirely dependent on CD19 as a target. Beyond CD19 CAR-T, CD22 directed CAR-T-cells are emerging as an important treatment modality, particularly in patients who relapse after CD19 targeted therapies. While CD22 CAR-T can be effective in inducing CR in patients with r/r B-ALL, including CD19 negative B-ALL, remission durability is limited due to rapid emergence of CD22 dim/negative disease, rendering CD22 CAR-T therapy as a bridge to HCT.¹³ However, having effective CAR-T for an alternative antigen target has also forged a path for the development of combinatorial treatment strategies, such as CD19 and CD22 dual targeting CAR-T.³⁴

While the current experience is based primarily on single infusions of single antigen targeted CD19 CAR-T and our discussion is informed by this experience, future iterations of CAR constructs will address the current barriers to durable remission.³⁵ Approaches such as CAR-T reinfusions for antigen-positive relapse,^18,19 dual-antigen targeting approaches to prevent antigen-escape,^36,37 and CAR-T boosting strategies,³⁸ are being actively studied and will continue to propel the field to optimize CAR-T as stand-alone therapy in B-ALL.

Use of HCT for post-CAR-T consolidation

While efforts are ongoing to optimize the use of CAR-T as a standalone therapy, HCT is currently the most reliable approach for relapse prevention. Indeed, prior to the introduction of CAR-T, the universal standard of care (SOC) for patients with high-risk, r/r ALL was to proceed to HCT following the achievement of a MRD negative CR.^39–41 HCT is still considered SOC for post-CAR-T remission consolidation in many centers.⁴² Experience remains limited (Table 1) and studies randomizing patients to HCT or observation post-CAR-T are lacking, however emerging data tends to support a role for HCT for post-CAR-T relapse prevention, particularly in patients who are HCT naive.

Table 1.

Outcomes of HSCT post CD19 CAR T-cell Therapy

Reference	Number of subjects	Age years. Median (range)	CAR costimulatory domain	Median time from CAR to HCT days (range)	EFS HCT vs No HCT at 2 yrs	Post-HSCT Relapse & TRM in HCT arm
Frey43	HCT N=9 No HCT N=15	HCT 39 (24–50) No HCT 36 (26–63)	4-1BB	78 (51–162)	HCT ∼85% No HCT ∼35%	Not specified
Hay17	HCT N=18	HCT 35 (22–73)	4-1BB	70 (44–138)	HCT 61%	Relapse 17% TRM 23%
Jiang44	HCT N=21 No HCT N=26	Not specified	4-1BB	Not specified	HCT ∼70% No HCT ∼30%	Not specified
Summers45	HCT N=23 No HCT N=27	HCT 15 (1–25) No HCT 12 (1–22)	4-1BB	Approximately 3 months	HCT 75% No HCT 40%	Relapse 5/23 (22%) TRM 1/23 (4.3%)
Zhang25	HCT N=75 No HCT N=27	(all) 12 (2–61)	CD28 and 4-1BB	63 (63–120)	HCT 76.9% No HCT 11.6%	Relapse 10/75 (13.3%) TRM 3/75 (4%)
Shah24	HCT N=21 No HCT N=7	(all) 13.5 (4.3–30.4)	CD28	54 (42–97)	HCT 62% No HCT 0%	Relapse 9.5% TRM 6/21 (29%)
Park12	HCT N=17 no HCT N=16	(all) 44 (23–74)	CD28	74 (44–312)	HCT ∼30% No HCT ∼30%	Relapse 6/17 (35%) TRM 6/17 (35%)
Curran15	HCT N=15 no HCT N=3	(all) 13.5 (1–22.5)	CD28	57 (30–200)	HCT 8/15 No HCT 0/3	Relapse 4/15 (26.7%) TRM 3/15 (20%)

