This Week in Hematology — Aug 26, 2026

Welcome to This Week in Hematology. This week we're covering 10 notable papers spanning plasma cell disorders and risk stratification, myeloid disease genetics and transplantation, lymphoma and rare leukemia management, and one hard look at anticoagulation monitoring in pregnancy. Let's dive in.

We start with plasma cell disorders, where the headline is long-term survival. In Leukemia, Facon and colleagues report the final analysis of the MAIA trial, following 737 transplant-ineligible patients with newly diagnosed myeloma for a median of just over seven years. Median overall survival with daratumumab plus lenalidomide and dexamethasone reached 90 months, compared with 64 months with lenalidomide and dexamethasone alone — that's a reduction in the risk of death of about a third, and it translates into a median survival of seven and a half years in a population who are, by definition, older and frailer. Roughly half of patients on the triplet were alive at seven years, versus about four in ten on the doublet. Just as important for practice, the median time to needing the next line of therapy was never reached in the daratumumab arm, against about three and a half years in the control arm, and deaths from adverse events were essentially identical between the arms at around one in eight patients. This is now the benchmark for frontline therapy in the transplant-ineligible setting.

But if this population is living seven years, prognostication has to account for more than tumour biology, and that is exactly the argument made in the American Journal of Hematology by Katodritou and colleagues. They built the Myeloma Elderly Prognostic Score from 881 newly diagnosed non-transplant-eligible patients in Greece, using just four routinely available variables — age of 75 or older, creatinine clearance or estimated glomerular filtration rate below 40, an ECOG performance status of 2 or worse, and ultra-high-risk cytogenetics. The separation is stark: median overall survival ran from 96 months for patients with none of these features down to just over 15 months for those with three or more, and each step up the score was associated with roughly a 70 percent higher risk of death. It out-discriminated the R2-ISS in both the Greek cohort and an independent Portuguese validation set of 300 patients. The message is that host factors are doing a lot of the prognostic work in elderly myeloma, and a disease-only staging system is leaving information on the table. Staying with plasma cells, Hellou and colleagues, also in the American Journal of Hematology, looked at 517 patients with newly diagnosed AL amyloidosis at the Mayo Clinic who had baseline plasma cell FISH. Myeloma-defined high-risk cytogenetics were present in just under a third of patients, and gain of chromosome 1q accounted for about four out of five of those cases. Patients with 1q gain had a heavier clonal burden and markedly worse long-term survival — median around 57 months, versus 120 months in standard-risk patients — with the adverse effect emerging mainly beyond two years from diagnosis, when the risk of death nearly tripled. Notably, the other high-risk lesions like translocation 4;14 or deletion 17p were not independently prognostic here. So in AL amyloidosis, 1q is the abnormality worth flagging on the FISH report, and it argues for sustained clonal surveillance rather than reassurance after an early haematologic response.

Turning to myeloid disease, two papers refine who does well and with what. In the British Journal of Haematology, Chang and colleagues analysed 411 patients with CEBPA-mutated acute myeloid leukaemia across six Chinese centres, testing the 2022 European LeukemiaNet shift from double CEBPA mutation to in-frame basic leucine zipper mutations as the favourable-risk definition. Patients with double-mutant, in-frame bZIP disease — the large majority of the cohort — were younger, more often de novo, enriched for GATA2 and depleted of NPM1 and DNMT3A co-mutations, and did best, with a composite complete remission rate of about 95 percent and three-year overall survival of 80 percent. Importantly, among intensively treated patients, the double-mutant group did better than single-mutant bZIP in-frame disease, which suggests the current single-bucket favourable-risk category may still be biologically heterogeneous. Then in the American Journal of Hematology, Warraich and colleagues examined 111 patients with newly diagnosed acute myeloid leukaemia treated with frontline venetoclax plus a hypomethylating agent who went on to allogeneic transplant, median age 70. Median post-transplant survival was not reached, and three-year survival was 57 percent overall — but a three-factor model built on age of 65 or older, STAG2 mutation, and DNMT3A mutation split three-year survival from 89 percent in the lowest-risk group to 19 percent in the highest. Counterintuitively, older age and those two mutations predicted better survival, which is a reminder that in the venetoclax era the biology that responds to hypomethylating therapy may be the biology that stays in remission after transplant. And on the donor question, Avenoso and colleagues report an EBMT registry study, also in the American Journal of Hematology, of 275 adults transplanted in first remission for acute myeloid leukaemia on a uniform treosulfan and post-transplant cyclophosphamide platform. Two-year overall survival was essentially identical for haploidentical and 9-out-of-10 mismatched unrelated donors, at around 72 to 73 percent, with comparable relapse and non-relapse mortality. The one difference was chronic graft-versus-host disease, which was nearly three times more common after haploidentical transplant, though not in extensive forms. Within this platform, HLA disparity looks largely neutralised, which widens donor choice considerably.

