Cost-effectiveness of Programmed Cell Death Ligand 1 Testing and Tumor Mutational Burden Testing of Immune Checkpoint Inhibitors for Advanced Non-Small Cell Lung Cancer
Lung cancer, particularly non-small cell lung cancer (NSCLC), remains the leading cause of tumor-related death worldwide. Despite advancements in treatment, the economic benefits of therapies for advanced NSCLC, especially immune checkpoint inhibitors (ICIs), remain a subject of debate. This study focuses on the cost-effectiveness of two widely used biomarkers for predicting ICI efficacy: programmed cell death ligand 1 (PD-L1) and tumor mutational burden (TMB). These biomarkers are evaluated through immunohistochemical methods and next-generation sequencing, respectively. However, their predictive values differ, leading to controversy over their comprehensive utility. This analysis aims to provide insights into the economic implications of these tests in both Chinese and US healthcare systems, using clinical data from the OAK trial (NCT02008227).
Background and Rationale
The OAK trial provided a unique opportunity to compare the predictive power of PD-L1 and TMB tests in the same patient population. Patients in this trial underwent both PD-L1 and blood TMB (bTMB) testing simultaneously, allowing for a direct comparison of their efficacy. The study aims to evaluate the cost-effectiveness of these biomarkers in guiding ICI treatment for advanced NSCLC, with the goal of improving precision medicine, providing multi-dimensional benefits to patients, and optimizing the allocation of medical resources.
Methodology
A Markov model was employed to assess the economic benefits of immunotherapy and its associated detection methods. The study population consisted of previously treated advanced NSCLC patients from the OAK trial, who were randomized into either the atezolizumab group (intravenous, 1200 mg every cycle) or the docetaxel group (intravenous, 75 mg/m² every cycle). Three subgroups were established: no-test, PD-L1, and TMB groups. The cut-off values for PD-L1 and TMB tests were set at 1% and 16 mutations/megabase (mut/Mb), respectively, based on available data and optimal results.
In the Markov model, patients were categorized into three mutually exclusive health states: progression-free survival, disease progression (DP), and death. The cycle length was set at 3 weeks. Data extraction from Kaplan-Meier survival curves, statistical calculations, and simulation of the decision tree model were performed using GetData Graph Digitizer, R software, and TreeAge, respectively. The Weibull survival model was used to fit the survival curves, and survival probabilities and transition probabilities were calculated accordingly.
Cost and Utility Parameters
The cost parameters included drug costs, testing costs, supportive care costs, adverse event (AE) costs, and DP state costs. For example, the cost of 1200 mg atezolizumab was $11,470.01 in China and $11,374.09 in the US, while the cost of 20 mg docetaxel was $97.00 in China and $113.27 in the US. PD-L1 tests cost $237.70 in China and $244.50 in the US, whereas TMB tests were significantly more expensive at $5675.68 in China and $5800.00 in the US. Supportive care costs were $337.50 in China and $146.32 in the US, and AE costs were $507.40 in China and $304.04 in the US. The DP state cost was $2500.00 in China and $5814.00 in the US.
Utilities were measured in quality-adjusted life years (QALYs). Health state utility values were obtained from published literature, including utility values for atezolizumab (0.7560), docetaxel (0.6520), and the DP state (0.4700). Disutility values for various AEs averaged -0.0609.
Outcomes and Analysis
The primary outcome was the incremental cost-effectiveness ratio (ICER), defined as the incremental cost per incremental QALY. Secondary outcomes included the average cost-effectiveness ratio (average CE) and net benefit, calculated as the willingness-to-pay (WTP) threshold multiplied by QALYs minus costs. The WTP threshold was set at three times the per capita GDP in China ($29,307/QALY) and $100,000/QALY in the US.
Model stability was assessed through one-way sensitivity analysis, varying cost parameters by ±30% and utility parameters by ±20%. Probabilistic sensitivity analysis was conducted using Monte Carlo simulation with 1000 iterations, fitting gamma and beta distributions to costs and utilities, respectively.
Results
The cost-effectiveness analysis was conducted in three steps. First, atezolizumab was compared with docetaxel in the no-test, PD-L1, and TMB groups. In China, the ICERs were $1,554,153/QALY, $1,495,485/QALY, and $1,340,718/QALY for the no-test, PD-L1, and TMB groups, respectively. Similar results were observed in the US, with ICERs of $1,560,996/QALY, $1,488,871/QALY, and $1,334,338/QALY. These ICERs exceeded the WTP thresholds, indicating that atezolizumab was not cost-effective compared to docetaxel in either country.
Second, the economic benefits of atezolizumab versus docetaxel were analyzed across different detection methods. In both China and the US, the PD-L1 and TMB test groups had lower ICERs than the no-test group, suggesting improved economic efficiency. The TMB test group had a lower ICER than the PD-L1 test group, indicating that the TMB test was more cost-effective. The average CE and net benefit results were consistent with these findings.
Finally, the cost-effectiveness of the biomarkers in selecting an atezolizumab-advantaged population was evaluated. Compared to the no-test group, the TMB test group achieved economic benefit with a negative incremental cost and an incremental effect of 0.0011 QALYs in both China and the US. In contrast, the PD-L1 test group was not cost-effective, with ICERs of $505,135/QALY in China and $295,962/QALY in the US. The ICER for the TMB test was $886,312/QALY in China and $371,731/QALY in the US, demonstrating that the TMB test provided greater economic benefits than the PD-L1 test.
Sensitivity Analysis
One-way sensitivity analysis revealed that chemotherapy costs had the greatest influence on model stability, followed by the DP state costs. Probabilistic sensitivity analysis indicated that if the WTP threshold exceeded $1,500,000, immunotherapy could become cost-effective. Biomarker testing could slightly reduce this threshold.
Discussion
This study is the first to balance the economic benefits of various detection methods in stratifying patients for ICIs. The use of data from the OAK trial ensured consistency across subgroups in terms of study population, treatment options, and statistical methods, enhancing comparability. TMB can be evaluated from tumor tissue (tTMB) and circulating tumor DNA (bTMB), with high consistency between the two methods. bTMB data were used in this study due to their non-invasiveness and suitability for long-term dynamic monitoring. The costs of tTMB and bTMB were considered, and no significant difference was observed.
However, the study has limitations. The synergistic effect of TMB and PD-L1 tests should be confirmed with independent predictive variables, as they may complement rather than replace each other. Given the high costs of ICIs and biomarker tests, combining TMB, PD-L1, or other tests for greater economic benefits warrants further exploration. Additionally, the study focused on atezolizumab as a single ICI to eliminate potential factor interference. Future research should verify whether these biomarker detections are cost-effective in combination therapy.
Conclusion
In conclusion, ICIs were not cost-effective compared to chemotherapy in both Chinese and US healthcare systems. However, the TMB and PD-L1 tests improved the cost-effectiveness of ICIs, with the TMB test being the most economical option. These findings provide valuable insights for improving precision medicine and optimizing the allocation of medical resources in the treatment of advanced NSCLC.
doi.org/10.1097/CM9.0000000000001120
Was this helpful?
0 / 0