Introduction

Over half (~ 3,000 cases) of acute lymphoblastic leukemia (ALL) diagnoses occur in patients ≥ 18 years old and ALL accounts for ~ 20% of all adult leukemia diagnoses. Each year, there are approximately 6,000 new cases of B and T-cell ALL in the United States causing an estimated 1,580 deaths annually [1].

For the last two decades, the 5-year overall survival (OS) for ALL has steadily improved across all age demographics with the largest gains in children 1–14 years of age who now achieve a > 93% 5-year OS following upfront intensive therapy. In contrast, the 5-year OS for adults 20–39 years of age is only 59% with further decrements in overall survival as patients age; with 43% 5-year OS in patients 40–59 years old, 29% in patients 60–69 years old, and only 13% in patients 70 years or older [2]. Despite initial complete response (CR) rates of ≥ 90%, over half of adult patients diagnosed with ALL will ultimately relapse with a median post-relapse OS of 4.5–6 months and a 5-year OS of less than 10% [1].

A major barrier for patients with relapsed ALL has been the limited utility of chemotherapy-based salvage regimens. Due to the low remission rates and short duration of the remissions achieved with salvage chemotherapy, only 10–30% of adult patients were previously able to proceed to a curative, allogeneic hematopoietic stem cell transplant (HSCT), irrespective of age [3, 4].

The limited efficacy of salvage chemotherapy drove the development of novel treatment approaches to try to improve long-term survival in relapsed ALL. Since 2017, four new immunotherapies have received FDA approval for relapsed B-ALL: blinatumomab, inotuzumab ozogamicin, and two chimeric antigen receptor (CAR) T-cell constructs (tisagenlecleucel and brexucabtagene autoleucel). These agents have drastically improved outcomes for relapsed, adult B-ALL and allow providers to choose a targeted approach based on patient and disease characteristics. Frustratingly, these advances have not yet made strides for patients diagnosed with relapsed T-ALL, with no new FDA approved therapiessince 2005 although active research with early phase clinical trials utilizing immunotherapy based approaches are ongoing.

Treatment Options for Patients Diagnosed With Relapsed B cell-ALL

CD19 Targeting

CD19 is essential for the function and maintenance of B-lineage cells, [5] and it is expressed on pre-B cells through their terminal differentiation into plasma cells [6]. Due to its almost universal presence and uniform expression on B-ALL, CD19 serves as one of the disease’s most reliable surface biomarkers and is the target for blinatumomab and the two recently FDA-approved CAR products for relapsed B-ALL [7, 8•, 9•, 10•]. Both blinatumomab and the CD19 CAR therapies function by engaging the cytotoxic potential of T-cells without the need for T-cell receptor specificity, antigen processing/presentation, or major histocompatibility complex context [11, 12].

Blinatumomab

Blinatumomab is a bispecific T-cell engager (BiTE) antibody for CD19 and CD3. By binding CD19 on B-cells and CD3 on primarily cytotoxic CD8 + T-cell (CTLs), blinatumomab results in T-cell activation and a cytotoxic T-cell response against CD19-expressing cells [13]. Blinatumomab received accelerated FDA approval in 2014 for Ph- r/r B-ALL based on initial phase II results [11, 14, 15] with full FDA approval for both Ph- and Ph + r/r B-ALL in 2017 following the TOWER and ALCANTARA trials [8•, 14] (Table 1).

Table 1 Pivotal trials for blinatumomab, inotuzumab, and CAR T-Cell in Relapsed B-ALL

The phase III randomized TOWER trial showed higher composite CR rates in patients with r/r Ph- ALL treated with blinatumomab compared to chemotherapy (44% vs 25%), and blinatumomab was more likely to result in a minimal residual disease negative (MRD-neg) CR (76% vs 48%). Median OS was 7.7 months with the greatest benefit for patients in first or second salvage and HSCT-naïve [8•]. The ALCANTARA trial demonstrated a similar CR/CRh rate (36%) with an MRD-neg CR in 88% of responders in patients with relapsed Ph + B-ALL or refractory to tyrosine kinase inhibitors (TKIs) [14].