Although one retrospective analysis of adult subjects treated with CD19/28ζ CAR-T showed no LFS or OS benefit to HCT with relatively poor EFS, relapse and non-relapse mortality (NRM) rates post-HCT (approximately 30%, 35% and 35% respectively),¹² three other reports in adult CD19 CAR-T recipients demonstrated an improved EFS with consolidative HCT, with 2-year EFS of approximately 61–85%.^17,43,44 Moreover, an analysis of 28 pediatric and young subjects who achieved an MRD negative CR following CD19/28ζ CAR-T, demonstrated improved LFS for the 21 patients who received consolidative HCT with a 5-year EFS of 61.9% and a cumulative incidence of relapse of 4.8%. All seven who did not undergo HCT relapsed.²³ Similarly, among 50 pediatric and young adult subjects treated with a CD19/4-1BB CAR-T who survived in remission until day 63 post CAR-T infusion, LFS was improved by HCT .²⁶ A recent meta-analysis incorporating eighteen CD19 CAR-T studies with a total of 758 adult and pediatric patients concluded that post-CAR-T HCT associated with a lower relapse rate.²²

While consensus is developing towards a probable benefit of a first HCT in HCT-naïve CAR-T recipients, outcomes following second HCT are generally not as good,^45,46 making the potential benefit of a second HCT following CAR-T less clear. In one study, for those receiving their first HCT post CD19 CAR-T (n=13) the 2-year LFS was approximately 86%. In contrast, for 10 patients who received a second HCT following a post-CAR-T induced remission for post-HCT relapse, the 2-year LFS was approximately 50%. Given the importance of TBI containing myeloablative conditioning in HCT for ALL,²¹ the inability to deliver intensive, high-dose TBI-based conditioning for a second HCT may be one factor contributing to the inferior outcomes. The typically higher NRM after second HCT may also increase the risk of second HCT following CAR-T and further evaluation is required.

Along these lines, separating out data on the role of HCT in children and young adults from older adults may serve to be critical, particularly given the need for intensive myeloablative HCT conditioning in ALL. As NRM is increased in older patients with myeloablative condition, and TBI-based conditioning is superior in HSCT for ALL, HCT outcomes may not be comparable. This may explain the discrepant results for the benefit of consolidative HCT observed between the adult and pediatric CD19/28ζ studies utilizing the same construct.^12,15

Risk-factors predictive of leukemia recurrence post CAR-T

With CD19/4-1BB CAR-T, where a desirable goal is to achieve and maintain a durable long-term remission without HCT, development of a standardized approach to risk-assessment and mitigation for post-CAR-T relapse is urgently needed. Risk factors that appear to be predictive of relapse include early loss of B-cell aplasia (BCA), detection of MRD by flow-cytometry or next-generation sequencing (NGS), disease burden prior to CAR-T infusion, prior treatment history and possibly leukemia cytogenetics. (Table 2)

Table 2 -.

Risk factors for post-CAR relapse and consolidative HCT considerations

Risk Factor	Considerations
Loss of B-cell aplasia (BCA)	• Shorter BCA has been associated with shorter EFS/OS. • Represents loss of functional CAR-T persistence/surveillance and predicts for future relapse • Early loss of BCA may serve as a trigger for consideration of proceeding to HCT.
Next-generation sequencing (NGS) measurable residual disease (MRD)	• Emerging data demonstrates the importance of NGS-MRD monitoring for early signs of disease recurrence. NGS-MRD is more sensitive than flow-cytometric MRD monitoring and does not rely on presence of surface antigen for disease detection. • Need to establish the identifying sequence in advance of achieving remission to facilitate tracking of residual/recurrent disease. • Detection of any NGS-MRD positivity is highly predictive of overt relapse and necessitates consideration of HCT.
Disease burden	• High-disease burden (≥5% marrow blasts) pre CAR-T is associated with worse EFS and OS post CAR-T. • High-disease burden associates with increased likelihood of CD19 negative relapse, implying that CD19 CAR T-cell reinfusion strategies may not be available for salvage after relapse. • Given the risk of poor outcomes after CAR-T in patients with high disease burden, HCT needs to be considered to improve durable remission. Further study is required.
Cytogenetics	• Patients with specific cytogenetics (e.g., KMT2a-r) may be predisposed to antigen escape and lineage switch thus HCT should be considered for relapse prevention following CAR-T.
Prior therapy	• Receipt of CD19 targeted therapy (e.g., blinatumomab) prior to CD19 CAR T-cells may predispose to antigen escape. Studies are ongoing. •For patients with prior HCT, a second HCT may be less effective and associated with higher toxicity.

