Executive Overview
For decades, disease relapse following allogeneic hematopoietic cell transplantation (allo-HCT) has represented one of the most formidable and heartbreaking hurdles in modern hematology. When acute leukemia or myelodysplastic syndromes (MDS) return after a transplant, treating physicians and their patients are thrust into a high-stakes clinical dilemma. Historically, standard salvage avenues have split down two primary paths: therapeutic donor lymphocyte infusion (DLI), designed to harness an immune-mediated graft-versus-leukemia effect, or a high-risk second allogeneic hematopoietic cell transplantation (allo-HCT).
Despite the widespread utilization of both strategies across global cancer centers, direct, high-quality comparative data evaluating these interventions from the exact moment of relapse have remained surprisingly scarce. Clinicians have long lacked definitive guidance, often forced to navigate this treacherous therapeutic landscape based on institutional preference, patient performance status, and anecdotal experience rather than robust empirical proof.
To address this critical knowledge gap, researchers at the University Medical Center Göttingen conducted a comprehensive, single-center retrospective cohort study evaluating 85 adult patients who experienced relapse of acute leukemia or MDS between 2010 and 2022. Published in Frontiers in Hematology, the study utilized sophisticated biostatistical methodologies—including propensity score matching (PSM), restricted mean survival time (RMST), and Fine-Gray competing risks models—to perform a rigorous head-to-head evaluation of therapeutic DLI versus a second allo-HCT.
The findings challenge conventional dogmas in post-transplant care. The data revealed no statistically significant difference in overall survival (OS) between patients undergoing a second allo-HCT and those receiving therapeutic DLI. While secondary transplantations offered superior disease control with a lower cumulative incidence of relapse-related mortality (RRM), this advantage was entirely counterbalanced by a trend toward higher non-relapse mortality (NRM). Consequently, the overall survival curves converged, demonstrating that neither strategy inherently secures a survival superiority. These insights underscore the imperative for individualized, patient-centric treatment planning rather than a one-size-fits-all salvage paradigm.
Detailed Chronology and Study Methodology
To understand the weight of these findings, one must examine the methodological rigor applied by the study’s authors (Alexander C. Angleitner, Moritz Maulhardt, Jan Hasenkamp, Jörg Büntzel, and Gerhard Wulf). Traditional analyses of post-relapse interventions have often been muddied by "immortal time bias"—a statistical distortion that occurs when the time elapsing between relapse and the administration of salvage therapy is improperly accounted for, artificially inflating the survival metrics of the intervention group.
To circumvent this, the Göttingen team established the exact date of relapse as time zero for all patients, ensuring a realistic, "real-world" estimation of clinical outcomes from the moment disease recurrence was documented.
Cohort Breakdown and Patient Selection
The retrospective cohort comprised 85 adult patients diagnosed with acute myeloid leukemia (AML), secondary AML (sAML), myelodysplastic syndromes (MDS), or acute lymphoblastic leukemia (ALL):
- Therapeutic DLI Group: 60 patients (71%) received DLI following overt hematologic relapse or clinical progression. Patients receiving DLI exclusively for declining donor chimerism or isolated minimal residual disease (MRD) positivity were intentionally excluded, ensuring the cohort strictly represented therapeutic intervention in active disease states.
- Second allo-HCT Group: 25 patients (29%) underwent a second stem cell transplantation.
In terms of baseline characteristics, patients selected for a second allo-HCT typically exhibited favorable performance status, an absence of uncontrolled graft-versus-host disease (GvHD) from their initial transplant, and were deemed fit enough to endure intensive cytoreductive conditioning protocols. Conversely, therapeutic DLI was more frequently directed toward patients with advanced relapse, lower physiological reserve, or those for whom a second major transplant was clinically unfeasible.
The baseline profile of the cohort reflected advanced, high-risk disease:
- Diagnosis: AML accounted for the majority of cases (63%), followed by sAML (24%), MDS (14%), and ALL (6%).
