This Week in Hematology — Sep 2, 2026

Welcome to This Week in Hematology. This week we're covering 10 notable papers spanning three broad territories: curative and disease-modifying therapy in the haemoglobinopathies, immunotherapy-era risk stratification in acute lymphoblastic leukaemia, and how we choose salvage and transplant strategies in myeloid disease and chronic lymphocytic leukaemia. Let's dive in.

We'll start with sickle cell disease and thalassaemia, where three papers ask very different questions about how far we should push disease modification versus transplant. In the American Journal of Hematology, Bello-Manga and colleagues report HOPE Kids 2, a phase 3 double-blind placebo-controlled trial of voxelotor in 236 children aged two to under fifteen with sickle cell anaemia and conditional transcranial Doppler velocities between 170 and just under 200 centimetres per second, with about 83 percent of participants enrolled in sub-Saharan Africa [1]. The primary endpoint was met: cerebral blood flow velocity fell by roughly 12 centimetres per second on voxelotor compared with about 4 on placebo, a difference of nearly 8 centimetres per second that was sustained through week 48, and haemoglobin rose as expected. But the trial cannot be read on efficacy alone. Dosing was paused in May 2024 after an imbalance of deaths emerged, 8 in the voxelotor arm versus 2 on placebo, and the study closed entirely when all active voxelotor trials were discontinued in September 2024. Sickle cell crisis as a treatment-emergent event was also considerably more common on drug, close to 60 percent versus just under 40 percent. The investigators judged all deaths unrelated to study drug, and the authors frame their findings as proof that inhibiting haemoglobin S polymerisation can lower stroke-risk velocities, while calling for work on regional risk of crisis and mortality. For practice, voxelotor is off the market; what this paper gives you is a mechanistic signal for the future and a sobering reminder that a favourable surrogate endpoint and a favourable safety profile are not the same thing.

Against that backdrop, two papers push in the curative and the pragmatic directions for transfusion-dependent thalassaemia. In the British Journal of Haematology, Lai and colleagues report a phase 4 single-centre trial of haploidentical transplant in 134 patients with a median age of under nine years, all conditioned uniformly with busulfan, cyclophosphamide, fludarabine and anti-thymocyte globulin [9]. Two-year overall survival was about 93 percent and event-free survival about 93 percent, though graft-versus-host-disease-free relapse-free survival was lower at roughly 78 percent, with grade three to four acute graft-versus-host disease in about one in eight patients. Donor-specific antibodies, a male recipient with a female donor, and marked hepatomegaly all identified higher-risk patients — which is exactly the checklist to run before offering a haploidentical graft when no matched donor exists. At the other end of the resource spectrum, Blood Advances publishes a randomised non-inferiority trial from four Indian centres by Kakkar and colleagues comparing thalidomide at one versus two milligrams per kilogram per day in 188 patients aged twelve and over [10]. Overall response, defined as at least a 25 percent reduction in transfusion requirement at 24 weeks, was 58 percent, and here the lower dose actually did better, roughly 67 percent versus 50 percent. Only about one in eighteen patients achieved transfusion independence, discontinuation for toxicity ran at about 11 percent, and only about half of initial responders sustained benefit to week 72. So low-dose thalidomide is a legitimate option where transplant and gene therapy are out of reach, with the practical point that going higher buys you nothing.

Turning to paediatric and young adult acute lymphoblastic leukaemia, two cohorts converge on the same message: in the immunotherapy era, our old risk factors are losing their grip. In Blood Advances, Vieira Martins and colleagues analysed 225 children and young people across the United Kingdom and Ireland who received frontline blinatumomab in place of selected chemotherapy components because of intolerance or resistance [3]. Among the 195 who continued with chemotherapy afterwards, none of the conventional variables — age, presenting white cell count, high-risk genetics, even end-of-induction measurable residual disease before blinatumomab — predicted relapse. What did predict relapse was detectable residual disease at the end of the first blinatumomab cycle, which raised relapse risk more than sixfold, along with IKZF1-plus and DUX4 rearrangement. An integrated model using those three features carved out a high-risk third of the cohort with an 18 percent two-year relapse rate, against essentially zero in everyone else. That is a strong argument for measuring residual disease after cycle one of blinatumomab and letting that, not the pre-immunotherapy assessment, drive intensification. Complementing this, Leukemia publishes a European real-world series from Moser and colleagues covering 220 children and young adults given tisagenlecleucel for relapse after allogeneic transplant across 31 centres [4]. Two-year event-free survival was about 44 percent with overall survival around 67 percent, and CAR-T failure occurred in more than half of patients. Three factors marked out the highest-risk group: relapse within six months of transplant, where two-year survival was only about 47 percent versus roughly 74 percent for later relapse; a prior matched sibling donor graft, which unexpectedly carried the worst outcomes; and disease burden at lymphodepletion, with survival around 82 percent for measurable-residual-disease-negative patients falling to about 55 percent for those not in remission. The actionable piece is disease burden — cytoreduce before lymphodepletion where you can.

