Current assessment and management of measurable residual disease in patients with acute lymphoblastic leukemia in the setting of CAR-T-cell therapy

Abstract

Chimeric antigen receptor (CAR)-modified T-cell therapy has achieved remarkable success in the treatment of acute lymphoblastic leukemia (ALL). Measurable/minimal residual disease (MRD) monitoring plays a significant role in the prognostication and management of patients undergoing CAR-T-cell therapy. Common MRD detection methods include flow cytometry (FCM), polymerase chain reaction (PCR), and next-generation sequencing (NGS), and each method has advantages and limitations. It has been well documented that MRD positivity predicts a poor prognosis and even disease relapse. Thus, how to perform prognostic evaluations, stratify risk based on MRD status, and apply MRD monitoring to guide individual therapeutic decisions have important implications in clinical practice. This review assesses the common and novel MRD assessment methods. In addition, we emphasize the critical role of MRD as a prognostic biomarker and summarize the latest studies regarding MRD-directed combination therapy with CAR-T-cell therapy and allogeneic hematopoietic stem cell transplantation (allo-HSCT), as well as other therapeutic strategies to improve treatment effect. Furthermore, this review discusses current challenges and strategies for MRD detection in the setting of disease relapse after targeted therapy.

Introduction

Acute lymphoblastic leukemia (ALL) is a major type of hematologic malignancies characterized by the abnormal proliferation of B-cell or T-cell lineage cells originating in the bone marrow (BM).^[1] The survival rate of patients with relapsed or refractory (r/r) ALL remains unsatisfactory, as the overall 5-year survival rate stands at only 10%–20%.^[2,3] There is ongoing research of new treatment strategies and medications in clinical practice aimed at improving the survival rate of these patients. Chimeric antigen receptor (CAR)-modified T-cell therapy is a novel immunotherapy that involves genetically modified T cells with CAR structures to specifically identify and eliminate tumor cells expressing specific antigens, and it has achieved excellent success in the treatment of hematological malignancies.^[4,5] As of June 2022, six CAR-T-cell products have been approved for marketing by the U.S. Food and Drug Administration. Among them, tisagenlecleucel (Kymriah) and brexucabtagene autoleucel (Tecartus) have been approved for the treatment of r/r ALL. In addition, the number of clinical trials related to CAR-T-cells has significantly increased.^[6] However, long-term follow-up data from clinical trials suggests that recurrence after CAR-T-cell treatment remains a major challenge.^[5]

Measurable/minimal residual disease (MRD) is defined as a small number of posttherapy tumor cells that cannot be detected by morphological or other traditional methods, but can be identified through highly sensitive detection methods, such as molecular biology and flow cytometry (FCM).^[7–9] The detection of MRD plays a significant role in prognosis evaluation of ALL during clinical treatment. By monitoring the status of MRD, treatment response can be more accurately assessed, and the risks of recurrence and survival can be predicted, thereby guiding further treatment decisions.^[10–12] The application of MRD in the field of CAR-T-cell therapy has been receiving growing attention with the advancement of targeted therapy. However, the detection of MRD after CAR-T-cell therapy has a great challengen due to tumor heterogeneity and off-target effects.

In this review, we summarize the common and novel methods of MRD monitoring in the era of CAR-T-cell therapy. We emphasize that MRD has unequivocal prognostic value for ALL patients, both before and after treatment. We focus on the MRD-directed combination therapy, especially CAR-T-cell therapy and allogeneic hematopoietic stem cell transplantation (allo-HSCT), including CAR-T-cell therapy bridging to allo-HSCT and a preemptive strategy of posttransplant. In addition, other promising MRD-directed interventions during CAR-T-cell therapy is also explored. In the end, this review highlights the new features of relapse after CAR-T-cell therapy, particularly the negative relapse associated with loss of CAR-T target antigen expression, posing new challenges to the existing MRD monitoring methods and strategies.

Assessment of MRD in Patients Receiving CAR-T-Cell Therapy

In clinical practice, we must focus on detection techniques, sample requirements, detection timepoints, data analysis, etc. The detailed MRD assessment procedure for patients with r/r ALL is described in Figure 1.

Figure 1.

The procedure of MRD monitoring for patients with r/r ALL. During CAR-T-cell treatment, patients with r/r ALL are recommended to monitor MRD at specific timepoints. Bone marrow is the main specimen for MRD monitoring, but peripheral blood is becoming a kind of promising and non-invasive sample. After finishing specimen collection, FCM, PCR, NGS, or other new technologies are performed to detect MRD. The final results of MRD are influenced by both the disease burden and the sensitivity of the detection methods. As the sensitivity of existing technologies ranges from 10^–4 to 10^–6, MRD status is defined as detectable positive when the disease burden is above the sensitivity; otherwise, it is defined as undetectable MRD or even truly negative MRD. ALL: Acute lymphoblastic leukemia; CAR: Chimeric antigen receptor; ddPCR: Droplet digital polymerase chain reaction; FCM: Flow cytometry; MRD: Measurable/minimal residual disease; NGF: Next-generation flow; NGS: Next-generation sequencing; PCR: Polymerase chain reaction; RNA: Ribonucleic acid; r/r ALL: Relapsed/refractory acute lymphoblastic leukemia. (Created with BioRender.com)

Techniques of MRD detection

The main methods for detection MRD, including FCM and polymerase chain reaction (PCR), have been widely adopted.^[13] These methods have established the groundwork for using MRD as an indicator of treatment efficacy. In recent years, there have been advancements in novel technologies in the context of precision medicine, such as droplet digital PCR (ddPCR), next-generation flow (NGF), next-generation sequencing (NGS), and RNA-sequencing, which offer more possibilities for MRD detection. However, it should be noted that these technologies are still in the research stage and are not routinely used in the clinic.^[14–17] The principles and characteristics of these methods are summarized in Table 1.