- Risk Stratification: Among patients with available molecular data, 69% presented with high-risk genetic profiles (such as TP53 mutations).
- Time to Intervention: Patients proceeding to a second allo-HCT experienced a significantly longer median interval from relapse to treatment compared to those receiving DLI (74.0 days versus 42.5 days, $p < 0.001$), reflecting the extensive time required for donor searches, workups, and intensive pre-transplant salvage chemotherapy.
Salvage Therapy and Treatment Protocols
The paths leading to cellular therapy varied starkly between the two groups. Patients bound for a second allo-HCT predominantly underwent intensive salvage chemotherapy regimens—such as FLAG-based protocols, HAM, ICE, or fludarabine-cytarabine combinations—frequently achieving temporary remission before conditioning. In contrast, the DLI cohort largely received lower-intensity interventions, including hypomethylating agent (HMA)-based therapies with or without venetologax, targeted therapies, or no intervening salvage therapy whatsoever.
For the second transplant procedures, reduced-intensity conditioning (RIC) was favored in 92% of cases, utilizing matched unrelated donors as the primary source, alongside matched related, mismatched unrelated, and haploidentical configurations. DLI protocols, meanwhile, involved stepwise dose escalations starting at $0.5 times 10^6$ CD3+ cells/kg, progressing up to $1 times 10^7$ cells/kg across a maximum of five scheduled infusions, provided active GvHD did not supervene.
Supporting Context and Empirical Metrics
The primary endpoint of the study was overall survival, tracked from the moment of relapse until death from any cause. The empirical results yielded critical quantitative insights into the limits of current post-transplant salvage modalities.
Overall Survival Analysis
Across the entire unadjusted cohort of 85 patients, the median overall survival was 0.92 years (95% CI: 0.71–1.54) in the second allo-HCT group compared to 0.61 years (95% CI: 0.45–0.94) in the DLI group. Despite a numerical advantage favoring the second transplant, this difference failed to achieve statistical significance (log-rank $p = 0.295$).
To minimize confounding variables, the researchers performed a 1:1 Propensity Score Matching (PSM) based on ECOG performance status and the time interval from the first allo-HCT to relapse. In the resulting matched cohort of 36 patients (18 per arm):
- Median OS: 1.01 years (95% CI: 0.60–1.80) for second allo-HCT versus 0.49 years (95% CI: 0.30–2.16) for therapeutic DLI.
- Statistical Significance: The survival curves remained statistically indistinguishable (log-rank $p = 0.519$), characterized by wide, overlapping confidence intervals that confirmed the absence of a definitive survival edge for either modality.
Furthermore, restricted mean survival time (RMST) analyses at 2 years corroborated these findings, showing a restricted mean survival of 0.95 years for DLI patients versus 1.04 years for secondary transplant recipients.
Relapse-Related Mortality (RRM) Versus Non-Relapse Mortality (NRM)
The most illuminating facet of the study emerged from the competing-risks regression models, which decoupled overall mortality into its underlying drivers: death due to active, recurring malignancy versus death resulting from treatment toxicities.
- At 1 Year:
- RRM was notably higher in the DLI group (43.3%) compared to the second allo-HCT group (24.4%), illustrating the superior anti-leukemic disease control afforded by a fresh, myeloablative or reduced-intensity re-transplantation.
- NRM inverted this trend, presenting a lower cumulative incidence in the DLI group (20.3%) versus a substantial 34.5% in the second allo-HCT group.
- At 2 Years:
- Cumulative RRM climbed to 50.4% for DLI and 38.3% for second allo-HCT.
- Cumulative NRM rose to 24.0% for DLI and 40.5% for second allo-HCT.
Fine-Gray competing risks regression confirmed that while a second allo-HCT successfully suppressed the subdistribution hazard (sHR) for relapse-related mortality (sHR = 0.63), this benefit was entirely neutralized by an increased hazard for non-relapse mortality (sHR = 1.81, driven primarily by severe infections, sepsis, and transplant-related organ failure).
