Maintenance strategies for relapse prevention and treatment Free

Posttransplant relapse after allogeneic stem cell transplantation (alloSCT) requires a highly personalized management plan based on the time from transplant to relapse, donor type, molecular features of the primary disease, history of graft-versus-host disease (GVHD), and other posttransplant complications.¹ Monitoring for evidence of early relapse and dropping donor hematopoiesis (ie, chimerism) can enable early intervention, though questions remain regarding which approaches are beneficial. Treatment in the posttransplant setting must strike a balance between the efficacy and toxicities of chemotherapies and targeted agents and the promotion of the graft-versus-leukemia (GVL) effect while avoiding excessive GVHD.

Here, we discuss recent evidence that bears on strategies to prevent and treat posttransplant relapse, using the framework provided by 3 clinical cases: 1) maintenance therapy for relapse prevention, 2) salvage therapy for relapse, and 3) second alloSCT (HSCT2) and other cellular therapies.

A 30-year-old man with Philadelphia chromosome-positive acute lymphoblastic leukemia (Ph⁺ ALL) with central nervous system involvement underwent haploidentical alloSCT with a total body irradiation-based myeloablative conditioning regimen and postgraft cyclophosphamide (PTCy). His bone marrow (BM) biopsy at posttransplant day 30 showed complete remission (CR) with negative BCR-ABL and 100% donor chimerism.

Maintenance therapy² is a low-intensity treatment in the early posttransplant period, when relapse is most likely to occur, that aims to prolong remission and overall survival. Growing evidence suggests that some types of maintenance therapy may lead to improved outcomes in patients with high-risk diseases. However, the potential benefits of maintenance therapy for disease control must be balanced against the risks of hematologic toxicities and GVHD (Table 1).

The FMS-like tyrosine kinase 3 (FLT3) inhibitors sorafenib and gilteritinib have each been studied as posttransplant maintenance therapies for FLT3-mutated acute myeloid leukmia (AML). A 5-year follow-up of a phase 3 randomized trial showed improved overall survival (OS) at 5 years (72% vs 57%) and lower cumulative incidence of relapse (CIR) (15% vs 36%) in patients who received 1 year of sorafenib maintenance compared to controls.³ The 5-year cumulative incidence of chronic GVHD (cGVHD) was similar in both groups. Interestingly, another randomized trial comparing 2 years of maintenance gilteritinib to placebo only showed a relapse-free survival (RFS) benefit (hazard ratio [HR] 0.515) among patients with detectable measurable residual disease (MRD) in the peritransplant period.⁴ 

In IDH1-mutant AML, a phase 1 study of ivosidenib maintenance for 1 year showed that ivosidenib was safe and well-tolerated.⁵ Efficacy outcomes were promising, with 2-year progression-free survival (PFS) of 81% and 2-year OS of 88%. For IDH2-mutant AML, enasidenib⁶ was well-tolerated as posttransplant maintenance therapy, with a 2-year PFS of 69% and 2-year OS of 74%.

In chronic myeloid leukemia (CML), a large registry study showed no benefit of posttransplant maintenance with a tyrosine kinase inhibitor (TKI), perhaps because most patients were beyond first complete remission (CR1) and potentially TKI- refractory.⁷ Conversely, TKI maintenance posttransplant has been suggested as the standard of care for Ph+ ALL based on a European Society for Blood and Marrow Transportation (EBMT) registry analysis that reported OS (HR 0.42) and leukemia-free survival (HR 0.44) benefits with TKI maintenance.⁸ For Ph+ ALL with a T315I mutation, ponatinib maintenance needs further investigation.⁹ 

Hypomethylating agent (HMA) monotherapy has been proposed as a maintenance therapy. In a retrospective, matched control study, patients with high-risk AML and myelodysplastic syndromes (MDS) had lower 3-year CIR rates after azacitidine maintenance compared to controls.¹⁰ However, a subsequent randomized trial failed to demonstrate a survival benefit, though the study may have been underpowered due to better outcomes in both groups.¹¹ 

Venetoclax is another potential candidate for maintenance therapy. A prospective study of AML patients¹² reported a 1-year OS of over 70% and 1-year RFS of 67% in patients who received venetoclax maintenance for 1 year. As combination therapy, maintenance venetoclax with low-dose azacitidine was recently evaluated in a phase 1 trial; 96% of this cohort had detectable pretransplant MRD,¹³ and 2-year OS and CIR were reported as 67% and 41%, respectively. The ongoing VIALE-T trial (NCT04161885), a phase 3 randomized trial, is evaluating the azacitidine/venetoclax combination as posttransplant maintenance therapy in AML patients.