In 2018, the FDA labeling was expanded to include MRD + B-ALL based on the single arm phase II BLAST trial for adult patients in morphologic remission but detectable MRD > 10–3 (Table 1). After one cycle of blinatumomab, 78% of MRD + patients achieved an MRD-neg remission with improved RFS (23.6 mo v 5.7 mo) and OS (38.9 mo v 12.5 mo) when compared to MRD non-responders [16•]. Similar to the TOWER trial, OS was superior in first salvage but there was no long-term survival advantage for patients undergoing consolidative HSCT [17]. Subsequent long-term follow-up of the TOWER trial also showed HSCT may not provide an OS benefit in patients achieving a CR/CRh [18].

The results of the pivotal blinatumomab trials are confirmed in real-world analysis with a CR/CRi rate of 65% in r/r patients (and a 73% MRD-neg) and 75% of the MRD + cohort achieving a remission that is MRD-neg. A third of patients (35.5%) were bridged to HSCT with a 2-year PFS and OS of 66% and 62% respectively [19]. For older patients (ages > 55 years old), efficacy and tolerability were also demonstrated including patients with relapse post prior allogeneic HSCT [20, 21].

Toxicity

Due to its unique mechanism of action, blinatumomab has lower rates of grade ≥ 3 neutropenia (37.8% vs 57.8%) and infection (34.1% vs 52.3%), but it can result in a global inflammatory response leading to serious toxicities of cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). Across trials, CRS has been reported in 15% of r/r B-ALL patients and 7% of MRD + patients with recommendations for dexamethasone prophylaxis and an initial low-dose infusion period of blinatumomab in order to decrease the incidence of CRS through a slower tumor debulking [22]. ICANS occurs in ~ 65% of all patients receiving blinatumomab but only 13% of events are grade ≥ 3. Rates of severe ICANS appear to be lower in pediatric patients and increased in individuals ≥ 65 years old [20].

Given its short half-life, toxicities like CRS and ICANS can be mitigated quickly with discontinuation of drug infusion and the use of corticosteroids. Compared to salvage chemotherapy approaches, blinatumomab results in a better quality of life and has no increased risk of post-HSCT complications when using blinatumomab as a bridge to transplant [23,24,25].

Limitations of Blinatumomab

A major limitation to blinatumomab efficacy is decreased efficacy in the presence of high disease burden with bone marrow blasts > 50% at time of treatment. Specifically, the TOWER trial showed a decreased CR/CRh (34.4% vs 64.5%) in patients with high vs low disease burden respectively. This finding was confirmed across multiple trials [26]. Although blinatumomab has been shown to have higher response rates in the MRD + setting compared to r/r B-ALL, the 3-year OS for patients with pre-blinatumomab MRD > 1%, 0.1–1%, and < 0.1% has been shown to be 33%, 58% and 86% respectively. Patients going directly to blinatumomab without bridging therapy also had shorter RFS and OS with 3-year OS of 66% in the bridging cohort compared to 16% in patients directly receiving blinatumomab [27]. Additionally, extramedullary involvement at the time of blinatumomab infusion has also been show to predict inferior treatment response, raising concern for its utility in patients with extramedullary disease (EMD). Of particular concern is the presence of CNS disease, which was identified in 29% of post-blinatumomab relapses [28].

Prior treatment with chemotherapy was shown to adversely affect outcomes in the TOWER trial with improved response rates in patients receiving ≤ 2 prior lines of therapy. [29] A pooled analysis found improved CR/CRh (54% vs 41%), median OS (10.4 vs 5.7 months) and median RFS (10.1 v 7.3 months) in patients receiving blinatumomab as first salvage [30]. Additional mechanisms of blinatumomab resistance may include decline and/or complete loss of CD19 [31, 32], increased activity and level of regulatory T cell (Tregs) in the tumor microenvironment [33], and upregulation of PD-L1 [34]. In the case of KMT2A-rearranged ALL, there have been several case reports that illustrate the potential for blinatumomab to induce lineage switch into an aggressive AML [35].

