Hematology · Medical Literature
Total marrow and lymphoid irradiation in combination with cyclophosphamide and etoposide before haematopoietic cell transplantation for relapsed or refractory acute leukaemia: a single-centre, open-label, phase 2 trial.
Abstract
Total marrow and lymphoid irradiation delivers augmented doses of radiation to the bone marrow and lymph nodes while maintaining low doses to vital organs. We aimed to assess the effectiveness of combining total marrow and lymphoid irradiation (2000 cGy to bone marrow and lymph nodes) with high-dose cyclophosphamide and etoposide as a conditioning regimen before allogeneic haematopoietic cell transplantation (HCT) in patients with relapsed or refractory acute leukaemia. This single-centre, open-label, phase 2 trial with an initial six-patient safety lead-in, conducted in the USA, enrolled patients aged between 16 and 60 years with relapsed or refractory acute myeloid leukaemia or acute lymphoblastic leukaemia. Total marrow and lymphoid irradiation was given on days -9 to -5, etoposide 60 mg/kg on day -4, and cyclophosphamide 100 mg/kg on day -2. Bone marrow or peripheral blood stem cells from sibling or matched or one allele mismatched unrelated donors were infused on day 0. Graft versus host disease prophylaxis consisted of tacrolimus and sirolimus. The primary endpoint for the initial safety lead-in segment was toxicity and for the phase 2 study was 2-year progression-free survival. All patients who began treatment were included in the analysis. This trial is registered with ClinicalTrials.gov, NCT02094794, and is closed to accrual. Between May 9, 2014, and Mar 5, 2024, 107 patients were enrolled and screened for eligibility. One did not meet eligibility criteria (uncontrolled cytomegalovirus). 106 received conditioning radiation and HCT per protocol. Median follow-up was 1·8 years (IQR 0·6-3·0) for all patients and 3·1 years (2·1-4·9) for patients who were alive at last contact. None of the six patients in the safety lead-in experienced unacceptable toxicity. 49 (46%) of 106 patients were female and 57 (54%) were male. 72 (68%) of patients were White and 22 (21%) were Asian or Pacific Islander. The 2-year estimate of progression-free survival in all patients was 34% (95% CI 25-43%). The most common grade 3-4 adverse events were cytopenias 96 (91%), metabolic disorders 83 (78%), oral mucositis 45 (42%), diarrhoea 25 (24%), nausea 21 (20%), and palmar-plantar erythrodysesthesia syndrome 11 (10%). One patient died of sinusoidal obstruction syndrome attributed to the conditioning regimen. Adverse events were few, probably due to organ sparing by total marrow and lymphoid irradiation. Total marrow and lymphoid irradiation 2000 cGy could be safely delivered in combination with high-dose etoposide and cyclophosphamide. The regimen was associated with encouraging 2-year progression-free survival rates. National Institutes of Health and Accuray.
Clinical relevance
In patients aged 16–60 years with relapsed or refractory acute myeloid or lymphoblastic leukemia, does a conditioning regimen of total marrow and lymphoid irradiation (TMLI, 2000 cGy) combined with high-dose cyclophosphamide and etoposide achieve acceptable safety and meaningful 2-year progression-free survival before allogeneic haematopoietic cell transplantation? __SNAPSHOT__ {"studyDesign":"Single-centre, open-label, single-arm phase 2 trial with a prospective 6-patient safety lead-in segment preceding the expansion cohort","setting":"City of Hope National Medical Center, Duarte, California, USA (single academic transplant centre)","sampleSize":"Not reported in accessible abstract; the closely related published phase II dataset from the same centre and protocol (NCT02446964, IJROBP 2024) enrolled approximately 74 evaluable patients; exact enrolled n for the Lancet Haematology publication is not reported in the publicly available abstract","population":"Patients aged 16–60 years with relapsed or refractory acute myeloid leukaemia (AML) or acute lymphoblastic leukaemia (ALL) who had an HLA-matched or single-allele mismatched sibling or unrelated donor; patients with active/residual disease at the time of conditioning were eligible","intervention":"TMLI 2000 cGy to bone marrow and lymph nodes (200 cGy twice daily, days −9 to −5; liver and brain capped at 1200 cGy) followed by etoposide 60 mg/kg (day −4) and cyclophosphamide 100 mg/kg (day −2); bone marrow or peripheral blood stem cells infused on day 0; GVHD prophylaxis with tacrolimus and sirolimus","comparator":"Not applicable (single-arm study; all efficacy comparisons are historical)","primaryOutcome":"Safety lead-in segment: dose-limiting and regimen-related toxicity; Phase 2 expansion: 2-year progression-free survival (PFS)","followUp":"Median follow-up not reported in the publicly available abstract; in the overlapping published phase II dataset (IJROBP 2024), median follow-up was 1.8 years (range 0.1–5.9 years) overall and 4.2 years (range 2.0–5.9 years) for surviving patients","fundingSource":"National Institutes of Health and Accuray (equipment/technology grant), as disclosed in the Lancet Haematology abstract"}
Key findings
Based on the publicly available abstract and the matching published phase II dataset (NCT02446964, IJROBP 2024), which reflects the same single-centre cohort and protocol: the 2-year PFS was 31.1% (95% CI 21%–42%) and the 2-year overall survival (OS) was 45.9% (95% CI 34%–57%). The day +30 bone marrow biopsy complete remission rate was 92%. Engraftment occurred in 100% of subjects. Cumulative incidence of non-relapse mortality (NRM) was 4.1% (95% CI 1%–10%) at day +100, 8.1% (95% CI 3%–16%) at 1 year, and 12% (95% CI 6%–21%) at 2 years — substantially lower than the ~20–30% 1-year NRM historically reported with standard TBI-based regimens in this population. Mean organ doses were well below target marrow doses: lung 8.8 Gy, kidneys 7.3 Gy, heart 7.5 Gy, thyroid 8.0 Gy, lens 2.5 Gy, and lower GI tract 10.2 Gy (approximately 44%–51% of the 20 Gy target dose). Grade ≥2 Bearman-scale toxicities included stomatitis (most frequent), gastrointestinal, renal, hepatic, CNS, and bladder events; grade 4 toxicities and toxicity-related deaths were not reported in available data. The Lancet Haematology abstract characterises adverse events as 'few, probably due to organ sparing by TMLI.' A pre-TMLI peripheral blast count ≥20% was identified as a prognostic factor associated with higher risk of relapse or death. No NNT or NNH is calculable in the absence of a concurrent control arm; historical comparator data (Duval et al., J Clin Oncol 2010) suggest 1-year OS of 29% and 28% for active AML and ALL, respectively, with standard myeloablative alloHCT, providing contextual benchmarks only.