Immunotherapies have also been tested in the posttransplant setting. As posttransplant maintenance, inotuzumab ozogamicin was safe with no cases of veno-occlusive disease in a phase 1 trial in Ph+ ALL.¹⁴ One-year nonrelapse mortality (NRM) and PFS were 6% and 89%, respectively. Posttransplant blinatumomab maintenance was also well-tolerated in high-risk ALL patients with a 1-year PFS of 71% and NRM of 0%.¹⁵ 

Posttransplant cytokine therapies have also been explored as strategies to enhance GVL. In a prospective trial in alloSCT recipients, prophylactic ultra-low dose interleukin-2 (IL-2) was well tolerated and promoted T_reg and natural killer (NK) cell expansion while preserving antigen-specific T-cells.¹⁶ A phase 1/2 trial showed that prophylactic pegylated interferon-α (IFN-α) was safe in alloSCT recipients with very high-risk AML, with a CIR of 42% and NRM of 13% at 6 months.¹⁷ 

The patient started posttransplant maintenance with dasatinib on day 30. He developed steroid-refractory gastrointestinal GVHD and was treated with systemic steroids and ruxolitinib, which was further complicated by invasive fungal infection requiring life-long posaconazole, leading to dasatinib dose adjustment. He discontinued dasatinib 2 years after alloSCT and remains in CR for 3 years after alloSCT.

A 44-year-old woman with NPM1+, FLT3 internal tandem duplication (FLT3-ITD)+ AML in an MRD-negative CR received a myeloablative human leukocyte antigen (HLA)-matched unrelated donor alloSCT. On day 30, myeloid chimerism was 100%, but T-cell chimerism was 79% and remained <80% despite tapering systemic immunosuppression. NPM1 transcripts became detectable 6 months posttransplant, and overt relapse developed after 10 months with 25% NPM1+ blasts, FLT3-ITD-negative by polymerase chain reaction, and CD3 chimerism of 78%.

Growing evidence suggests that posttransplant MRD monitoring can predict clinical outcomes. The premise behind such monitoring is that early low-level relapse will be more successfully treated than overt relapse. In the FIGARO trial,¹⁸ posttransplant MRD-positivity and mixed donor T-cell chimerism were independently associated with poor OS. Early complete myeloid-lineage (CD33⁺) donor chimerism before day 60 was reported to correlate with lower relapse rates.¹⁹ This observation supports the importance of monitoring both MRD and donor chimerism. The optimal MRD monitoring method differs based on the underlying disease and genetic abnormalities.²⁰ We recommend checking patient-specific MRD assays from BM and blood lineage-specific chimerism on days 30, 100, 180, and 360 after alloSCT.

At the time of overt AML relapse, we recommend performing next-generation sequencing to search for targetable mutations (ie, FLT3, IDH1). For ALL, checking for CD19 or CD22 expression is critical for selecting immunotherapies. HLA loss of heterozygosity (LOH) at chromosome 6, which contains HLA genes, is observed in up to one-third of patients who relapse after haploidentical alloSCT and is another important feature to evaluate at relapse, as it impacts the utility of donor lymphocyte infusions (DLIs) and donor selection for HSCT2.^21,22 Commercial assays for this are under development.

There is no FDA-approved or consensus approach for treating posttransplant relapse. Thus, treatment choices are commonly made based on prior therapies, targetable mutations, and patient fitness (Table 2).