Table 1.

Characteristics of current main MRD detection methods.

Detection methods	Target	Diagnostic sample	Sample origin	Sensitivity	Applicability	Whether need patient-specific primers/probes	Advantages	Disadvantages
FCM	LAIP/DFN	LAIP: mandatory DFN: not mandatory	Fresh viable cells	3–4 colors: 10–3–10–4; 6–8 colors: 10–4; >8 colors: 10–6	>90%	No	Fast, inexpensive, high sensitivity	Potential immunological phenotypic shift, confounders, such as normal and regenerative B-cell precursors during post-treatment follow-up, require significant expertise
qPCR	IGH/TCR rearrangements	Mandatory	DNA	10–4–10–5	>90%	Yes	Good standardization, high sensitivity, applicable in virtually all B-ALL, easy-to-use	Time-consuming, require significant expertise, require constructed primers, potential to miss small subclones and clone evolution
RT-qPCR	Fusing genes	Mandatory	RNA	10–4–10–5	35%–45% (age dependent)	Yes	High sensitivity, inexpensive, easy-to-use	Not application in every case, limited standardization
ddPCR	Target genes (e.g., IGH/TCR rearrangements, fusing genes)	Mandatory	DNA/RNA	10–4–10–5	>90%	Yes	Absolute quantitative and very high sensitivity, independent of standard curve, suitable for complex matrix samples	Needs high technical ability, relatively expensive, limited standardization
NGF	LAIP	Mandatory	Cells	10–4–10–6	>95%	No	Very high sensitivity, fast and reproducible, highly standardized	Relatively expensive, require fresh material
NGS	Target genes (e.g., IGH/TCR rearrangements, fusing genes)	Mandatory	DNA	10–6	>95%	No	Relatively fast, very high sensitivity, potential to identify small subclones and clone evolution	Expensive, require a good grasp of complex bioinformatics, limited standardization
RNA-sequencing	Target genes (e.g., IGH/TCR rearrangements, fusing genes)	Mandatory	RNA	10–5–10–6	>90%	No	High sensitivity, potential to identify small subclones and clone evolution	Not suitable for detecting non-transcribed unproductive and incomplete gene rearrangements, require significant expertise, limited clinical validation and standardization

B-ALL: B-cell acute lymphocytic leukemia; ddPCR: Droplet digital polymerase chain reaction; DFN: Different from normal; DNA: Deoxyribonucleic acid; FCM: Flow cytometry; IGH: Immunoglobulin heavy chain; LAIP: Leukemia-associated immunophenotypes; MRD: Measurable/minimal residual disease; NGF: Next-generation flow; NGS: Next-generation sequencing; qPCR: Real-time quantitative polymerase chain reaction; RNA: Ribonucleic acid; RT-qPCR: Reverse transcriptase quantitative polymerase chain reaction; TCR: T-cell receptor.

Sample requirements, detection timepoints, and thresholds in MRD monitoring

Bone Marrow (BM) was the first choice for MRD detection in B-cell ALL (B-ALL) and was not recommended to be replaced by peripheral blood (PB).^[18–20] In contrast, several studies have suggested that PB samples are alternative sources in T-cell ALL (T-ALL).^[18] BM specimens used for MRD detection have some defects. The uneven distribution of tumor cells in BM can lead to false testing results, and it is difficult for patients to accept frequent operations since BM puncture is an invasive operation. Studies have suggested that non-invasive liquid biopsy, which analyzes cell-free DNA (cfDNA) and other peripheral blood components by PCR or NGS, seems to be an alternative approach.^[21,22] But in published clinical studies of CAR-T-cell therapy, liquid biopsy has only been applied for MRD monitoring in diffuse large B-cell lymphoma (DLBCL).^[23] In addition, it is also recommended that cerebrospinal fluid or other affected tissues that can be sampled should be detected given the potential for extramedullary relapse.