CAR-T: Tisagenlecleucel and Brexucabtagene autoleucel

As a form adoptive cell therapy, CARs are genetically engineered T cells expressing a synthetic antigen receptor construct that reprogram the T-cell’s specificity and function. Like blinatumomab, the two FDA approved CARs, tisagenlecleucel (tisa-cel) and brexucabtagene autoleucel (brexu-cel), result in a CTL response through CD19 targeting leading to extensive proliferation of CAR T cells. Following the release of tumor antigens, non–CAR T cells are also recruited producing further antitumor effect through the process of cross priming [12]. Tisa-cel and brexu-cel are both second generation CARs differentiated primarily by their co-stimulatory domain (4-1BB and CD28, respectively) neither of which have demonstrated clear superiority in B cell-ALL [36].

Based on the results of the phase 2 ELIANA trial, tisa-cel is approved for patients < 25 years of age who have received two prior lines of treatment which have failed to establish disease response/control or with post-HSCT relapse (Table 1). Following Phase I results in 60 patients showing a CR rate of 93%, the ELIANA trial of 75 patients resulted in an 81% CR/CRi rate with 95% of responding achieving MRD-negativity by day 28 [9•]. Among evaluable patients, tisa-cel resulted in a 5 year OS and RFS of 55% and 44%, respectively [37]. A recent 3-year update of the ELIANA trial, with a median follow-up of 38.8 months, confirmed these initial findings, and illustrated the potential of a single CAR infusion to be a curative treatment option for heavily pre-treated young adult patients (3-year RFS of 76% in responders) [38••].

Approved for patients > 18 years old with relapsed ALL, brexu-cel was FDA approved in October 2021 following the results of the phase I/II ZUMA-3 trial [10•] (Table 1). Brexu-cel is the same CAR construct as axicabtagene autoleucel previously evaluated in pediatric patients at the National Cancer Institute [39] although it has altered manufacturing in order to decrease early ex‐vivo activation and exhaustion of CAR cells by removing leukemic cells from the leukapheresis product [30]. A total of 55/71 enrolled patients were infused on phase II of ZUMA-3 Trial which demonstrated a CR/CRh rate of 71% (and a 97% MRD-negative rate). In this study, the patients were heavily pre-treated with 45% having relapsed post HSCT, 42% of whom had prior exposure to blinatumomab and 22% of whom had prior inotuzumab. With a median follow-up of 16.4 months, the median DOR and RFS following brexu-cel infusion was 12.8 months and 11.6 months, respectively, with 10% of patients proceeding to consolidative HSCT. Of the 55 patients treated on ZUMA-3, 15% were age 65 or older with all 8 patients achieving a CR (compared to 71% in patients < 65 years old) [40]. In the 2-year follow-up, the median OS in responders was 25.4 months versus 5.5 in historical controls [41••]. Compared to tisa-cel, there is stronger support for the use of post-CAR allogeneic HSCT as consolidative therapy for brexu-cel based on experience with CD28 co-stimulatory domains [39, 42,43,44].

Toxicity

Like blinatumomab, the primary toxicity of concern for CD19 CAR constructs is CRS and ICANS. However, unlike blinatumomab’s short half-life and the ability to abort exposure with drug discontinuation, CAR T-cells are a “live” agent with continued exposure that persists in the body. This leads to the potential for increased severity and duration of CAR mediated toxicities (i.e. there is no immediate “kill switch” to stop the ongoing cascade of CRS/ICANS other than steroids/supportive approaches). Rates of grade ≥ 3 CRS and ICANS were 46% and 13%, respectively, with tisa-cel in the Phase 2 ELIANA trial. For brexu-cel, rates of grade ≥ 3 CRS and ICANS were 24% and 25% in the ZUMA-3 study [41••]. While there has been some indication of higher CAR toxicity in older adults treated with CD19 directed therapy for lymphoma, ZUMA-3 did not report differences in CAR toxicity comparing patients ≥ 65 years and < 65 years of age [40].

Aside from the risk of CRS and ICANS, the need for a pre-infusion lymphodepletion regimen can potentially add additional toxicities and patients are also at risk for prolonged cytopenias and infectious complications post-CAR [45, 46].