For myeloid malignancies, HMA monotherapy achieves a 9%-20% CR rate, with median survival of only 3-6 months.^23–26 The combination of HMA with venetoclax does not offer substantial improvement over monotherapy, with a 30% CR rate and median survival of only 6 months.^27–31 Although not yet documented in the posttransplant relapse setting, treatment with HMA and venetoclax may be reasonable for patients with IDH1/2 mutations, as this subgroup showed favorable outcomes in the frontline setting.³² Higher-intensity regimens containing fludarabine, cytarabine, and idarubicin, with or without venetoclax, have also been used for younger and fit patients. However, even these only yielded a CR in up to 36% of patients, and responses were not durable, with a median survival of 6 months.^27,33–35 Given these disappointing outcomes, the benefits of high-intensity cytotoxic therapies must be weighed against their serious side effects, including end-organ damage, prolonged cytopenias, opportunistic infections, and early mortality rates as high as 10%.²⁷ 

Targeted therapies can be applied to selected patients with actionable mutations. TKIs are effective salvage therapies for chronic myeloid leukemia that has relapsed after alloSCT (3-year adjusted OS of 54% when used as monotherapy).³⁶ Ivosidenib and enasidenib were well tolerated without GVHD^37,38 in alloSCT recipients; however, they have not been prospectively studied specifically for posttransplant relapse. FLT3 inhibitors were reported to enhance GVL through IL-15 induction^39,40 and reduce expression of exhaustion markers in alloreactive CD8 T-cells⁴⁰ in preclinical models. Retrospective studies reported that sorafenib is tolerable, with CR rates of 38% to 53% and 1-year OS of 22% to 66%.^41,42 Revumenib, a menin inhibitor, showed clinical efficacy with 23%-30% CR rates in relapsed or refractory lysine methyltransferase 2A (KMT2A) rearranged or NPM1-mutant AML, including the patients relapsing after alloSCT.^43,44 

DLIs were initially introduced and found to be successful in treating chronic phase CML relapse.⁴⁵ However, for high-risk diseases (AML, ALL, MDS, and blast-phase CML), DLI monotherapy rarely controls disease and is therefore commonly combined with salvage chemotherapies. Several clinical trials have explored the role of prophylactic or preemptive DLI, especially with in vivo or ex vivo T cell depletion, but whether this reduces relapse is inconclusive due to a lack of prospective controlled studies.⁴⁶ Nevertheless, large retrospective registry data consistently support the observation that long-term survival is only achieved in patients who receive DLI or HSCT2.^34,47 Thus, DLI should at least be considered as a treatment option in combination with cytoreductive salvage therapies or other experimental approaches.

Although immunotherapy has been proposed as a strategy to promote GVL, the GVHD risk, which is related to the robust activation of alloreactive T-cells, can be significant. A phase 1/2b trial reported acceptable safety outcomes after ipilimumab for posttransplant AML relapse, with 14% incidence of GVHD and 21% incidence of immune-related adverse events (irAEs); the overall response rate (ORR) was 31%, predominantly effective in extramedullary AML. However, subsequent multicenter trials of combined ipilimumab and decitabine showed minimal clinical effects (ORR 20%) and relatively high incidences of GVHD (36%) and irAEs (16%). In general, the use of checkpoint inhibitors in the posttransplant setting is limited to Hodgkin's disease (CR 40%-50%),⁴⁸ as the risk of GVHD outweighs the benefit for other hematologic malignancies.

Lenalidomide posttransplant maintenance as monotherapy caused excessive acute GVHD (aGVHD) (60% grade III-IV GVHD within 2 cycles of therapy).⁴⁹ Lenalidomide was then tested in combination with azacitidine in the phase 1 VIOLA trial, which showed only a 10% rate of aGVHD.⁵⁰ The subsequent phase 2 Azalena trial confirmed the safety of lenalidomide and azacitidine when combined with DLI, with an ORR of 56% and median OS of 21 months in relapsed AML/MDS after alloSCT.⁵¹ 

IFN-α has been used to potentiate the effect of DLI in CML.⁵² In a recent phase 1/2 trial of pegylated interferon-2α in combination with DLI, GVHD incidence was relatively high (62%) with 2-year OS and PFS of 31% and 24%, respectively.⁵³ Patients who developed GVHD survived longer than patients without GVHD (median OS 284 vs 43 days), suggesting a link between GVHD and GVL in this approach.

IFN-γ has been suggested as a rational strategy for potentiating GVL in myeloblastic leukemias based on preclinical studies showing that IFN-γ restores HLA expression on myeloid blasts after alloSCT.^54,55,56 A phase 2 trial evaluating the safety of IFN-γ and DLI for posttransplant relapse is ongoing (NCT06529731).

Bi-specific T-cell engagers (BiTEs) are approved to treat B-cell malignancies. Blinatumomab was reported to safely salvage patients with relapsed B-cell acute lymphoblastic leukemia (B-ALL) after alloSCT, achieving a 45% CR rate within 2 cycles and a 36% OS at 1 year.