The optimal timepoints for MRD detection have not been determined since the data on targeted therapy are limited. A deeper understanding of CAR-T-cell dynamics is helpful for identifying timepoints. Researchers have suggested that patients who do not receive further treatments after CAR-T-cell therapy should undergo BM puncture to assess MRD every 3 months for 6 to 12 months.^[24] In addition, MRD detection is recommended at arbitrary timepoints when disease progression is suspected in combination with clinical manifestations.^[25] The threshold value of MRD is influenced by many factors, including detection timepoints, treatment methods, and specimens. Thus, there is no unified standard for the MRD threshold, but 0.01% has been the most widely used threshold in many clinical trials of CAR-T-cell therapy.^[4,5]

Performance of MRD detection methods in CAR-T-cell therapy

Previous studies have suggested the superiority of NGS over FCM and PCR in monitoring MRD status for patients receiving chemotherapy or HSCT.^[26–28] To explore a more sensitive MRD assay in the setting of CAR-T-cell therapy, Huang et al^[29] retrospectively enrolled 27 patients with B-ALL who had achieved complete remission (CR) after CAR-T-cell therapy, and they measured MRD in 63 samples. Using a threshold of 0.01%, discordance between NGS and FCM was identified in 17 of 63 samples (27%). The controversial group with NGS-MRD-positive and FCM-MRD-negative patients had inferior leukemia-free survival (LFS) compared to those with NGS-MRD-negative and FCM-MRD-negative patients (P = 0.037). These results indicated that NGS predicted tumor load dynamics and prognosis better than FCM. Similarly, Pulsipher et al^[30] compared the results of FCM-MRD with those of NGS-MRD in the same samples from patients with r/r ALL receiving tisagenlecleucel therapy. The research showed that both methods detected a high percentage of original cells in BM samples, but NGS-MRD was more sensitive than FCM-MRD. By analyzing the detection results of FCM-MRD, NGS-MRD with a sensitivity of 10^–4, and NGS-MRD with a sensitivity of 10^–6, in qualified samples from relapsed patients, it was observed that the effectiveness rates were 50%, 69%, and 100%, respectively. In addition, NGS-MRD was also more sensitive than FCM-MRD in PB samples, revealing that MRD measurement by NGS was a more sensitive biomarker to date for defining the risk of relapse after CAR-T-cell therapy.^[30]

Prognostic Value of MRD Assessment in CAR-T-Cell Therapy

The prognostic significance of MRD in chemotherapy, HSCT, and other treatments have been confirmed for patients with r/r ALL.^[31–34] Here, we focus on the prognostic value of MRD before and post CAR-T-cell therapy.

MRD assessment before CAR-T-cell therapy predicts treatment prognosis

Several studies have illustrated that disease burden before CAR-T-cell therapy was a remarkable predictor of remission duration and survival.^[4,35,36] Wudhikarn et al^[35] reported the results of a single-center, phase I study at Memorial Sloan Kettering Cancer Center (MSKCC) that enrolled 56 adult B-ALL patients treated with cluster of differentiation (CD)19-CAR-T cells. They showed that a higher disease burden, as indicated by bone marrow blasts ≥5% or extramedullary disease, rather than MRD prior to CAR T-cell infusion, was the only significant risk factor associated with the risk of disease progression (P = 0.02). However, in a prospective clinical trial involving 28 r/r B-ALL patients who underwent humanized CD19-targeted CAR-T-cell therapy, Chen et al^[36] indicated that a higher disease burden, defined as MRD ≥65.6% prior to intervention, was associated with poor overall survival (OS) (P = 0.017) and LFS (P = 0.001). These findings emphasize the importance of accurately assessing MRD before initiating CAR-T-cell therapy and highlight the need for timely intervention to improve treatment outcomes in patients with r/r ALL.

Continuous MRD monitoring after intervention predicts prognosis in CAR-T-cell therapy

The significance of continuous MRD monitoring following CAR-T-cell therapy as a robust prognostic indicator has been demonstrated.^{[4,30,37–43]}Table 2 presents the results of recent clinical studies regarding the association of MRD status after CAR-T-cell treatment with outcomes for ALL patients, with a majority of studies focusing on FCM and PCR methods for MRD detection. For instance, Park et al^[4] performed a prospective study in 53 B-ALL patients treated with CD19-CAR-T cells at Memorial Sloan Kettering Cancer Center. Event-free survival (EFS) (P <0.001) and OS (P <0.001) were significantly longer among patients with MRD-negative CR than among those with either MRD-positive CR or no response after treatment. Hay et al^[37] compared patients who achieved MRD-negative CR after CAR-T-cell therapy with those who did not respond, and indicated that patients with MRD-negative CR were associated with better EFS (P <0.0001) and OS (P = 0.014). Similarly, Zhang et al^[40] published the results of a phase I trial in which 110 r/r B-ALL patients received an infusion of CD19 CAR-T-cells with a 4-1BB costimulatory, and revealed that there was a trend toward higher OS (P = 0.136) and LFS (P = 0.135) for patients with MRD-negative CR vs. MRD-positive CR after treatment. However, the small number of patients with MRD-positive status prevented the identification of statistically significant difference. Furthermore, recent research has revealed that the use of highly sensitive NGS detection methods to assess MRD has shown promising application value. Specifically, a negative BM NGS-MRD status at 3 and 6 months after tisagenlecleucel infusion has been found to significantly correlates with a reduced risk of relapse and an extended survival period for patients (P <0.05).^[30]

Table 2.

Recent studies regarding the association of MRD status after CAR-T-cell treatment with outcomes for patients receiving CAR-T-cell therapy.

