This Week in Hematology — Aug 19, 2026

Welcome to This Week in Hematology. This week we're covering 10 notable papers spanning bone marrow failure and myeloproliferative disease, blood-based disease monitoring in myeloma and lymphoma, and real-world outcomes with the targeted and cellular therapies we now use every day. Let's dive in.

Let's start with two randomised trials in Blood that both challenge a treatment ceiling. The EMAA trial, led by Höchsmann and colleagues, is the first placebo-controlled phase three study in moderate aplastic anemia, a group for whom no consensus first-line therapy has existed. Eighty-five patients with clinically significant cytopenia were randomised to cyclosporine A with either eltrombopag or placebo, and among the seventy-five patients evaluable at week twenty-four, the overall response rate was about seventy-one percent with eltrombopag versus roughly forty-three percent with cyclosporine alone — a statistically significant difference on the primary endpoint [1]. What makes this trial practically useful is what happened after unblinding. Placebo-arm patients who had not achieved a complete response were crossed over to eltrombopag, and that group reached an overall response rate of about seventy-four percent by week forty-eight, again a significant improvement. Relapse rate and failure-free survival did not differ between arms, and there were no new safety signals. So the actionable message is that adding eltrombopag up front improves trilineage response in moderate disease, and adding it later still rescues a substantial number OF PATIENTS who fail cyclosporine monotherapy. Also in Blood, Palmer and colleagues report the phase 2a IMPRSSION study of sapablursen in polycythemia vera — an antisense oligonucleotide that suppresses TMPRSS6, raises hepcidin, and restricts iron availability for erythropoiesis. Forty-nine phlebotomy-dependent patients received subcutaneous dosing every four weeks. Over the efficacy window, the median number of phlebotomies fell from five at baseline to zero in the higher-dose cohort and about one and a half in the forty-milligram cohort [2]. Symptom burden on the MPN symptom assessment score improved significantly in the higher-dose cohort but not in the lower-dose cohort. Adverse events were mostly mild or moderate, with anemia and fatigue the most common — unsurprising given the mechanism. This is early-phase data, but it points toward hematocrit control without cytoreduction or repeated venesection.

Now to a theme that is quietly reshaping follow-up: can we monitor disease from blood instead of marrow? In Blood, Dejoie and colleagues used the structured sampling schedule of the CASSIOPEIA trial to compare serum MALDI-TOF mass spectrometry against bone marrow measurable residual disease by next-generation flow and next-generation sequencing in 237 myeloma patients with paired samples from post-induction through maintenance [3]. At every time point, negativity by any of the three methods predicted longer progression-free survival, with broadly comparable landmark performance. Agreement between blood-based mass spectrometry and marrow improved during maintenance, with strong positive and negative predictive value against a ten-to-the-minus-five marrow reference — but rule-out performance was more limited against sequencing at ten-to-the-minus-six. Sustained negativity over one and two years identified favourable-outcome patients regardless of platform. The authors frame this as complementary and phase-adapted: use blood for frequent non-invasive monitoring during maintenance, and reach for marrow when you genuinely need maximum sensitivity. A parallel story comes from Haematologica, where Gould and colleagues report real-world use of a commercially available immunoglobulin-rearrangement circulating tumour DNA assay in large B-cell lymphoma [9]. Median turnaround from sample collection to result was nine days, which matters for feasibility. After frontline immunochemotherapy in 102 patients, twelve-month progression-free survival was about fifteen percent when residual disease was detectable versus about eighty-five percent when it was undetectable. That is a stark separation, and it was seen in routine practice rather than a trial-embedded biobank. The caveat is that we still lack randomised evidence that acting on a positive result changes outcome — so for now this is prognostic information, not yet a treatment trigger.