Abbreviations: BCA: B-cell aplasia; EFS: event-free survival; OS: overall survival; CAR: chimeric antigen receptor; HCT: hematopoietic cell transplantation

Loss of B-cell aplasia

CD19 CAR-T deplete normal as well as malignant B-cells and therefore BCA, defined as <1% CD19⁺ cells among total lymphocytes, can be used as a measure of in vivo CD19 CAR-T functional activity. Conversely, lack of BCA indicates inadequacy of CD19 CAR-T functional activity and early loss of BCA predicts relapse after CAR-T.^47,48 An analysis of BCA was conducted in participants in two studies testing tisagenlecleucel in B-ALL (ELIANA, (NCT02435849) and ENSIGN, (NCT02228096) trials).⁴⁹ Specifically, a landmark analysis was performed to assess the outcome of patients with continued BCA compared to those who recovered B-cells between 1–3 months, 3–6 months, 6–9-months, and 9–12 months after tisagenlecleucel. The analysis showed that patients with loss of BCA during the first year had a significantly higher risk of relapse, compared to those without B-cell recovery (HR=4.5). Notably, there was an association between EFS and duration of BCA. For patients with persistence of BCA at 3, 6, 9, and 12 months the corresponding 2 years EFS were 63%, 72%, 83% and 88%, respectively. Relapses occurring with persistence of BCA were invariably CD19-negative. The experience at Seattle Children’s similarly demonstrated that consolidative HCT improved outcomes in patients with short BCA (<63 days), even in those with a history of prior HCT.²⁶

Next-generation sequencing (NGS) Measurable Residual Disease

Sensitive detection of MRD in patients with leukemia in remission (defined as <5% bone marrow blasts) plays an important role in the management of pediatric ALL. In North America, the most common technology for MRD detection is currently multiparametric flow cytometry (MFC). The Children’s Oncology Group (COG) has used a standardized MFC assay, which detects disease reliably in marrow or peripheral blood at levels greater than or equal to 0.01% of mononuclear cells.⁵⁰ Monitoring MRD by MFC to determine response to initial induction and consolidation chemotherapy and subsequent treatment allocation is now standard in pediatric ALL.⁵⁰ MRD by MFC is also used to evaluate remission depth prior to HCT and detection of MRD by MFC in children with ALL immediately pre-HCT is associated with post-HCT relapse and inferior survival.^51,52

A newer more sensitive method for MRD detection builds on the premise that clonal blast populations can be defined by specific IgH or TCR rearrangements. NGS MRD can reliably detect the presence of blasts with a sensitivity exceeding 1 in 10⁶ cells.^53,54 NGS MRD pre-HCT carries an increased risk for post-HCT ALL relapse and NGS MRD positivity post-HCT is even more prognostic of ultimate overt relapse.⁵⁵ An additional benefit of NGS MRD testing is that it reliably detects CD19 negative B-ALL as well as CD19 positive B-ALL.

Based on the hypothesis that MRD measurements at various timepoints after achieving remission with tisagenlecleucel could identify patients at high or low risk of relapse after CAR-T, an analysis of MRD and BCA was conducted using blood and marrow samples from patients enrolled on one of two studies testing tisagenlecleucel in B-ALL (ELIANA, (NCT02435849) and ENSIGN, (NCT02228096) trials).⁴⁹ A total of 474 samples from 109 patients were analyzed for NGS-MRD, and 387 were informative for NGS and MFC. Multivariate analysis at day 28 after infusion showed independent associations of NGS MRD>0 (HR 4.9) and B-cell recovery (HR 3.3) with relapse. By 3 months, the NGS-MRD HR increased to 12, while B-cell recovery was not independently significantly predictive (HR 1.27). Detectable NGS-MRD reliably predicted risk with sufficient time to implement management of relapse prevention such as HCT or second CAR- T infusion.⁴⁹

Disease burden

High disease burden, which has generally been classified as ≥ 5% blasts in the marrow, prior to CD19 CAR-T is emerging as a risk-factor for non-response or B-ALL relapse after CAR-T.^23,33,47 High-disease burden is also associated with risk of CD19 negative relapse in particular, which further limits therapeutic salvage strategies.⁵⁶ Additional studies will be needed to clarify the role of disease burden as a formal risk factor for poor CAR-T outcomes.