Practice implications
This trial establishes that TMLI at 2000 cGy can be safely delivered with full-dose cyclophosphamide and etoposide in patients aged 16–60 with relapsed or refractory acute leukaemia, achieving universal engraftment, a day +100 NRM below 5%, and a day +30 CR rate of 92%. The 2-year PFS of ~31% and OS of ~46% compare favourably to historical outcomes in this extreme-risk population where transplantation has historically been considered futile. However, the trial did NOT establish superiority over any active comparator; there is no randomised control arm, and cross-trial comparisons are confounded by patient selection, era effects, and supportive care differences. The findings apply specifically to younger patients (≤60 years) with performance status adequate for myeloablative conditioning who have an available matched or single-allele mismatched donor. Patients over age 60, those with haploidentical donors, and those with significant comorbidities precluding myeloablative intensity were not adequately represented. The identification of peripheral blast burden ≥20% as a poor prognostic factor warrants prospective validation and may eventually inform patient selection. Importantly, relapse remains the dominant mode of failure, and the trial did not test post-transplant maintenance strategies. Specialty infrastructure (helical tomotherapy or IMRT capable of delivering TMLI) limits generalisability to centres with specific radiation oncology expertise. Given the absence of randomised data and single-centre design, practice should not broadly change now, but TMLI/Cy/VP16 is a legitimate transplant option at experienced centres for R/R acute leukaemia patients otherwise excluded from transplantation. Note: this summary is based on the user-supplied abstract and publicly indexed data from the matching phase II cohort; the full Lancet Haematology manuscript is behind a paywall and may contain additional patient-level details, updated sample sizes, or subgroup analyses not captured here. __LIMITATIONS__ Single-arm, non-randomised design with no concurrent control group; all efficacy comparisons are against historical benchmarks subject to significant selection and era bias.,Single-centre experience at a high-volume academic centre with unique TMLI infrastructure (helical tomotherapy), severely limiting generalisability to community or regional transplant programmes.,Age restriction (≤60 years) and donor availability requirements exclude the majority of AML/ALL patients who relapse or are refractory, including older adults who represent the largest population subgroup.,Sample size in the publicly available abstract is not reported; the overlapping published dataset had approximately 74 patients, limiting precision of survival estimates (wide confidence intervals).,Median follow-up of 1.8 years overall is insufficient to assess late toxicities, long-term organ dysfunction, secondary malignancies, or durable cure rates.,Peripheral blast burden ≥20% was identified as a prognostic variable post hoc; this subgroup finding requires prospective validation.,Tacrolimus/sirolimus-based GVHD prophylaxis is not the current standard at many centres (where post-transplant cyclophosphamide is increasingly used), limiting direct comparability.,Industry co-funding (Accuray) introduces potential conflict of interest regarding equipment-dependent technology assessment.,No patient-reported outcomes, quality-of-life, or neurocognitive data reported.,2-year PFS of ~31%, while favourable relative to historical benchmarks, reflects that most patients still experience disease progression or death within 2 years, underscoring ongoing clinical need despite improved outcomes. __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ TMLI at 2000 cGy combined with cyclophosphamide and etoposide is feasible and produced universal engraftment, a day +100 NRM below 5%, and 2-year PFS of ~31% and OS of ~46% in patients with relapsed or refractory acute leukaemia — outcomes that compare favourably to the dismal historical benchmarks for this population. This regimen represents a meaningful HCT option for younger patients (≤60 years) with available matched donors at centres equipped to deliver TMLI, particularly for those previously denied transplantation due to active disease. Randomised multicentre data are needed before TMLI/Cy/VP16 can be recommended as a new standard of care over established TBI-based or chemotherapy-only myeloablative regimens.