Trials	Sample size	Age (years), median (range)	Target antigen/costimulatory domain	Methods for MRD	Cutoff value, %	Outcomes
Park et al[4]	53	44 (23–74)	CD19/CD28	FCM	0.01	MRD-negative CR was associated with longer EFS (P <0.001) and OS (P <0.001) than MRD-positive CR or no response after treatment
Hay et al[37]	53	39 (20–76)	CD19/4-1BB	FCM	0.01	Patients who achieved MRD-negative CR compared with those who did not respond were associated with better EFS (P <0.0001) and OS (P = 0.014).
Jiang et al[38]	58	NR	CD19/4-1BB	FCM	0.01	MRD-negative CR was independently associated with higher OS and EFS (P <0.05)
Pan et al[39]	90	6 (0.6–18)	CD19/CD28 CD22/4-1BB	FCM	0.01	Positive MRD was associated with higher EFS (P <0.001)
Zhang et al[40]	110	12 (2–61)	CD19/4-1BB	FCM	0.01	There was a trend toward higher OS (P = 0.136) and LFS (P = 0.135) for patients with MRD-negative CR than those with MRD-positive CR after CAR-T-cell therapy
Yan et al[41]	50	23.5 (6–51)	CD19/NR CD22/NR	FCM	0.01	Positive MRD was associated with lower EFS (P <0.001) and higher 1-year CIR (P <0.001)
Ma et al[42]	9	34.1 (16–57)	CD19/4-1BB	FCM/qPCR	0.01	MRD-negative CR was associated with higher 6 months OS rate (76% vs. 14%)*
Dourthe et al[43]	51	17 (1–29.2)	CD19/4-1BB	qPCR	0.01	The absence of complete MRD response was associated with an increased CIR (P = 0.006)
Pulsipher et al[30]	109	12 (3–25)	CD19/4-1BB	NGS	0.0001	A similar predictive power for less relapse (P <0.001) and better OS (P <0.05) was shown for patients with negative NGS-MRD at 3 and 6 months

^*No significant difference. CAR: Chimeric antigen receptor; CIR: Cumulative incidence of relapse; CR: Complete remission; DOR: Duration of response; EFS: Event-free survival; FCM: Flow cytometry; LFS: Leukemia-free survival; MRD: Measurable/minimal residual disease; NGS: Next-generation sequencing; NR: Not reported; OS: Overall survival; PFS: Progression-free survival; qPCR: Real-time quantitative polymerase chain reaction.

MRD-Directed CAR-T-Cell Therapy: Enhancing Precision and Efficacy in ALL

The integration of MRD assessment into the decision-making process is crucial due to its ability to accurately evaluate treatment response and prognosis. MRD status can guide the optimal timing of CAR-T cell therapy, and patients who are MRD negative CR prior to targeted therapy tend to experience better outcomes, as mentioned above. Moreover, considering previous studies have proven that MRD are of great value in adapting chemotherapy and HSCT therapy, MRD assessment also has the potential to guide the adjustment of CAR-T cell therapy strategy.^[44] It is worth exploring whether patients with positive MRD should be recommended to undergo more intensive preparative regimens and higher doses of cell infusions compared to MRD-negative patients. However, there is currently limited research data available in this regard.

Currently, research is primarily focused on the combination of CAR-T-cell therapy with other treatment modalities. On the one hand, CAR-T-cell therapy can be applied to reduce tumor load in r/r ALL patients or even clear MRD in patients with morphological CR prior to allo-HSCT. On the other hand, it can be used as a consolidation or maintenance therapy to reduce the risk of posttransplant recurrence in high-risk B-ALL patients. Given that CAR-T-cell therapy and allo-HSCT therapy are both critical and effective treatments for r/r ALL, it is particularly important to optimize the combination of these therapies in clinical practice, and MRD assessment can play a role in guiding this approach.

CAR-T-cell therapy bridging to allo-HSCT

Consolidative allo-HSCT following CAR-T-cell therapy is almost an effective treatment model for patients with r/r B-ALL. However, which patients should receive allo-HSCT therapy and when allo-HSCT therapy should be applied have not been verified. Many previous studies have emphasized the significance of pre-HSCT MRD status after targeted therapy as a prognostic factor.^[36,45–49] In addition, researches have revealed that MRD negativity predicted favorable outcomes and even no increased the risks of treatment-related toxicities.^[47,50]Table 3 presents clinical studies regarding the association of pre-transplant MRD status with outcomes for patients receiving CAR-T-cell therapy followed by allo-HSCT. Therefore, it is important to investigate whether MRD status could serve as a useful therapeutic indication for consolidative allo-HSCT.

Table 3.

Recent clinic studies regarding the association of pre-transplant MRD status with outcomes for patients receiving CAR-T-cell therapy followed by allo-HSCT.