KMT2a

In some studies, the subgroup of pediatric B-ALL patients who have a KMT2a gene rearrangement (KMT2a-r) in their leukemia were more likely to relapse after CD19 CAR-T. In particular, the KMT2a-r B-ALL patients appeared more likely to have a CD19 negative recurrence and to relapse with a myeloid phenotype.^57–59 Although subsequent studies have not reproduced these initial findings there remains concern about the durability of CAR-T efficacy for KMT2a-r B-ALL. Furthermore, recent data confirms a beneficial role of HCT in adult patients with KTM2a-r B-ALL,³⁹ further supporting the notion of HCT. However, as KMT2a-r B-ALL is more common in very young children, who are particularly vulnerable to HCT late effects, there is also a strong desire to avoid HCT in this context. Strategies incorporating frequent NGS MRD monitoring post CAR-T with a plan to recommend HCT if and when the risk of relapse becomes excessive may be of particular benefit to young children.

Prior CD19-targeting therapy

CD19-targeting agents, including bi-specific T-cell engaging monoclonal antibodies, such as blinatumomab, have now become common components of the management of pediatric B-ALL.² As blinatumomab has been associated with CD19 loss/down-regulation on the B-ALL cell surface, there have been concerns that its administration prior to CD19 CAR-T infusions may diminish the depth and, or duration of the CAR-T therapeutic effect. In a recent study, four of five patients who had received blinatumomab prior to receiving a CD28 co-stimulatory CD19 CAR-T were CAR-T non-responders.²³ In another study of children receiving CD19/4-1BB CAR-T, patients with prior blinatumomab had a trend towards lower CD19 expression and had a significantly higher rate of CD19 negative relapse or MRD persistence when compared with patients without prior blinatumomab exposure, although the cohort only included 19 patients (12%) that were classified as having CD19 dim expression.⁶⁰ A larger study of 420 children and young adults who received CD19 CAR-T showed a similar association of worse outcomes for patients receiving prior blinatumomab.⁶¹

Future directions

In the present era, we have at our disposal highly effective and novel immunotherapeutic approaches for remission induction. With an established potential for durable remission using CAR-T alone, selectively incorporating HCT following CAR-T induced remission for those at risk of post-CAR relapse should serve to optimize the curative potential of CAR-T while sparing exposure to intensive HCT if not needed. With availability of ultra-sensitive tools to detect very low levels of MRD, a strategy incorporating risk-based stratification will allow us to leverage the full curative potential of CAR-T therapy. Patients who receive CAR-T with limited persistence or CAR-T targeting molecules with a high-potential for down-regulation such as CD22, are likely to continue to be routinely referred to HCT for relapse prevention, whilst those who receive CD19 CAR-T with a 4-1BB costimulatory domain, including tisagenlecleucel, are likely to benefit from a risk-based approach. Risk stratification is the foundation upon which pediatric and young adult ALL therapy has been built and establishing a fine-tuned systematic approach incorporating individualized considerations for post-CAR-T HCT is needed to improve survival and limit long-term complications.

Key points.

• A substantial fraction of patients may achieve prolonged survival with CD19 CAR-T alone, however, approximately 50% of patients will experience relapse following CAR-T therapy.

• The curative potential of CD19 CAR-T may be optimized by appropriate utilization of consolidative HCT for relapse prevention in patients at high-risk of post-CD19 CAR-T relapse.

• Loss of B-cell aplasia, detection of post-CAR-T minimal residual disease (MRD) and high-disease burden prior to CAR-T infusion are risk factors for relapse following CD19 CAR-T therapy

• Patient specific risk factors, such as underlying leukemic cytogenetics and prior therapy, may also impact post-CD19 CAR-T outcomes and further study is warranted.

• A risk-based stratification is needed to help identify high-risk patients for whom consolidative HCT is likely to improve overall survival and to distinguish low-risk patients for whom HCT can be avoided.