Limitations of CAR-T

The efficacy of CAR-T therapy, like blinatumomab, is limited by the presence of high disease burden prior to lymphodepleting chemotherapy. High leukemia burden pre-CAR is associated with both a decreased overall survival and an increased risk for CAR-mediated toxicities [47,48,49]. The ZUMA-3 trial found a CR/CRi rate of 42% for patients with bone marrow blasts > 75% compared to a CR/CRi of 80–91% in patients with lower disease burden treated with brexu-cel [30]. A pooled, real-world analysis of tisa-cel showed superior outcomes in patients with undetectable or low burden disease compared to patients with pre-infusion bone marrow blasts ≥ 5%, CNS-3 disease, or non-CNS EMD. The 12-month OS for low burden and undetectable disease was 85% and 95%, respectively, with a similar EFS of 70–72%. In comparison, high disease burden patients had a 12 month OS of 58% and EFS of 31%. Importantly, patients with bone marrow blasts ≥ 5% were also shown to have higher rates of grade ≥ 3 toxicities including CRS, ICANS, and ICU monitoring [50].

Other clinical features associated with impaired response to CD19 targeted CAR therapy include: blinatumomab non-response, prior chemotherapy exposure, and presence of non-CNS EMD. Compared to patients that had a prior response to blinatumomab, nonresponders to blinatumomab had inferior CR rates in response to tisa-cel (65% versus 93%), as well as worse EFS, RFS and OS [49]. Prior chemotherapy exposure was found to portend a worse OS with a hazard ratio (HR) of 1.4, possibly due to the inferior fitness of the T cells collected to manufacture the CAR T product [50, 51]. Like blinatumomab, the presence of EMD at the time of infusion was independently associated with inferior EFS (HR 1.9) [49]. Although CAR therapy appears to be agnostic to cytogenetic risk group with similar outcomes regardless of karyotype [52, 53], the clinical benefit in specific high risk (cytogenetic or mutation based)ALL, such as rearranged KMT2A, mutated TP53 and hypodiploid ALL is limited [53,54,55,56].

In addition to disease specific limitations, the time required [3,4,5,6] to engineer an autologous CAR-T product is be a restricting factor, particularly in proliferative leukemic disease. The initial clinical trials had high dropout rates ranging from 18–32% due to time needed for manufacturing juxtaposed with the infectious complications of leukemia and or death due to disease progression before therapy could be delivered [9•, 10•, 47].

CD22 Targeting

CD22 is present on 90% of B-ALL cells and functions both as an adhesion molecule for other leukocytes and to increase the threshold for antigen/receptor stimulation [57, 58]. Importantly, it remains detected in most cases of CD19 antigen loss following CD19 directed therapy with blinatumomab or CAR T-cell therapy [59, 60]. Unlike CD19, CD22 is rapidly internalized following antibody binding, a feature that is exploited by the antibody–drug conjugate (ADC) Inotuzumab. Inotuzumab is an anti-CD22 monoclonal antibody that is bound to the cytotoxic anti-tumor antibiotic calicheamicin [61]. Following the binding of inotuzumab to the CD22 antigen on B-ALL cells, the complex is internalized, releasing calicheamicin that attaches to the minor grove of the leukemia DNA resulting in DNA cleavage and subsequent apoptosis [62, 63].

Inotuzumab

FDA approval for inotuzumab is based on results of the randomized, phase III INO-VATE trial comparing inotuzumab to the standard of care in r/r B-ALL (Table 1). The CR/CRi rate in the intention-to-treat analysis was 80.7% compared to 29.4% with chemotherapy [64•]. Inotuzumab was also more likely to achieve CR regardless of bone marrow blast (BMB) percentage with no statistically significant difference in CR rate between low (BMB < 50%), moderate (BMB 50–90%) or high (BMB > 90%) disease burden. [65] The majority of responses to inotuzumab occurred following cycle 1 (73%) with 78.4% achieving a MRD-negative CR [64•]. With a median follow-up of 29.6 months, the 3-year OS with inotuzumab was 20% versus 6.5% with standard of care therapy. The survival benefit of inotuzumab was in part due to patients being more likely to proceed with HSCT following inotuzumab (48% vs. 22%) [66••]. Patients with MRD-negative CR consolidated with HSCT after inotuzumab had a median OS of 19.2 months compared to 11.1 months in patients that were MRD-positive [67]. These results with single-agent inotuzumab were confirmed in a multicenter real-word analysis of adult patients diagnosed with ALL treated with inotuzumab outside of clinical trials [68].