Authors	Sample size	Age (years), median (range)	Target antigen/Costimulatory domain	Lymphodepletion before HSCT	Transplant donors	Methods for MRD	Cutoff value (%)	Outcomes
Yan et al[41]	50	23.5 (6–51)	CD19/NR CD22/NR	Bu/Cy or TBI/Cy or Flu	HID/MSD/MUD	FCM	0.01	Compare with patients with negative MRD (<0.01%) before HSCT, patients with positive MRD had higher CIR (P = 0.038) and lower EFS (P = 0.008)
Zhao et al[46]	55	26 (3–65)	CD19/CD28	Bu/Cy or TBI/Cy	HID/MSD/MUD	FCM	0.01	MRD positivity at transplantation was an independent factor associated with poor DFS (P = 0.005), OS (P = 0.035), and high CIR (P = 0.045)
Hu et al[47]	52	8 (1–17)	CD19/4-1BB CD19/CD28	Ara-c/Bu/Cy	HID	FCM	0.01	MRD-positive status was associated with poor OS (P = 0.045)
Zhao et al[48]	27	11 (3–44)	CD19/4-1BB	TBI/Cy/Flu or Bu/Cy/Flu	HID/MUD/MSD	FCM	0.01	Negative MRD was associated with lower CIR (P = 0.032), superior LFS (P = 0.024), and OS (P = 0.02)
Chen et al[36]	10	19 (6–54)	CD19/4-1BB	Bu/Cy	HID/MUD	FCM	0.01	Negative MRD was associated with superior LFS (P = 0.032)
Li et al[49]	137	9.7 (1–56)	CD19/4-1BB CD22/4-1BB	TBI/Flu or Bu/Flu	HID/MUD/MSD	FCM/qPCR	0.01	MRD positivity were independent risk factors for LFS (P = 0.033)

allo-HSCT: Allogeneic hematopoietic stem cell transplantation; Ara-C: Cytarabine; Bu: Busulfan; CAR: Chimeric antigen receptor; CIR: Cumulative incidence of relapse; Cy: Cyclophosphamide; DFS: Disease-free survival; EFS: Event-free survival; FCM: Flow cytometry; Flu: Fludarabine; HID: Haploidentical donor; HSCT: Hematopoietic stem cell transplantation; LFS: Leukemia-free survival; MRD: Measurable/minimal residual disease; MSD: Human leukocyte antigen (HLA)-matched sibling donor; MUD: HLA-identical unrelated donor; NR: Not reported; OS: Overall survival; TBI: Total body irradiation.

Achieving MRD negativity: selective bridging to allo-HSCT for high-risk patients

We focus on the influence of MRD status on the benefits of consolidative allo-HSCT and attempt to further clarify the indications for consolidative allo-HSCT based on MRD status. Several studies have demonstrated that allo-HSCT can provide significant clinical benefits in patients with MRD-negative CR after CAR-T-cell therapy.^[37,38] For instance, a study in China enrolled 47 r/r B-ALL patients with or without consolidative allo-HSCT following CAR-T-cell therapy during MRD-negative CR. The study found significant differences in EFS and relapse-free survival (RFS) between the patients who received allo-HSCT and those not (P <0.05).^[38] To identify the candidates who may potentially benefit from consolidative allo-HSCT, the patients were subgrouped based on their pre-infusion MRD and genetic phenotypes. The findings indicated that consolidative allo-HSCT significantly prolonged EFS and RFS (P <0.05) in the subgroups with either high pre-infusion MRD (≥5%) or poor prognostic markers. However, there was no difference in the subgroup with low pre-infusion MRD and without poor prognostic markers regardless of whether they received allo-HSCT. Besides, it is crucial to take into account the presence of adverse prognostic markers when evaluating the potential benefits of consolidative transplantation in patients with MRD-negative CR to avoid overmedicalization. Other studies have indicated that pretreatment lactate dehydrogenase levels, platelet counts, lymphodepletion regimens, phenotypic characteristics of CAR-T cells, and so on also influence the efficacy of allo-HSCT.^[37,51,52]

Achieving MRD positivity: no definite conclusion for consolidative allo-HSCT

The benefits of allo-HSCT in r/r B-ALL patients with MRD-positive CR or morphological relapse after CAR-T-cell therapy remain a matter of debate due to limited research data.^[36,53] In a study conducted by Song et al^[53], 23 r/r B-ALL patients who were in non-remission (NR) or MRD-positive after CAR-T-cell treatment to evaluate the efficacy and safety of salvage allo-HSCT. The 1-year OS and LFS values were 85.7% and 35.7%, 68.1% and 40.9%, in NR and MRD-positive groups, respectively. Thus, the researchers considered allo-HSCT to be a safe and effective salvage modality. Zhao et al^[46] performed a multicenter, retrospective study involving 122 patients who received CAR-T-cell therapy, including 67 patients without subsequent HSCT (non-transplant group) and 55 patients with subsequent haplo-HSCT (transplant group). The study found that the transplant group was associated with higher 2-year OS (P <0.001) and LFS (P <0.001) compared to the non-transplant group. Further analysis showed that although the MRD-negative group had better outcomes than the MRD-positive group and non-transplant group, the CIR (P = 0.139), LFS (P = 0.305), and OS (P = 0.231) did not differ significantly between the MRD-positive group with allo-HSCT and non-transplant group.

