This Week in Hematology — Sep 30, 2026

This AudioScholar briefing is generated by artificial intelligence for healthcare professionals and trainees. It is not medical advice.

Welcome to This Week in Hematology. This week we're covering 10 notable papers spanning lymphoid malignancy therapy and prognostication, thrombosis and antibody-mediated coagulopathy, and a new generation of prediction tools built on ordinary blood tests. Let's dive in.

We start with lymphoid malignancies, where two papers push at the frontline and the relapsed setting from opposite ends. In Blood, Luskin and colleagues report a phase 1 trial combining the BCL-2 inhibitor venetoclax with standard-dose inotuzumab ozogamicin in adults with relapsed or refractory CD22-positive B-cell acute lymphoblastic leukemia and lymphoblastic lymphoma. Twenty-three patients enrolled, and the recommended dose settled at 400 milligrams of venetoclax daily for 21 days per cycle. Of 22 evaluable patients, 21 achieved complete remission, most of them after a single cycle, and measurable residual disease cleared by flow cytometry in close to nine out of ten of those assessed. Two-year disease-free and overall survival were both 48 percent, with just under two thirds of patients proceeding to transplant. The important safety signal is sinusoidal obstruction syndrome, which occurred in four patients, about one in six, a familiar hazard with inotuzumab that this combination does not abolish. Correlative work pointed to acquired MCL-1 dependence, CD22 antigen escape and drug efflux as routes of progression. This is a single-arm phase 1 study in a small cohort, so it establishes feasibility and a striking remission rate rather than comparative benefit, but it supports randomized evaluation of the combination as a bridge to transplant.

Staying with lymphoma, British Journal of Haematology publishes the largest real-world comparison yet of the two new frontline regimens for advanced-stage Hodgkin lymphoma. Chohan and colleagues assembled 646 patients across 17 United States centres, around four fifths treated with brentuximab vedotin plus AVD and the remainder with nivolumab plus AVD. The toxicity profiles diverged sharply and predictably. Neuropathy affected well over half of the brentuximab group versus about one in five of those given nivolumab, and dose reductions or omissions were more than twice as common with brentuximab. Nivolumab, on the other hand, brought substantially more neutropenia, affecting roughly three quarters of patients at any grade, and more infections overall, though severe infections and febrile neutropenia were comparable. On efficacy, two-year overall survival was 97 percent with brentuximab and 100 percent with nivolumab, and two-year progression-free survival was 84 versus 88 percent. Neither difference reached statistical significance, and that is worth flagging, because the authors describe the numerically improved progression-free survival as supporting nivolumab-AVD as a frontline standard. What these data firmly show is comparable survival at just over two years of follow-up, with a genuine trade-off between neuropathy and myelosuppression, which is a conversation about patient characteristics rather than a settled hierarchy.

At the other end of the age spectrum, also in British Journal of Haematology, Dubois and colleagues address a group that every existing prognostic index handles badly: patients over 80 with large B-cell lymphoma treated with curative intent. Using 438 trial patients treated with anti-CD20 plus mini-CHOP, they built the SENIOR-IPI from four readily available variables — the number of age-adjusted IPI predictors, lactate dehydrogenase above three times the upper limit of normal, albumin below 35 grams per litre, and the number of years over 80. In an independent real-world validation cohort of 195 patients it outperformed the standard IPI, the age-adjusted IPI and the NCCN-IPI for both progression-free and overall survival, with discrimination for overall survival rising from around 0.6 to 0.67. That is a modest but consistent gain, achieved without any new test, and it gives clinicians a better-calibrated way to discuss expected outcomes with the very elderly, though it has not been prospectively tested as a treatment-allocation tool.

Two papers this week deal with states that sit between health and disease. In Leukemia, Kewan and colleagues characterise clonal monocytosis of undetermined significance and its cytopenic counterpart, drawing on 958 patients with sustained monocytosis after stringent exclusion of overt myeloid neoplasm. The 74 patients meeting precursor criteria looked genomically distinct from oligomonocytic chronic myelomonocytic leukemia, which was enriched for TET2, ASXL1 and SRSF2 mutations with higher variant burdens, whereas the precursor states more often carried DNMT3A and PPM1D. Clinically, over a median of about two and a half years there were no progression events among patients with clonal monocytosis alone, while roughly one in eight of those with accompanying cytopenias progressed, almost always to chronic myelomonocytic leukemia. Overall survival was significantly worse in oligomonocytic disease. The message the authors draw is that the cytopenia, not the monocytosis, carries the risk, and that morphology and genomics need to be read together. In Blood Advances, El Jurdi and colleagues report a randomized, double-blind, placebo-controlled phase 2 trial of clobetasol 0.05 percent as an aqueous oral solution for oral chronic graft-versus-host disease. Among the 35 subjects reaching day 28, the overall response rate was 91 percent, with about a fifth achieving complete response, including patients who had already failed clobetasol ointment, and paired mucosal biopsies showed reduced effector and proliferating T cells alongside falls in salivary interleukin-6 and CXCL10. The caution is systemic: cosyntropin testing at day 28 demonstrated new adrenal insufficiency in 14 patients not otherwise on supraphysiologic steroids. So a topical therapy with measurable tissue-level activity, but not one that is free of systemic glucocorticoid exposure.