For older patients (≥ 55 years of age) who were treated on the INO-VATE trial, there was no statistically significant difference in ORR or MRD-negativity rates compared to younger patients. Older patients were more likely to achieve a CRi (43%) compared to a CR (27%), but there was no difference in median DOR (4.7 mo vs 5.4 mo) in either patient population. OS was shorter in the older cohort (5.6 vs 8.6 months; HR 0.610) likely due to these patients being less likely to proceed to HSCT (28% vs 58%) [69]. Like blinatumomab, outcomes with inotuzumab improved when given as first salvage with a CR/CRi rate of 76% compared to 50% when given as a second or later relapse treatment [70, 71]. The delivery of inotuzumab as a short IV infusion in the outpatient setting allows for the ability to administer the therapy in community oncology practices. Importantly, from a quality of life standpoint, patients had lower rate of hospitalization when compared to traditional salvage chemotherapy [72].

Toxicity

Aside from an association with increased risk of myelosuppression and neutropenic fever [64•], the most significant toxicity associated with inotuzumab is veno-occlusive disease (VOD). VOD occurred in 13% of patients treated with inotuzumab on the INO-VATE trial compared to 3% in the standard of care arm. Of those with inotuzumab related VOD, 82% were grade III or higher. For those patients receiving a consolidative HSCT, there was an increased rate of VOD if receiving inotuzumab (18.8–22%) compared to standard chemotherapy [73, 74]. For patients ≥ 55 years of age, VOD post-HSCT occurred in 41% of patients compared to 7% in patients 18–29 years of age and 24% in patients 30–54 years of age [69]. Risk of inotuzumab related VOD post-HSCT can been reduced by several approaches: (1) using fractionated dosing, (2) limiting the number of cycles to ≤ 2, (3) avoiding dual alkylator therapy, (4) confirming normal bilirubin pre-HSCT, (5) allowing for a minimum of 4 weeks between inotozumab and HSCT conditioning regimen initiation, and (6) using ursodiol through day + 90 post HSCT [18, 73,74,75,76]. For these reasons, a real-world analysis in adults found lower VOD rates post-HSCT that were comparable to historical rates in patients without inotuzumab exposure[68].

Limitations of Inotuzumab

Subgroup analysis from the INO-VATE trial found benefit for all patients achieving CR except Ph + B-ALL and patients with < 70% CD22-positivity. Further analysis following the widespread incorporation of TKI therapy for Ph + B-ALL, showed superior median OS in the Ph + patients compared to other cytogenetic abnormalities (NR vs 9.4 mo) [68]. KMT2A-r B-ALL is a high-risk population, which is known to have a lower average CD22 antigen density and the existence of sub-population of blasts that lack CD22 expression [58]. KMT2A-r B-ALL has been shown in multiple studies to be independently associated with inferior OS [68].

Unlike blinatumomab and CAR, high disease burden does not affect outcomes with inotuzumab, but there also appears to be no significant difference in OS for patients with primary refractory disease highlighting the lack of long-term durability of these therapies [68]. There also appears to be evidence that inotuzumab is effective in relapsed patients with non-CNS EMD with CR rate of 55% and 2-year OS of 18% in this challenging to treat population [77]. Despite the ability to use inotuzumab in the setting of high tumor burden or EMD, durability of remissions is short necessitating consolidative HSCT. Patients consolidated with allogeneic HSCT had a 2-year OS of 39% compared to 13% in those not proceeding to HSCT [66••].

T-ALL

Tremendous progress has been made in the treatment of relapsed adult B-ALL, however, relapsed T-ALL remains a serious challenge. The only FDA approved therapy for relapsed T-ALL remains Nelarabine. This drug was approved in 2005 following demonstration of its efficacy as a single agent in adult relapsed T-ALL with a modest CR rate of 31% and a 1-year OS of only 28% [78]. Recent incorporation of nelarabine in the Children’s Oncology Group’s upfront pediatric T-ALL (including patients up to 31 years of age) resulted in an improved 5-year DFS of 91% but showed no difference in OS [79]. Based on these results, nelarabine is expected to move to the upfront setting in younger adults and fit older adults likely decreasing its utility in the relapsed setting.