Based on the current research advancements listed, it has been observed that patients who achieved MRD negativity after targeted therapy and received subsequent HSCT tended to experience better clinical outcomes than patients with positive MRD. It is important to note that MRD status can change over time from negative to positive or even relapse after CAR-T-cell treatment.^[4] Therefore, achieving MRD-negative CR after targeted therapy and choosing a suitable therapeutic window for allo-HSCT are critical considerations for patients without contraindications. Some researchers have suggested consolidative allo-HSCT within 3 months of CAR-T-cell therapy.^[54] In addition, for patients who cannot achieve MRD negativity after therapy, it is recommended to explore new strategies to intervene MRD before proceeding with consolidative allo-HSCT.

Meanwhile, considering that allo-HSCT will increase the economic burden and can cause treatment-related complications, such as life-threatening graft-versus-host disease (GvHD), it is not applicable in all patients with r/r B-ALL after CAR-T-cell treatment.^[55] Based on this perspective, it is recommended to comprehensively evaluate patient factors (e.g., age, previous transplant history, complications) and disease factors (e.g., pretreatment lactate dehydrogenase level, platelet count, lymphodepletion regimen) during CAR-T-cell therapy and perform risk stratification in patients based on MRD monitoring to determine whether consolidative allo-HSCT therapy should be performed. Hence, we established an algorithm for MRD-directed consolidative allo-HSCT in B-ALL [Figure 2]. However, there remain limitations of this algorithm and more long-term outcome data are needed to constantly redefine the optimal therapeutic indications.

Figure 2.

Proposed algorithm for MRD-directed consolidative allo-HSCT after CAR-T-cell therapy (A) and factors associated with high risk of disease recurrence (B). Relapsed/refractory B-ALL patients receiving CAR-T-cell therapy are recommended to monitor MRD at specific timepoints. For patients with MRD-negative CR, it is necessary to evaluate factors associated with high risk of disease recurrence to determine whether consolidative allo-HSCT therapy should be performed. In addition, salvage therapeutic strategies should be considered for patients with MRD-positive CR or even no CR. allo-HSCT: Allogeneic hematopoietic stem cell transplantation; B-ALL: B-cell acute lymphoblastic leukemia; CAR: Chimeric antigen receptor; CR: Complete remission; MRD: Measurable/minimal residual disease.

CAR-T-cell infusion after allo-HSCT

It is vital to apply CAR-T-cell therapy appropriately after transplantation despised allo-HSCT is one of the primary treatments for patients with ALL. Multiple studies have demonstrated that CAR-T-cell infusion is a promising therapeutic option for patients who relapse after allo-HSCT, but it does not always yield satisfactory outcomes.^{[4,51,56–59]} Hence, researchers explored advancing the intervention time to maximize the benefits of targeted therapy. Table 4 presents the results of current clinical trials of patients receiving CAR-T-cell infusion after allo-HSCT in B-cell malignancies. Study has been proved that tumor burden will affect the efficacy of CAR-T-cell therapy, and MRD positivity after allo-HSCT indicates an increased risk of subsequent recurrence.^[58] Therefore, investigating whether preemptive CAR-T-cell infusion in B-ALL patients with positive MRD is superior to salvage CAR-T-cell intervention after allo-HSCT is worthwhile. Zhang et al^[59] conducted a retrospective study on 43 B-ALL patients who experienced relapse after transplantation and underwent CD19-CAR-T cell therapy. The study found that 34 patients (79%) achieved complete histological remission. This included 12 out of 13 patients with a bone marrow blast rate of 0.01%–5% and positive MRD detection, 14 out of 20 patients with a bone marrow blast rate of 5%–50%, and 8 out of 10 patients with a bone marrow blast rate of ≥50%. Though the study did not find any significant differences between the treatment groups, it is worth noting that this comparison was not adjusted for other factors such as pre-infusion conditioning regimens or the types of CAR-T cells used. In this regard, Zhao et al^[60] performed a prospective clinical trial at PUIH, enrolling 12 MRD-positive B-ALL patients who subsequently accepted preemptive donor-derived CD19-CAR-T-cell infusion. All patients achieved MRD-negative CR without experiencing acute GvHD. The CIR, disease-free survival (DFS), and OS were 42.8%, 65.6%, and 100%, respectively, with a median follow-up time of 424.5 days. Generally, these results indicated that a preemptive strategy based on monitoring of MRD status might be a better therapeutic pattern than the outcomes of other clinical trials listed in the table, which mostly infused CAR-T-cells after relapse. Overall, limited by the insufficient follow-up duration, the efficacy of CAR-T-cell therapy after allo-HSCT remains under further investigation, and prospective, multicenter studies are necessary to provide further evidence for preemptive CAR-T-cell infusion strategies posttransplantation in the future.

Table 4.

Recent clinical trial registrations on CAR-T cells infusion after allo-HSCT in ALL.

