Evaluating the Impact of Subsequent Anticancer Therapy Use on Life-Year and QALY Estimates in Late-Line Metastatic Colorectal Cancer: A Health Economic Analysis of FRESCO-2
Author(s)
Victoria F. Paly, MHS1, Jesper Rimestad, MSc2, Lidziya Vanahel Ulvenes, MPhil2, Jannick Beck, MD3, Rameet Sachdev, MPH4, Pratishtha Khanduri, MSc5, Denise Zou, MS6.
1Takeda Pharmaceuticals America Inc., Cambridge, MA, USA, 2Takeda AS, Asker, Norway, 3Takeda Danmark, Vallensbæk, Denmark, 4Evidera, Bethesda, MD, USA, 5Evidera, Toronto, ON, Canada, 6Evidera, Vancouver, BC, Canada.
1Takeda Pharmaceuticals America Inc., Cambridge, MA, USA, 2Takeda AS, Asker, Norway, 3Takeda Danmark, Vallensbæk, Denmark, 4Evidera, Bethesda, MD, USA, 5Evidera, Toronto, ON, Canada, 6Evidera, Vancouver, BC, Canada.
OBJECTIVES: In FRESCO-2 (NCT04322539), fruquintinib+best supportive care (BSC) demonstrated significantly improved overall survival (OS) versus placebo+BSC in patients with refractory metastatic colorectal cancer. After discontinuation of fruquintinib versus placebo, 29% versus 34% of patients received subsequent anti-cancer therapy (ST). In real-world practice, ST options and use in late-line are limited. Post-hoc analyses demonstrated that the incremental benefit of fruquintinib was greater when adjusting OS for ST than in the intent-to-treat (ITT) analysis (hazard ratios = 0.43-0.49 versus 0.66 for ITT) (Lonardi, et al. 2025). This analysis aimed to evaluate the impact of adjusting for ST to reflect real-world practice on extrapolated survival and cost-effectiveness of fruquintinib versus BSC alone.
METHODS: Parametric survival models were fitted to OS for the ITT population and three scenarios censoring patients at time of ST initiation: naïvely censored, inverse-probability-of-censoring weighting model, and marginal structural model. Proportional hazards and accelerated failure time assumptions were assessed. A 10-year time horizon was applied to estimate life-years (LYs) and quality-adjusted life-years (QALYs) for each arm based on the best fitting distributions, using Norway (NOR) and Denmark (DEN) as reference cases. Differences in incremental cost-effectiveness ratios (ICERs) were explored.
RESULTS: In the ITT analysis, total LYs (NOR/DEN) and QALYs for fruquintinib were 0.83/0.84 and 0.56/0.64, showing an incremental benefit versus placebo+BSC of 0.17/0.17 and 0.12/0.14, respectively. In all three adjusted scenarios, incremental LYs and QALYs were consistently nearly twice as high, ranging from 0.28-0.32 LYs and 0.19-0.24 QALYs, across perspectives. ICERs fell by a proportionate amount across the three adjusted scenarios.
CONCLUSIONS: Removing the effect of ST from OS consistently showed a greater magnitude of incremental benefit for fruquintinib, indicating ST being a confounder in favour of placebo+BSC in ITT-based cost-effectiveness analyses. Adjusting trial data to better reflect real-world practice should be taken into account in Health Technology Assessment decision-making.
METHODS: Parametric survival models were fitted to OS for the ITT population and three scenarios censoring patients at time of ST initiation: naïvely censored, inverse-probability-of-censoring weighting model, and marginal structural model. Proportional hazards and accelerated failure time assumptions were assessed. A 10-year time horizon was applied to estimate life-years (LYs) and quality-adjusted life-years (QALYs) for each arm based on the best fitting distributions, using Norway (NOR) and Denmark (DEN) as reference cases. Differences in incremental cost-effectiveness ratios (ICERs) were explored.
RESULTS: In the ITT analysis, total LYs (NOR/DEN) and QALYs for fruquintinib were 0.83/0.84 and 0.56/0.64, showing an incremental benefit versus placebo+BSC of 0.17/0.17 and 0.12/0.14, respectively. In all three adjusted scenarios, incremental LYs and QALYs were consistently nearly twice as high, ranging from 0.28-0.32 LYs and 0.19-0.24 QALYs, across perspectives. ICERs fell by a proportionate amount across the three adjusted scenarios.
CONCLUSIONS: Removing the effect of ST from OS consistently showed a greater magnitude of incremental benefit for fruquintinib, indicating ST being a confounder in favour of placebo+BSC in ITT-based cost-effectiveness analyses. Adjusting trial data to better reflect real-world practice should be taken into account in Health Technology Assessment decision-making.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
EE445
Topic
Economic Evaluation
Disease
Gastrointestinal Disorders, Oncology