Hematology

No survival differences by neighborhood deprivation or distance among CAR T recipients at a single center

In a retrospective single-center cohort of 124 B-NHL and 45 MM patients treated with CAR T, median ADI was 62.5 and median distance was 42.5 miles. Longer distance correlated with higher area deprivation (p<0.001), but OS and PFS did not differ by ADI or distance for B-NHL or MM. Dichotomizing the highest 15% ADI versus lower 85% likewise showed no outcome difference. Authors note referral and selection biases likely influence which patients reach the center, limiting generalizability. 1

Practical framework supports selective TDM and model‑informed dosing for some oral targeted agents

This narrative review synthesizes exposure–response and exposure–toxicity data for oral TKIs and PARP inhibitors and proposes a clinical TDM framework. High-level evidence supports exposure-guided dosing for imatinib and sunitinib with predefined trough targets and improved outcomes. Alectinib, cabozantinib, trametinib, and lenvatinib have consistent exposure relationships supporting selective implementation. For agents such as osimertinib, brigatinib, olaparib, and niraparib, monitoring is more useful for toxicity-driven management than routine efficacy optimization. 2

Bayesian networks show promise but need robust external validation before clinical use

Systematic review identified 52 oncology prognostic BN studies with median development cohort size 438 patients. Hybrid BN approaches (machine learning plus expert knowledge) often performed well and sometimes outperformed Cox models on validation sets. More than half the studies had unclear or high risk of bias and external validation was uncommon. Authors recommend high-quality external validation, prospective evaluation, and health-technology assessment before clinical deployment. 3