Our third theme is real-world confirmation, and occasional refinement, of therapies we already prescribe. In the British Journal of Haematology, the GIMEMA AML2320 study prospectively followed 193 newly diagnosed patients unfit for intensive chemotherapy who received venetoclax with azacitidine or decitabine [4]. Median age was seventy-four, and over four in ten patients were seventy-five or older. Composite complete remission was reached in about three quarters OF PATIENTS by the end of cycle four, and median overall survival was thirteen months. Two findings deserve attention: achieving remission within four cycles was associated with roughly double the median survival, about nineteen months versus nine; and patients who received the full four-hundred-milligram venetoclax dose without concomitant azoles did better than those on reduced doses with azoles, about eighteen months versus eleven. A matching-adjusted indirect comparison with VIALE-A showed essentially identical median survival. That azole finding is observational and confounded by infection risk, but it should at least prompt you to revisit whether prophylactic azoles and dose reduction are automatic in every patient. Staying with acute myeloid leukaemia, Kunadt and colleagues in the same journal report 156 relapsed or refractory FLT3-mutated patients treated with gilteritinib across twenty-five centres — a heavily pretreated group in which more than half had relapsed after a prior transplant and nearly half had prior FLT3 inhibitor exposure [7]. The remission rate was about forty-five percent, with median overall survival of ten months. The sequencing signal is the important part: patients treated with gilteritinib alone had one-year survival of about thirty-four percent, whereas those who proceeded to allogeneic transplant had one-year survival between roughly seventy-six and ninety-two percent depending on whether gilteritinib was resumed afterwards, and resuming it post-transplant produced a remission rate of about eighty-six percent. Selection bias obviously favours the transplanted group, but the practical framing holds — treat gilteritinib as a bridge and then as maintenance, not as a destination.

In myeloma relapse, EMN Italy contributes 306 patients who progressed on lenalidomide maintenance after autologous transplant, a population increasingly common and poorly characterised [8]. Most received isatuximab, carfilzomib and dexamethasone, and anti-CD38-based combinations outperformed the alternatives, with isatuximab-carfilzomib-dexamethasone achieving the longest progression-free survival at nearly twenty-three months. High lactate dehydrogenase, higher stage, and relapse within twelve months of starting maintenance all predicted worse progression-free and overall survival. This effectively documents the pre-BCMA standard of care against which newer immunotherapies will be judged.

Finally, three papers on immunotherapy delivery. From Haematologica, Rocha and colleagues analysed 1,191 EBMT-registered patients aged seventy or older receiving CAR T-cell therapy for large B-cell lymphoma, comparing tisagenlecleucel with axicabtagene ciloleucel, with median follow-up of about twenty-three months [6]. Progression-free survival favoured axicabtagene ciloleucel, at the cost of more cytokine release syndrome at day thirty. For a seventy-four-year-old, that is a real trade-off to discuss explicitly rather than defaulting on age alone. In the British Journal of Haematology, Townsend and colleagues report phase 1b glofitamab added to obinutuzumab or rituximab plus CHOP in thirty-one patients with relapsed or refractory B-cell non-Hodgkin lymphoma [5]. The complete metabolic response rate was about eighty-four percent and median progression-free survival was not reached after forty-four months, but toxicity was substantial — grade three or four events in roughly four in five patients, cytokine release syndrome in about half, and four fatal events, mostly infections. And the phase 2 DISTINKT study of danburstotug, an anti-PD-L1 antibody, in relapsed or refractory extranodal natural killer T-cell lymphoma reported an objective response rate of about seventy-nine percent with complete responses in roughly two thirds OF PATIENTS, and median progression-free survival of about twenty-nine months in a very small cohort of twenty-three [10]. In a disease with few options, that durability is noteworthy, and PD-L1 membrane specificity looks like a candidate predictive biomarker.

If you only have time for one paper this week, make it the EMAA trial in Blood [1]. It is the first placebo-controlled randomised evidence in moderate aplastic anemia, a setting where practice has been entirely opinion-driven, and it gives you both a first-line answer and a salvage strategy.

Here are the key takeaways from this week in Hematology. First, in moderate aplastic anemia, adding eltrombopag to cyclosporine improves trilineage response, and late addition still rescues non-responders. Second, blood-based monitoring is maturing on two fronts — mass spectrometry in myeloma and immunoglobulin-based circulating tumour DNA in large B-cell lymphoma both carry prognostic weight, but neither yet replaces marrow at maximum sensitivity, and neither has randomised evidence that acting on a positive result helps. Third, in unfit acute myeloid leukaemia, early remission on venetoclax plus a hypomethylating agent predicts longer survival, and reflexive azole-driven dose reduction deserves scrutiny. Fourth, in relapsed FLT3-mutated disease, gilteritinib works best as a bridge to transplant with post-transplant resumption. And fifth, in patients over seventy receiving CAR T-cells, axicabtagene ciloleucel gave better progression-free survival but more cytokine release syndrome than tisagenlecleucel — a discussion to have patient by patient.

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 a quick rating on Apple Podcasts or Spotify helps other physicians discover the show.

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