COST-EFFECTIVENESS ANALYSIS OF DPYD-GUIDED FOLFOX IN UNRESECTABLE METASTATIC COLORECTAL CANCER IN THE US: A PATIENT-LEVEL MICROSIMULATION MODEL
Author(s)
Hyunwoo Koo, MS, PharmD1, Pianpian Cao, PhD1, Anita Turk, MD2, Todd Skaar, PhD2;
1Purdue University, West Lafayette, IN, USA, 2Indiana University School of Medicine, Indianapolis, IN, USA
1Purdue University, West Lafayette, IN, USA, 2Indiana University School of Medicine, Indianapolis, IN, USA
OBJECTIVES: To evaluate the cost-effectiveness of DPYD-guided FOLFOX versus universal dosing for unresectable metastatic colorectal cancer from the payer’s perspective.
METHODS: We developed a patient-level microsimulation model leveraging a three-state semi-Markov structure with a tunnel phase to capture time-dependent changes in NCI CTCAE grade 3+ adverse events (diarrhea/colitis, nausea/vomiting, or hematologic toxicity). Time horizon was 10-year. Costs and outcomes were discounted by 3%. All cost values were adjusted to 2025 USD, and outcomes include Quality-Adjusted Life Years (QALYs). Kaplan-Meier curves from the clinical trials were digitized using WebPlotDigitizer v5 to parametrically fit the survival outcomes: time to progression, progression-free survival, or overall survival. Life table for the total population (US, 2023) was used to evaluate background death in progression-free state. Costs were obtained from the Payment Allowance Limits for the Medicare Part B drugs, the CMS Physician Fee Schedule, published papers, and GoodRx. Costs related to adverse events from claims data were modeled with a gamma distribution to capture heterogenous resource use (e.g., hospitalization and G-CSF). Frequency of DPYD heterozygous carriers was derived from the US-based real-world study and homozygous carriers were estimated by Hardy-Weinberg. Probability distribution of mortality and incidence of adverse events corresponding to each genotype group were approximated with real-world data through Markov Chain Monte Carlo. To evaluate uncertainty, probabilistic sensitivity analysis (PSA) and deterministic one-way sensitivity analyses were conducted.
RESULTS: Compared to universal dosing, DPYD-guided therapy decreased costs but gained QALYs, resulting in an ICER of -$71,684.81 per QALY and ~85% probability of being cost-effective in the PSA. One-way sensitivity analysis suggested that the cost of bevacizumab is the most influential variable on the NMB.
CONCLUSIONS: DPYD-guided therapy is likely cost-effective compared to universal dosing for patients with unresectable metastatic colorectal cancer in the US, supporting its implementation in clinical practice.
METHODS: We developed a patient-level microsimulation model leveraging a three-state semi-Markov structure with a tunnel phase to capture time-dependent changes in NCI CTCAE grade 3+ adverse events (diarrhea/colitis, nausea/vomiting, or hematologic toxicity). Time horizon was 10-year. Costs and outcomes were discounted by 3%. All cost values were adjusted to 2025 USD, and outcomes include Quality-Adjusted Life Years (QALYs). Kaplan-Meier curves from the clinical trials were digitized using WebPlotDigitizer v5 to parametrically fit the survival outcomes: time to progression, progression-free survival, or overall survival. Life table for the total population (US, 2023) was used to evaluate background death in progression-free state. Costs were obtained from the Payment Allowance Limits for the Medicare Part B drugs, the CMS Physician Fee Schedule, published papers, and GoodRx. Costs related to adverse events from claims data were modeled with a gamma distribution to capture heterogenous resource use (e.g., hospitalization and G-CSF). Frequency of DPYD heterozygous carriers was derived from the US-based real-world study and homozygous carriers were estimated by Hardy-Weinberg. Probability distribution of mortality and incidence of adverse events corresponding to each genotype group were approximated with real-world data through Markov Chain Monte Carlo. To evaluate uncertainty, probabilistic sensitivity analysis (PSA) and deterministic one-way sensitivity analyses were conducted.
RESULTS: Compared to universal dosing, DPYD-guided therapy decreased costs but gained QALYs, resulting in an ICER of -$71,684.81 per QALY and ~85% probability of being cost-effective in the PSA. One-way sensitivity analysis suggested that the cost of bevacizumab is the most influential variable on the NMB.
CONCLUSIONS: DPYD-guided therapy is likely cost-effective compared to universal dosing for patients with unresectable metastatic colorectal cancer in the US, supporting its implementation in clinical practice.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
PT30
Topic
Economic Evaluation
Disease
SDC: Oncology, STA: Personalized & Precision Medicine