Register number/reference	Sample size	Age (years), median (range)	Pre-HSCT	Source of CAR-T cells	Target antigen/costimulatory domain (No.)	Status before CAR-T cells infusion (No.)	Follow-up	aGvHD ≥ grade 2 (No.)	CR (%)	Survival
NCT00840853 Cruz et al[56]	8	38.5 (9–59)	MSD/MUD	Donor	CD19/CD28	Relapse (6) No relapse (2)	NR	0	37.5	NR
NCT01865617 Turtle et al[57]	11	29.5 (23–58)	MSD/MUD/HID/DUCB	Recipient	CD19/4-1BB	Relapse (7)MRD-positive (4)	NR	0	90.9	NR
NCT02028455 Gardner et al[51]	28	12.8 (1.3–23.2)	NR	Recipient	CD19/4-1BB	Relapse	NR	NR	93	NR
NCT01044069 Park et al[4]	19	NR	NR	Recipient	CD19/CD28	Relapse	NR	NR	84	NR
NCT03173417 NCT03050190 NCT02813837 Chen et al[58]	35	21 (2–55)	HID/MSD/MUD	Donor/ Recipient	CD19/4-1BB (26) CD19/CD28 (6) CD19/CD28 + 4-1BB (3)	Relapse	12.7 (1.5–41.5) months	17.6%	85.7	1 year CIR: 60.4% 1 year OS: 53.3%18 months CIR: 68.3% 18 months OS: 30.0%
ChiCTR-OOC-16008447 Zhang et al[59]	43	24 (4–60)	HID/MSD	Donor	CD19/CD28 (18) CD19/4-1BB (25)	Relapse (30) MRD-positive (13)	17 (6–47) months	0	79	1 year CIR: 57% 1 year EFS: 43%
NCT03327285 NCT04336501 Zhao et al[60]	11	39 (18–56)	HID/MSD	Donor	NR	MRD-positive	424.5 (105–1137) days	0	100	CIR: 42.8% PFS: 65.6% OS: 100%

aGvHD: Acute graft-versus-host disease; allo-HSCT: Allogeneic hematopoietic stem cell transplantation; CAR: Chimeric antigen receptor; CIR: Cumulative incidence of relapse; CR: Complete remission; DUCB: Double umbilical cord blood; EFS: Event-free survival; HID: Haploidentical donor; HSCT: Hematopoietic stem cell transplantation; MSD: Human leukocyte antigen (HLA)-matched sibling donor; MUD: HLA-identical unrelated donor; NR: Not reported; OS: Overall survival; PD: Progressive disease; PFS: Progression-free survival.

New Challenges for MRD Detection in CAR-T-Cell Therapy

Recently, clinical studies have indicated that a significant proportion of disease progression and even relapse persists after remission of CAR-T-cell therapy.^[4,61,62] In addition, immune escape and even clonal evolution events increase rapidly under high immune pressure.^[61] These dilemmas present new challenges and impose higher requirements for the existing MRD monitoring methods and strategies. Based on the immunophenotypes of relapse, recurrence can be roughly classified into two patterns: target-antigen-positive relapse and target-antigen-negative relapse.^[62]

Target-antigen-positive relapse

Antigen-positive relapse usually occurs early after CAR-T-cell therapy and constitutes the majority of relapses. Hay et al^[37] reported that the 2-year CIR was 34% for CD19-positive and 14% for CD19-negative relapse in patients with r/r B-sALL after CD19-CAR-T-cell infusion. In addition, they revealed that the lymphodepletion regimen and the magnitude of CAR-T-cell expansion or persistence were associated with relapse phenotypes. This type of recurrence is mainly related to rapid exhaustion and dysfunction of CAR-T-cells.^[63] Further research has suggested that the clinical status of patients (e.g., age, previous treatment, disease burden), CAR costimulatory domain, immunogenicity, and initial phenotypes of CAR-T-cells influence the ability of CAR-T-cells to recognize and kill tumor cells.^[45] In this setting, the immunophenotypes of tumor cells do not change generally after CAR-T-cell therapy, and it is still suitable to detect MRD by conventional FCM or PCR. However, it is necessary to pay attention to the possible subclones that may not be detected at initial diagnosis or new clones that may evolve during treatment, resulting in relapse. Therefore, a “Different from Normal” (DFN) approach or NGS technology with higher sensitivity is further needed to measure MRD in patients with target-antigen-positive relapse.

Target-antigen-negative relapse

Antigen-negative relapse often occurs in the middle or late stage after CAR-T-cell therapy. In B-ALL, CD19-negative relapse accounts for approximately 7%–22% of disease relapses.^{[61,51,64,65]} Furthermore, the loss of other target antigens, such as CD22 and B-cell maturation antigen (BCMA), has also been observed with the application of more CAR-T-cell therapies targeting different antigens.^[66,67] Relapse with negative antigen can arise from a variety of mechanisms, including mutation or alternative splicing of target antigen, lineage conversion, clonal evolution and so on.^[68–73]