In lymphoma and rare leukaemia, the theme is codified practice. The British Society for Haematology has published an updated guideline on primary mediastinal B-cell lymphoma in the British Journal of Haematology, led by Osborne and colleagues, and the headline change is a PET-adapted approach: patients achieving a complete metabolic response after dose-intensive chemoimmunotherapy can have consolidative radiotherapy omitted, with a defined framework for managing Deauville 4 and 5 residual disease, including CAR T-cell therapy and checkpoint inhibition. Alongside that, Blood carries the International Consensus Criteria for T-large granular lymphocytic leukaemia from Brammer and colleagues — the first uniform diagnostic, treatment-initiation, and response criteria for a disease that has been diagnosed and assessed inconsistently across centres, which has real implications for interpreting the growing number of trials. And filling a genuine evidence gap, Martin-Moro and colleagues in Blood Advances report real-world outcomes with zanubrutinib in 118 Spanish patients with relapsed or refractory marginal zone lymphoma — a median age of 75, two thirds with cardiovascular comorbidity, and about a quarter who would not have qualified for the MAGNOLIA trial. The overall response rate was 76 percent, with one-year progression-free survival of about 79 percent and overall survival of about 87 percent, and responses did not vary by subtype, age, or trial eligibility. Adverse events occurred in about a third of patients, grade 3 or higher in about one in seven, and bleeding was more frequent in those on anticoagulants. Responses were deeper with fewer prior lines, supporting earlier use.

Finally, a cautionary note on anticoagulation in pregnancy. In the Journal of Thrombosis and Haemostasis, Ashraf and colleagues systematically reviewed 94 studies of low molecular weight heparin dosing during pregnancy, comparing fixed, weight-based, and anti-factor-Xa-guided strategies. Pooled event rates for thromboembolic complications and major bleeding overlapped broadly across all three strategies, and in women with mechanical heart valves the risks were substantial — roughly 8 percent for thromboembolism and 8 percent for major bleeding. But the authors are explicit that with mostly single-arm data, small samples, and low to very low certainty of evidence, this is not evidence of equivalence. The honest reading is that we still do not know whether anti-Xa monitoring helps, and that mechanical valves in pregnancy remain a high-risk situation demanding specialist care.

If you only have time for one paper this week, make it the MAIA final analysis in Leukemia [1]. A median overall survival of seven and a half years in transplant-ineligible newly diagnosed myeloma resets the expectations you set with your older patients at diagnosis, and it settles the case for daratumumab-based triplet therapy up front.

Here are the key takeaways from this week in Hematology. Frontline daratumumab, lenalidomide and dexamethasone now delivers a median survival of seven and a half years in transplant-ineligible myeloma, with no excess in fatal adverse events. In that same older population, host factors — age, renal function and performance status — outperform disease-centred staging for predicting survival, so consider a composite score. In AL amyloidosis, gain of chromosome 1q is the cytogenetic lesion that matters, and its effect appears late, so keep watching the clone. In acute myeloid leukaemia, refining CEBPA and post-transplant risk by mutation profile is becoming actionable, and within a treosulfan plus post-transplant cyclophosphamide platform, haploidentical and mismatched unrelated donors give comparable survival, so donor availability need not limit you. And zanubrutinib works in real-world relapsed marginal zone lymphoma, including in patients trials would have excluded, but watch bleeding in those on anticoagulants.

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 think of one colleague — in any specialty — who never has time to keep up with the literature. Tell them about AudioScholar: a free ten-minute weekly for every specialty, to listen to in any podcast app, or to read at audioscholar dot C C.

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