Turning to thrombosis and antibody-mediated coagulopathy, three papers converge on mechanism and salvage. In Blood, Hulshof and colleagues interrogate why VEXAS syndrome is so thrombogenic, studying 40 patients with defined pathogenic UBA1 variants, just over half of whom had experienced thrombotic events. They found elevated von Willebrand factor and factor VIII, raised von Willebrand factor propeptide and angiopoietin-2 consistent with Weibel-Palade body exocytosis, accumulation of high molecular weight multimers, elevated soluble VCAM-1 and thrombomodulin, and both increased thrombin generation and an activated protein C resistance phenotype. This is a mechanistic cohort study, not a therapeutic one, but it reframes VEXAS thrombosis as an endotheliopathy and raises the von Willebrand-ADAMTS13 axis as a candidate target for future study. In the Journal of Thrombosis and Haemostasis, Thaler and colleagues report seven patients with autoantibody-mediated coagulation disorders — two with acquired von Willebrand disease, two with acquired factor XI deficiency and three with catastrophic antiphospholipid syndrome — treated with the BCMA-directed bispecific teclistamab on the logic that long-lived plasma cells escape conventional B-cell depletion. All seven achieved complete remission, at a median of 24 weeks, with bleeding cessation or antibody disappearance as appropriate, and toxicity limited mainly to grade 1 cytokine release syndrome, with all patients receiving immunoglobulin replacement. Seven retrospective cases cannot establish efficacy, and the authors say exactly that, but as a proof of concept in refractory disease it is striking.

Also in the Journal of Thrombosis and Haemostasis, Singh and colleagues examine intravenous immunoglobulin in heparin-induced thrombocytopenia across a decade of single-network practice. Immunoglobulin was reserved for the sickest patients, with a median platelet nadir of 26 and thrombosis present at diagnosis in about three quarters of them. Platelet recovery occurred in 93 percent at a median of six days from the first dose. In a propensity-matched comparison, 24 patients given immunoglobulin within three days of diagnosis recovered platelets faster than matched controls, a median of three days versus five, roughly doubling the rate of recovery, but there was no difference in thrombosis, bleeding or mortality. This is retrospective, single-network and small, so it supports the hypothesis that early administration accelerates platelet recovery without yet showing that it changes hard outcomes.

Finally, two prediction papers. In British Journal of Haematology, Christensen and colleagues built CBC-HEMA from full blood counts, age, sex and C-reactive protein in more than 850,000 Danish primary care patients, achieving an area under the curve of 0.81 for any hematological malignancy within six months. Prediction was excellent for chronic lymphocytic leukemia and myeloproliferative disease but poor for non-Hodgkin lymphoma and plasma cell dyscrasia, and performance degraded markedly in UK Biobank, which the authors read as evidence that the tool is unfit for screening healthy populations. In Thrombosis and Haemostasis, Fehlmann and colleagues pooled just over 5,000 emergency department patients with suspected pulmonary embolism from four prospective European studies and tested a two-step machine learning algorithm. The best model missed about 1.4 percent of pulmonary embolisms, keeping the upper bound below the prespecified 2 percent safety threshold, ruled out disease with no testing at all in roughly one in ten patients, and produced an imaging rate of about 54 percent — essentially matching the 4PEPS rule and better than PERC. The honest conclusion is parity, not superiority, on retrospective data, and prospective external validation is still required.

If you only have time for one paper this week, make it the multicentre real-world comparison of nivolumab-AVD and brentuximab vedotin-AVD in advanced-stage Hodgkin lymphoma in British Journal of Haematology [2]. It is the first large comparative look at two regimens now competing for the same frontline slot, and it reframes the choice as a toxicity trade-off rather than an efficacy one, since the survival differences did not reach significance.

Here is what this week's evidence adds up to in Hematology. First, in relapsed B-cell ALL, venetoclax added to inotuzumab produced complete remission in nearly every evaluable patient in a phase 1 trial, with deep measurable residual disease clearance but a persistent sinusoidal obstruction syndrome signal, and it now needs randomized testing [1]. Second, in advanced Hodgkin lymphoma the two modern regimens look equivalent for survival at just over two years in real-world practice, with neuropathy weighted toward brentuximab and neutropenia and infection toward nivolumab [2]. Third, risk stratification is getting more granular at both ends of the disease spectrum, with a validated prognostic index for large B-cell lymphoma in patients over 80 [4] and evidence that clonal monocytosis without cytopenia carries essentially no short-term progression risk while the cytopenic form does [5]. Fourth, targeting plasma cells with a BCMA bispecific achieved remission in all seven patients with refractory autoantibody-mediated coagulopathy — hypothesis-generating, not practice-defining [6] — while mechanistic work in Blood positions VEXAS thrombosis as an endothelial disease [7]. And fifth, machine learning applied to routine bloods and to suspected pulmonary embolism performed respectably but has not yet outperformed existing clinical rules, and both models degraded outside their development settings [8][10].

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.