Recently, inclusion of the BCL2 inhibitor Venetoclax has shown promise in combination regimens for r/r T-ALL, although these studies have enrolled small numbers of patients. The CR rate for relapsed T-ALL was 60% when venetoclax was added to a traditional chemotherapy backbone, and a phase I trial combing it with the BCL-XL inhibitor navitoclax resulted in a CR rate of 52.6% [80, 81]. However, the movement of these BH3 mimetics (venetoclax/navitoclax) into frontline setting may limit their efficacy in relapsed T-ALL.

Unfortunately, unlike B-ALL, the development of immune targeting approaches in T-ALL remains in early developmental stages. Given the high expression of CD38 on T-ALL, monoclonal antibodies like daratumab and isatuximab, with proven efficacy in multiple myeloma, have been tried with modest results [82, 83]. A trial with a bispecific T-cell–recruiting antibody, similar to blinatumomab, is underway, which also targets CD38 (NCT05038644). Early phase clinical trials testing CD5 and CD7 CAR-T constructs are also looking to overcome the challenge of fratricide caused by the target antigen being expressed both on the T-ALL cells and the CAR-T product [84,85,86].

Challenges and Future Directions

Treatment Selection Between Novel Agents in B-ALL and Growing Upfront Exposure

To date, no randomized head-to-head comparison exists between blinatumomab, inotuzumab and CAR-T in B-ALL, with the only proposed trial (NCT03628053) previously withdrawn. Concerns exist for the interplay of these treatment approaches and the influence of subsequent therapies. Thankfully, real-world analysis has shown comparable efficacy whether patients receive inotuzumab after blinatumomab or the reverse with response rates of 58% vs 52%, respectively [87]. In regards to CAR, neither prior Intozumab exposure nor blinatumomab have been show to effect outcomes, as long as a patient’s leukemia had a prior response and retained CD19 expression [30, 38••, 42, 49].

While some practice preferences have emerged in relapsed B-ALL (CD19 CAR for post-HSCT relapse or HSCT ineligible, inotuzumab for high disease burden or proliferative disease, and blinatumomab for MRD + disease with an available transplant donor), patient and logistical preferences must be taken into account when selecting therapy. Importantly, as blinatumomab, inotuzumab, and CAR-T therapy all are incorporated into the upfront treatment setting, the majority of patients may be exposed to one or more agents earlier in their disease course and this has the potential to decrease future effective therapy in the relapsed setting [88]. Clinical outcomes for patients with B-ALL with non-response or relapsing post-CAR-T, as well as after relapse following sequential treatment with blinatumomab and inotuzumab, remain dismal [89, 90].

Role of Allogeneic HSCT for Relapsed B-ALL

Previously, the greatest limitation to durable remissions in relapsed, adult B-ALL was the inability to achieve an MRD-negative remission in order to proceed to allogeneic HSCT. However, with the significant improvement in CR rates in patients treated with blinatumomab, inotuzumab, or CAR-T therapy, the question emerges if there is the same necessity for consolidative HSCT. However, based on the available data, most patients still require HSCT for durable remissions regardless of the product used.

While some patients with MRD + B-ALL treated with blinatumomab were able to achieve durable remissions without HSCT, the majority of patients (61%) not undergoing HSCT ultimately relapse [17]. As a single agent, blinatumomab produces superior outcomes for patients when bridging to a consolidative HSCT [18, 91]. In a retrospective analysis of the inotuzumab trials the OS was significantly improved with HSCT (51% vs 22.8%), and remissions achieved with inotuzumab similarly appear to be short lived without a consolidative HSCT [75].

Compared to blinatumomab and inotuzumab, the possibility for durable remission without HSCT exists with CAR T-cell therapy, but it is dependent on the CAR construct and ongoing CAR T cell persistence. The 3-year RFS on the ELIANA trial for patients responding to the tisa-cel but receiving no subsequent therapy was 76% [38••]. Predictive models for post-CAR relapse based on retrospective analyses are being developed with ongoing b-cell aplasia and next generation sequencing (NGS)-MRD as the sole predictive factors identified to date [92•]. The upcoming CAR-CURE trial (NCT05621291) will seek to answer this question prospectively. For CAR products using a CD28 domain, like brexu-cel, HSCT is still felt to be of central importance for meaningful, long-term disease remissions [39, 47].