CELL INTELLIGENCE · 肿瘤细胞治疗研究

BACKGROUND AND OBJECTIVES: Treatment decision-making in relapsed/refractory diffuse large B-cell lymphoma (R/R DLBCL) is increasingly complex, especially with the emergence of novel, high-cost therapies such as chimeric antigen receptor (CAR) T-cell therapy. Understanding how physicians weigh various clinical and nonclinical treatment attributes is essential for aligning healthcare decisions with both clinical value and economic sustainability. This study aims to quantify Japanese physicians' preferences in health outcomes and patient experience outcomes through eliciting trade-offs in a choice experiment for specific treatment characteristics in R/R DLBCL. METHODS: We performed a discrete choice experiment (DCE) with Japanese hematologists and oncologists. Participants completed an online best-worst scaling three alternative discrete choice experiment featuring a series of hypothetical treatment outcomes for DLBCL. The DCE included two patient vignettes: patient Q, a 70-year-old male with two prior treatment lines and Eastern Cooperative Oncology Group (ECOG) performance status 1; and patient R, a 58-year-old female with three prior treatment lines and ECOG performance status 2. Predefined attributes included: overall survival (OS) rates at 12 and 24 months, change in ECOG status at 6 months, CAR-T related side effects (risk of severe cytokine release syndrome (CRS), and risk of severe neurological events), duration of hospitalization, time until treatment initiation, and total treatment cost. A conditional logit model was used to calculate odds ratios (OR) for each attribute's influence on selecting the preferred treatment recommendation. ORs were translated into willingness-to-pay (WTP) estimates for the health outcomes, changes in adverse event risks and waiting time/hospitalization duration. RESULTS: A total of 231 Japanese hematologists and oncologists participated, each making six treatment decisions for two patient vignettes. Across both profiles, 12- and 24-month overall survival (OS), hospitalization duration, waiting time, and treatment cost significantly influenced preferences. For patient Q, reduced CRS risk was significant, while for patient R, ECOG improvement was influential. Willingness to pay (WTP) for a 1% increase in 24-month OS was approximately JPY 1.4 million (~USD 9520) for patient Q and JPY 1.8 million (~USD 12,240) for patient R; corresponding estimates for 12-month OS were slightly lower. A 1-week reduction in hospital stay was valued at JPY 1.1-1.9 million (USD 7480-12,920), and a 1-week reduction in waiting time at JPY 600,000-1.5 million (USD 4080-10,200). Overall, physicians prioritized long-term survival more strongly for the younger, more severe patient, whereas preferences were more balanced for the older patient. Clinical improvement (ECOG) weighed more heavily for the younger patient, while safety concerns were more salient in the older case. CONCLUSIONS: Our findings suggest that Japanese hematologists are cost-aware in their treatment decision-making. While improved OS remains the most valued outcome, reducing hospital length of stay is associated with substantial WTP, highlighting the importance of logistical and economic factors. The severity of prognosis shifted physician priorities to short-term survival and rapid access to care with more tolerance of adverse event risks. Future research should explore how these physician preferences align with patient perspectives.