A nationwide Japanese cohort study of 1,311 adults with high-risk acute myeloid leukemia undergoing first single-unit cord blood transplantation found rapid, comparable neutrophil engraftment across methotrexate dose groups, with two-year overall survival of 48.5% and disease-free survival of 44.1%. Low-dose methotrexate was associated with more chronic graft-versus-host disease and poorer chronic GVHD-free and relapse-free survival in chemotherapy-based myeloablative and fludarabine-based reduced-intensity regimens, without reducing relapse. In total-body-irradiation-based myeloablative regimens, greater methotrexate exposure reduced grade II–IV acute GVHD and improved GVHD-free and relapse-free survival. Standard-dose methotrexate generally offered the best balance between GVHD prevention and survival in the other conditioning groups, supporting regimen-specific rather than uniformly reduced dosing.

Experiments in Arabidopsis identify histone deacetylase HDA19 as a regulator of histone H3.3 diacetylation at lysines 27 and 36. The mark declines in salt-stressed wild-type plants but remains elevated in hda19-deficient mutants. Mimicking persistent diacetylation increases late embryogenesis abundant (LEA) stress-response proteins and improves seedling salt tolerance, while removing key LEA proteins eliminates the mutants’ tolerance advantage. The findings link HDA19-mediated chromatin regulation to LEA protein accumulation and plant resilience under salinity.

A controlled One Health study analyzed 2,358 E. coli isolates collected in the Sapporo area of Japan during 2021–2022 from patients, healthy humans, companion animals, livestock, sewage, and river water. Whole-genome sequencing of 235 fluoroquinolone-resistant isolates identified 88 ST131 isolates, predominantly clade C1, with closely related strains shared across humans, companion animals, and sewage. Very small core-genome SNP distances between household pet-owner pairs and among human isolates suggest asymptomatic human carriage, human-to-human spread, and human–companion animal transmission. ST131 isolates carried multiple antimicrobial-resistance and virulence genes and were genetically similar to clinical strains, indicating a potential role in hospital-associated infections. Long-read sequencing also found highly similar blaCTX-M-27 ESBL plasmids across human, animal, and sewage isolates, suggesting that plasmid dissemination contributes alongside clonal expansion. The findings place Sapporo’s local transmission networks within the broader global expansion of ST131 and support community- and household-level interventions alongside hospital-focused antimicrobial-resistance control.

A large multi‑center Japanese cohort study evaluated whether having a history of solid tumors, and especially the types of treatments received for those tumors, influences outcomes after allogeneic hematopoietic stem cell transplantation. After adjusting for baseline differences using propensity score matching, overall survival, relapse, and non‑relapse mortality were similar between patients with and without prior solid tumors. However, patients previously treated with combined chemotherapy and radiation therapy showed significantly worse survival and higher non‑relapse mortality, driven mainly by post‑transplant complications. Recurrence of prior solid tumors after transplantation was rare. The study concludes that treatment modality—particularly combined chemo‑radiation exposure—should be incorporated into risk assessment for transplant eligibility, while most other prior solid tumor treatments do not adversely affect transplant outcomes.

Military clashes along the Cambodia-Thailand border have led to significant humanitarian challenges, with thousands seeking refuge in camps. The conflict, triggered by a shooting incident in disputed territories, resulted in numerous casualties and displacements. Despite a ceasefire, tensions remain high, with Cambodian soldiers detained by the Thai military. The article highlights the plight of those affected, emphasizing the desire for peace among the Cambodian people.

A retrospective analysis of 6,551 Japanese patients aged 16–59 with acute myeloid leukemia undergoing first allogeneic hematopoietic stem cell transplantation found that FLU/BU4 and conventional myeloablative conditioning produced similar survival in patients younger than 50. Among patients aged 50–59, FLU/BU4 improved three-year overall survival from 42.8% to 47.0%, largely through lower treatment-related mortality without increasing relapse. The benefit was concentrated in patients transplanted while in complete remission, in whom FLU/BU4 was associated with better adjusted survival and lower treatment-related mortality; no meaningful advantage was observed in non-remission patients. The authors suggest FLU/BU4 may be preferred for transplant-eligible AML patients aged 50–59 in complete remission, while noting that the retrospective Japanese design and variation in busulfan administration limit generalizability.

Analysis of more than 3,400 adult acute myeloid leukemia patients undergoing cord blood transplantation in Japan shows that myeloablative conditioning delivers better survival, lower relapse, and stronger engraftment outcomes than reduced-intensity conditioning, particularly for recipients under 55. Although myeloablative regimens improve disease control across all ages, transplant-related mortality rises sharply in patients over 55 due to higher rates of acute graft-versus-host disease and treatment toxicity. For patients aged 56–60, reduced-intensity conditioning offers similar overall survival with less toxicity regardless of risk factors. The study identifies age as the strongest determinant in selecting conditioning intensity and establishes 55 as a practical threshold for optimizing outcomes in cord blood transplantation for AML.

The study developed and evaluated a machine learning approach for predicting Timed Up and Go test performance in cancer patients using inertial measurement unit data captured during routine in-clinic-like movements. From 23 body sensor keypoints, pelvic velocity variability emerged as the most informative feature, enabling a linear regression model to achieve a mean absolute error of 1.322 seconds and a correlation of 0.713. The findings indicate that a single-sensor setup near the pelvis can effectively estimate physical function, potentially reducing clinical resource demands. Limitations include a small and homogenous cohort and lack of external validation. Future directions include broader patient populations, integration of video and audio features, and longitudinal prediction of functional decline.

Analysis of over 4,600 adult acute lymphoblastic leukemia patients from a nationwide Japanese registry showed substantial heterogeneity in the effectiveness of intensified myeloablative conditioning prior to allogeneic stem cell transplant. A machine learning Bayesian causal forest model identified high-benefit patient subgroups who gained improved disease-free survival and reduced one-year mortality without increasing non-relapse mortality. The high-benefit targeting approach outperformed traditional strategies based on age or disease risk index, demonstrating a roughly six-percentage-point reduction in one‑year mortality. Younger patients, males, CMV‑negative individuals, T‑cell phenotype, and higher disease risk were more frequently represented in the high-benefit group. Results suggest machine‑learning‑guided individualized conditioning decisions may optimize transplant outcomes compared with conventional methods, though external validation and incorporation of additional clinical variables remain necessary.

The article reviews research on applying AI and machine learning to predict objective physical function metrics such as SPPB, TUG, grip strength, and walking speed using inputs from sensors, videos, images, and clinical tabular data. Current physical assessments are informative but difficult to implement routinely due to space, time, equipment, and staffing constraints. Studies increasingly employ AI to estimate these measures without requiring dedicated testing, potentially enabling frequent, automated monitoring in clinical settings. Walking speed prediction is the most extensively studied area, while grip strength prediction is less common due to its ease of manual measurement. Sensor‑based input, especially IMUs, remains dominant, though video and image‑based approaches are emerging and may offer non‑contact alternatives. Most reviewed studies use small datasets and lack external validation, underscoring the need for larger, standardized datasets. AI‑enabled assessment could support personalized treatment decisions, particularly for older cancer patients, by enabling earlier detection of functional decline.