Independent Predictors of Survival
Multivariate Cox proportional hazards modeling identified key independent prognostic factors governing survival post-relapse:
- ECOG Performance Status: Exhibited a strong trend toward significance (HR = 1.32, $p = 0.072$), confirming that baseline functional fitness dictates salvage capacity.
- Time to Relapse: Longer latency periods between the initial transplant and disease recurrence showed a trend toward protective association with overall survival (HR = 0.87 per year, $p = 0.058$).
- Conditioning Intensity: Myeloablative conditioning during the first allo-HCT demonstrated an initial univariate association with improved overall survival (HR = 0.58, $p = 0.027$), though it did not retain independent significance in the fully adjusted multivariate model.
Official Statements and Ethical Compliance
The study was conducted in strict compliance with the ethical principles outlined in the Declaration of Helsinki. Formal approval was granted by the institutional ethics committee of the University Medical Center Göttingen (Approval Number: 12/8/25). Due to the retrospective nature of the study and the utilization of rigorously anonymized clinical records, the requirement for individual patient informed consent was formally waived by the reviewing committee.
In their official statements, the study authors emphasized the pragmatic implications of their research for everyday clinical oncology:
"We did not detect an overall survival difference between second allo-HCT and therapeutic DLI, including after propensity score matching for key confounders such as performance status and time to relapse," stated lead author Alexander C. Angleitner. "Given the high intensity and profound toxicity profile associated with a second allogeneic transplantation, treatment selection must remain strictly individualized. Clinicians must weigh performance status, remission duration, and patient preferences, as our data confirm no universal overall survival benefit for either strategy."
The research team transparently acknowledged the inherent limitations of their work, including its retrospective, single-center design, the modest sample size within matched subsets, and biological heterogeneity across distinct disease subtypes (AML, sAML, MDS, ALL). Furthermore, the investigators noted that generative AI tools were utilized exclusively for language editing and grammar refinement, with all scientific content, data curation, and clinical interpretations remaining under the direct, uncompromised responsibility of the human authors.
Future Outlook and Clinical Takeaways
The findings published by Angleitner and colleagues signal a vital maturation in the management of post-transplant relapse. For decades, oncologists have wrestled with the temptation to escalate therapy aggressively at the first sign of disease recurrence, assuming that a second allogeneic transplant represents a definitive rescue.
This study provides a sobering corrective: while a second allo-HCT undeniably packs a harder punch against retreating leukemia cells (effectively cutting 1-year relapse-related mortality nearly in half), its heavy toll in non-relapse mortality—driven by opportunistic infections, organ failure, and treatment-related toxicity—wipes out any statistical gain in overall survival.
Moving forward, the oncology community must pivot away from dogmatic treatment escalation and toward precision-guided salvage algorithms. Key imperatives for future research include:
- Prospective Multicenter Trials: Validating these single-center observations in large-scale, prospective cohorts designed to eliminate immortal time bias entirely.
- Biomarker-Driven Selection: Utilizing advanced measurable residual disease (MRD) tracking, single-cell genomics, and immune-reconstitution profiling to identify the rare patient subsets who genuinely derive long-term curative benefit from secondary transplantation versus those who are better served by lower-toxicity cellular therapies or novel immunotherapies (such as CAR-T cell therapy or bispecific antibodies).
- Optimizing Pre-Relapse Interventions: Emphasizing preemptive strategies—such as early intervention during molecular or micro-chimerism decline—before overt, treatment-resistant hematologic relapse takes root.
Until such prospective data mature, clinicians facing the devastating reality of post-transplant relapse are urged to step back from rigid treatment algorithms. Treatment must be tailored to the individual, balancing the aggressive biology of the recurring malignancy against the patient’s physiological resilience, personal goals of care, and the grim reality that survival outcomes remain unacceptably modest regardless of the cellular rescue route chosen.