In acute myeloid leukaemia, three papers address treatment choice at diagnosis and at molecular relapse. The largest, in the American Journal of Hematology, is the PETHRATIFY analysis by Romero Riquelme and colleagues of 1,658 adults aged fourteen to eighty-five with newly diagnosed FLT3-mutated disease from 129 Spanish registry centres [5]. Composite complete remission was about ten points higher with midostaurin added to intensive chemotherapy, day-thirty mortality was roughly a third of that seen with chemotherapy alone, at about 2 percent versus 7 percent, and median overall survival was 47 months versus 19. Importantly, the authors are candid that in 261 propensity-matched pairs the effect attenuated — event-free survival remained significantly better, but overall survival became only a non-significant trend. Benefit held across mutation type, allelic burden and cytogenetic risk, and was smaller in NPM1 wild-type and secondary disease. Read it as real-world confirmation of standard of care rather than a new effect size. Where the news is genuinely neutral, Leukemia reports the COMMAND registry study from Badar and colleagues in TP53-mutant disease, comparing decitabine- with azacitidine-based induction with or without venetoclax in 321 of 652 patients [6]. Despite the theory that decitabine clears TP53 clones more deeply, propensity-matched analysis showed no difference in event-free or overall survival between any of the four regimens, with median survival in the six-to-nine-month range across the board. Multi-hit TP53 status, not agent choice, drove outcome. So choose your hypomethylating agent on logistics and tolerability, and keep steering these patients toward trials. Finally, in Blood Advances, Orvain and colleagues examined 121 adults with core-binding-factor or NPM1-mutated disease at first molecular relapse [7]. Across upfront allogeneic transplant, intensive chemotherapy, venetoclax plus azacitidine and other approaches, three-year survival ranged from about 64 to 84 percent with no statistically significant difference, and 81 percent of the whole cohort ultimately reached transplant. Upfront transplant is a reasonable option when feasible, but a bridging strategy with lower-intensity therapy does not appear to cost patients their transplant or their survival.

That leads into our last theme: who still needs an allograft, and how do we widen the donor pool. In the American Journal of Hematology, Shaffer and colleagues report the ACCESS trial expansion cohort — 193 adults receiving reduced-intensity or non-myeloablative HLA-mismatched unrelated donor peripheral blood transplants with post-transplant cyclophosphamide prophylaxis [2]. Median age was 65, a third of patients had a high comorbidity index, and a third had donors matched at less than seven of eight loci. One-year survival was about 80 percent, non-relapse mortality about 13 percent, relapse about 17 percent, and outcomes were the same for the more mismatched donors and consistent as the trial expanded from 13 centres to 33 — which is the generalisability signal that matters. The caution is infection: severe grade three to five infection occurred in about 39 percent of patients within a year. Conversely, in the British Journal of Haematology, Awan and colleagues compared ibrutinib with allogeneic transplant in relapsed or refractory chronic lymphocytic leukaemia with deletion 17p, and after adjustment ibrutinib was associated with roughly a 60 percent lower risk of death, along with better progression-free survival [8]. That is retrospective and confounded by selection, but it reinforces that in this genomic subgroup targeted therapy, not transplant, is the default first move.

If you only have time for one paper this week, make it HOPE Kids 2 in the American Journal of Hematology [1]. It is the definitive account of why voxelotor came off the market, and it is the paper you will need when a family asks what happened to the drug and what it means for their child's stroke risk.

Here are the key takeaways from this week in Hematology. First, voxelotor lowered conditional transcranial Doppler velocities in children with sickle cell disease, but with more crises and more deaths on drug — polymerisation inhibition remains a target, not a current option. Second, after frontline blinatumomab, measure residual disease at the end of cycle one and look for IKZF1-plus and DUX4; chemotherapy-era risk factors no longer stratify. Third, before CAR-T for post-transplant relapse, reduce disease burden and recognise early relapse as high risk. Fourth, in FLT3-mutated disease midostaurin remains standard of care across the adult age spectrum, while in TP53-mutant disease azacitidine and decitabine are interchangeable. And fifth, mismatched unrelated donor transplant with post-transplant cyclophosphamide and reduced-intensity conditioning is reproducible across many centres, with infection as the main price of admission.

That's your roundup for This Week in Hematology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if it saves you time, a quick rating in your podcast app helps other physicians find the show.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.