Anjo Kosei Hospital

This study aimed to identify patient groups in which myeloablative conditioning (MAC) or reduced-intensity conditioning (RIC) regimens induced superior progression-free survival (PFS) in patients with acute myeloid leukemia (AML) in complete remission (CR) using a machine-learning approach. Our study included 3273 patients aged 40–69 with AML in CR. The patients were divided into training (N = 2020) and validation cohorts (N = 1253). We employed a machine learning-based group identification model in the training cohort. Subsequently, in the validation cohort, we estimated the impact of the optimal conditioning group compared with the non-optimal conditioning group on PFS using an inverse probability weight analysis. The developed model was consistent with the eight factors and combinations, and the high score suggested that RIC was more appropriate than MAC. In the validation cohort, 127 patients with high scores and who received RIC and 769 patients with low scores and who received MAC were categorized into the optimal conditioning group (896, 71.5%). The weighted hazard ratio for PFS was 0.73 (95% confidence interval: 0.57–0.94) in the optimal conditioning group compared with the non-optimal conditioning group (P = 0.016). In conclusion, we developed an easy-to-use model that helps the physician choose a patient-specific conditioning regimen for patients with AML in CR.

Elevated systolic blood pressure (SBP) increases myocardial oxygen demand, whereas low diastolic blood pressure (DBP) can impair coronary perfusion. The prognostic impact of this high-SBP/low-DBP (HSLD) profile in patients with ST-elevation myocardial infarction (STEMI) remains unknown. We analyzed 696 consecutive patients with STEMI undergoing primary percutaneous coronary intervention (PCI) from a prospective 11-center registry in Japan. Patients were categorized based on pre-procedural SBP (≥ 120 mmHg) and DBP (< 70 mmHg) into four groups: high-SBP/low-DBP (HSLD) high-SBP/high-DBP (HSHD), low-SBP/high-DBP (LSHD), and low-SBP/low-DBP (LSLD). The primary endpoint was a 1-year composite of cardiovascular death, non-fatal myocardial infarction, or stroke. Median follow-up was 369 days. The 1-year cumulative incidence of the primary endpoint was 7 (9.5%) in the HSLD group, 19 (4.2%) in the HSHD group, 2 (2.8%) in the LSHD group, and 5 (6.0%) in the LSLD group (log-rank p = 0.007). In multivariable Cox regression, the HSLD was independently associated with the primary endpoint compared to the HSHD group (hazard ratio 3.20, 95% CI 1.04-9.87; p = 0.043). The prognostic value of HSLD was consistent across major subgroups. A pre-procedural blood pressure pattern of SBP ≥ 120 mmHg combined with DBP < 70 mmHg independently predicts a higher risk of adverse cardiovascular events at 1 year in patients with STEMI. This simple measurement at admission identifies a previously unrecognized high-risk phenotype and presents a valuable opportunity for early risk stratification.