The patient received reinduction therapy with fludarabine and cytarabine and achieved an MRD-negative CR. She received DLI with recovery of donor T-cell chimerism. She subsequently developed mild skin and genital chronic GVHD, which was controlled with topical therapies. She has remained in NPM1-negative CR for 2 years.

A 36-year-old man with high-risk MDS with monosomy 7 and an EZH2 mutation (R-IPSS score 5.2) received a haploidentical alloSCT with myeloablative conditioning and PTCy. His BM showed detectable MRD by flow cytometry at 6 months after transplant. He received decitabine and venetoclax for 5 cycles; treatment was held for 3 months due to preseptal cellulitis and fungal pneumonia. A subsequent BM showed evolution to AML with 20% myeloblasts and new mutations (ASXL1, NF1, and GATA2) with 81% donor chimerism. He was treated with CPX (cytarabine and daunorubicin) liposome-351; a posttreatment BM sample had 2.2% residual abnormal blasts with 97% donor chimerism.

HSCT2 is a salvage option for physically fit patients, particularly those who have relapsed more than 6 months after the first alloSCT and achieved CR after salvage therapy with no prior high-grade GVHD.^57,58 A retrospective study reported that AML/MDS patients who underwent HSCT2 had significantly higher 5-year OS rates (26%) than those who did not (7%).⁵⁹ Moreover, survival after HSCT2 has improved over the past 2 decades, particularly in younger patients, whose 2-year OS is reported to be around 30%.⁶⁰ 

The benefit of using the same or a different donor for HSCT2 is still unclear. Theoretically, a different HLA-matched donor could target different minor histocompatibility antigens. Large cohort studies, however, showed comparable outcomes in HSCT2 using the same donors vs different donors.^58,61 Switching to a different haploidentical donor enables the targeting of mismatched HLA molecules, which could enable GVL.⁶² A retrospective registry analysis did not reveal the clinical benefits of using a new haploidentical donor in HSCT2. Rather, a switch was associated with higher NRM.⁶³ Using HLA-LOH information to guide optimal donor selection for HSCT2 is biologically reasonable, so prospective studies are required to validate the utility of HLA-LOH for second donor selection.

Donor-derived adoptive cell therapies have a potential role in preventing or treating posttransplant relapse. Donor-derived CD19 CAR-Ts (chimeric antigen receptor T-cells) induced deep and durable remissions without GVHD in relapsed B-ALL.^64,65 When infused prophylactically, donor-derived anti-WT1 T-cell receptor-transduced T-cells (TCR-T) achieved 100% relapse-free survival at 3 years.⁶⁶ Donor-derived multiple leukemia antigen-specific T-cells (mLSTs) were safely infused as adjuvant therapy prophylactically or for relapsed patients who achieved CR after salvage therapy with 2-year OS of 77%.⁶⁷ More recently, TCR-T targeting the minor histocompatibility antigen HA-1 TCR-T was tested in patients with relapsed leukemia/MDS occurring after alloSCT; this was safe with a preliminary efficacy signal.⁶⁸ Clinical trials testing a wide spectrum of donor-derived adoptive cellular therapies are expected to launch in the coming years (Table 3).

The patient received HSCT2 from a haploidentical donor who does not share haplotypes with his original donor using a myeloablative conditioning regimen with PTCy. He has been in MRD-negative CR with full-donor chimerism for 6 months without GVHD.

Current evidence is insufficient to provide optimal posttransplant approaches for every patient. Prospective, well-controlled studies are critically needed to investigate new strategies for risk-adapted prophylaxis or effective salvage therapies for posttransplant relapse. Clinical trials should always be considered for patients at high risk for or who have overt relapse.

Emily Geramita: no competing financial interests to declare.

Jing-Zhou Hou: no competing financial interests to declare.

Warren D. Shlomchik is a co-founder, option holder, and paid consultant for Bluesphere Bio. Warren D. Shlomchik is also an option holder and consultant for Orca Bio.

Sawa Ito has received the research funding from BlueSphere Bio.

Emily Geramita: Nothing to disclose.

Jing-Zhou Hou: Nothing to disclose.

Warren D. Shlomchik: Nothing to disclose.

Sawa Ito: Nothing to disclose.