早期结外 NK/T 细胞淋巴瘤的风险适应性联合模式治疗：基于马尔可夫模型的成本-效果分析

Combined-modality therapy (CMT) improves survival in patients with early-stage extranodal natural killer-/T-cell lymphoma (ENKTCL) compared with radiotherapy (RT) alone. However, the effect is inadequate for low-risk patients as defined by nomogram-revised risk index (NRI). As such, it remains unclear whether the survival benefits outweigh the additional costs.

A Markov model was constructed to compare CMT versus RT alone for patients with early-stage ENKTCL, according to five risk groups defined by NRI model. Transition probabilities, effectiveness, and cost data were derived from the China Lymphoma Collaborative Group cohort, while health utility data were estimated from adverse effects. Life-years, costs, quality-adjusted life-years (QALYs), and incremental cost-effectiveness ratios were calculated from the perspective of Chinese payers. Evaluations for customized countries or settings can be accomplished using a web-based tool.

Over the 6-year horizon, CMT increased life-years by 5.47, 5.19, 4.82, 4.62, and 4.49 years at $517,472 US dollars (USD)/QALY, $22,871 USD/QALY, $7,865 USD/QALY, $4,598 USD/QALY, and $2,278 USD/QALY for the low-risk (NRI = 0), intermediate-low-risk (NRI = 1), intermediate-high-risk (NRI = 2), high-risk (NRI = 3), and very high-risk (NRI = 4) groups, respectively. The probabilities of cost-effectiveness at a willingness-to-pay threshold of $5,208 USD/QALY were 0.00%, 0.01%, 7.40%, 72.07%, and 99.10% for each risk group. Over the lifetime horizon, all risk groups, except for low-risk group, had a probability of over 90% of being cost-effective. Estimates were varied according to country settings, integrated through a web-based customized analysis.

CMT is unlikely to be cost-effective for low-risk patients but highly likely to be cost-effective for high-risk and very high-risk patients. As for intermediate-low or intermediate-high-risk patients, the cost-effectiveness of CMT varies depending on the time horizon and willingness-to-pay threshold.