Considering the target antigen could be a crucial indicator in the identification of tumor cells, it is essential to modify gating strategies when applying FCM to monitor MRD. Some studies have suggested extending FCM-MRD testing to other antigens expressed in early B-cell precursors (BCPs) besides CD19 to track B-ALL clones more sensitively.^[65,74–76] Cherian et al^[76] explored new gate markers in B-ALL MRD analysis after CD19-targeted therapy. They employed the DFN approach, defined the rough B-cell gate by identifying the expression of CD22 or CD24 (without CD66b), and combined it with markers aberrantly expressed in B-ALL, including CD10, CD20, CD34, CD38, and CD45, to differentiate abnormal B-lymphoid blast populations. This approach successfully identified CD19-positive and CD19-negative residual tumor cells at levels ranging from 0.003% to 98.8% of the white blood cells. However, the researchers found that both CD22 and CD24 were temporarily downregulated in one patient during CD19-CAR-T-cell therapy, resulting in a negative FCM outcome in the case of positive PCR. To address this issue, the utilization of cytoplasmic B-specific antigens like cytoplasmic CD79a could be beneficial in detecting the recurrence of CD19-negative B-ALL cells, since these antigens demonstrate a significant degree of lineage-specificity for B-cell differentiation.^[77] Mikhailova et al^[78] developed a single-tube, 11-color panel for FCM-MRD detection and chose the combination of CD22 and iCD79a as the main substitution for CD19, which proved to be a sensitive and reliable approach even in the case of possible CD19 loss. Similarly, Chen et al^[79] revealed that the MFC panel with CD79a gating can monitor MRD sensitively and predict the prognosis. Moreover, myeloid markers, such as CD33, CD65, and CD15, should also be considered in the setting of lineage switching, which has been described in patients with KMT2A rearrangements.^[80] In short, the DFN approach, which required more indicators labeled in a single tube and a more sophisticated analysis strategy, might be a better recommended FCM pattern in the age of CAR-T therapy. A more precise set of markers for predicting the presence of residual tumor cells is needed in the future with further studies in larger cohorts of patients.

PCR assays can provide reliable results regardless of the expression of the target antigen when clonal IGH/TCR gene rearrangements or fusing genes remain stable in tumor cells during treatment, which can help distinguish such normal very-early BCPs and leukemia cells with negative target antigen.^[81] However, PCR is unable to identify important changes in negative antigens, which could influence the effectiveness of subsequent treatment.^[82] Remarkably, clonal IGH/TCR gene rearrangements are not always consistent with the initial diagnosis if relapses are induced by lineage transformation or clonal evolution. For example, Gardner et al^[71] identified 7 B-ALL patients harboring rearrangement of mixed lineage leukemia (MLL) gene, and 2 patients among them developed acute myelocytic leukemia (AML) within 1 month of CD19-CAR-T-cell infusion. Specifically, both two patients maintained persistent MLL gene rearrangement during lineage conversion, while IGH gene rearrangement was not always conserved, bringing difficulty to PCR assays in MRD monitoring. In this regard, combining PCR assays and FCM could improve the sensitivity of MRD detection.

In comparison to FCM and PCR, NGS might be a better detection tool because of its ability to identify subclones and clone evolution. Consequently, it is recommended that the MRD status of clinically high-risk patients, especially those suspected of relapse with a negative antigen, be evaluated using NGS. Studies have demonstrated that clonotypic rearrangements utilized to monitor NGS-MRD remain consistent, even in instances of CD19 expression loss or lineage switch.^[30]

Conclusions and Prospectives

Future studies should focus on the following aspects to achieve MRD directed CAR-T-cell therapy or combination therapy: (1) Optimization and combination of current assessment methods. Efforts should be made to refine existing MRD monitoring techniques and explore the synergistic use of different methods to enhance the sensitivity of MRD detection. (2) Association between MRD status and patient prognosis. Further data is needed to establish a clearer understanding of the relationship between MRD status and the prognosis of patients with ALL. (3) Prospective clinical trials for MRD-guided personalized management. How MRD assessments guide individual and personalized management of patients with ALL requires further prospective clinical trials. (4) Evaluation of MRD as a surrogate endpoint. Exploring whether MRD can replace OS and DFS as a new surrogate endpoint of clinical trials would have significant clinical implications. Such an endpoint would accelerate the approval of drugs while reduces trial costs and better guides future treatment strategies for patients. (5) Exploring other promising MRD-directed interventions during CAR-T-cell therapy besides HSCT. Besides allo-HSCT, antibodies, immunomodulatory drugs, targeted therapies, and other preemptive strategies aimed at reducing or eliminating MRD during CAR-T-cell therapy seem to be a research direction worthy of further exploration since they have been successfully used to intervene in MRD peri-transplantation.

In conclusion, MRD monitoring has shown great promise in the field of CAR-T-cell therapy. Numerous studies have highlighted the high prognostic significance of pre- and post-treatment MRD status in ALL. In addition, continuous MRD assessment will contribute to better treatment strategies. MRD-guided combined therapy, involving the integration of CAR-T cell therapy with other approaches, has the potential to reduce the risk of relapse. Especially, consolidative allo-HSCT or preemptive CAR-T-cell infusion based on the MRD status could achieve better treatment outcomes. Although recurrence with negative target antigen in response to the selection immune pressures present challenges for MRD measurements in clinical practice, NGS and other novel detection techniques have shown promising potential.

Funding

This study was supported by grants from National Natural Science Foundation of China (Nos. 81870140 and 82070184), Peking University People’s Hospital Research and Development Funds (No. RDL2021-01), Beijing Nova Program (No. 20220484235), and Beijing Life Oasis Public Service Center (No. CARTFR-01).

Conflicts of interest

None.