Inavolisib received marketing approval in China in March 2025 and was subsequently added to the National Reimbursement Drug List in January 2026 following national insurance price negotiations that cut its price by approximately 61%. That negotiation substantially reduced the financial burden on patients and broadened clinical access, but the new study suggests that even this steep concession was not enough to bring the triple-combination regimen within the formal economic limits that guide national reimbursement decisions. The findings matter because China’s health technology assessment framework uses a willingness-to-pay threshold defined as twice the national per capita gross domestic product, a figure that anchors whether a treatment’s added cost is justified by its added health benefit.

The clinical foundation for the economic model came from the phase III INAVO120 trial, which tested inavolisib against placebo, both in combination with palbociclib, a cyclin-dependent kinase 4 and 6 inhibitor, and fulvestrant, a selective estrogen receptor degrader. The trial enrolled patients with PIK3CA-mutated, hormone receptor-positive, HER2-negative locally advanced or metastatic breast cancer whose disease had recurred or progressed during or within twelve months of adjuvant endocrine therapy. Results showed a clear survival advantage: median overall survival reached 34.0 months with inavolisib versus 27.0 months with placebo, and median progression-free survival extended to 17.2 months versus 7.3 months. The regimen also maintained a manageable safety profile, which earned it a Class I recommendation in the Chinese Society of Clinical Oncology guidelines for endocrine therapy-resistant disease.

What distinguishes this analysis from earlier pharmacoeconomic evaluations of inavolisib is its reliance on the Asian cohort of the INAVO120 trial rather than the global population. Previous Chinese economic studies had used survival data from the overall multinational cohort, which did not account for heterogeneity across ethnic populations. In the Asian subgroup, median overall survival was 32.7 months with inavolisib versus 27.0 months with placebo, and median progression-free survival was 14.8 months versus 6.8 months. The Asian cohort’s somewhat smaller survival gains compared with the global population directly affect the value calculation, because fewer incremental quality-adjusted life-years mean less room for a high drug price to remain economically defensible.

Methodologically, the researchers built a partitioned survival model with a 30-year time horizon using TreeAge Pro 2022. Patients entered the model in the progression-free survival state and could transition to progressive disease or death. Survival data points were extracted from the overall survival and progression-free survival curves of the Asian cohort using WebPlotDigitizer, then reconstructed and extrapolated in R with standard parametric distributions including exponential, Weibull, gamma, log-normal, log-logistic, Gompertz, and generalized gamma functions. The optimal distributions were selected using the Akaike Information Criterion and the Bayesian Information Criterion, supplemented by visual inspection. The model ran on 28-day cycles matching the trial’s dosing schedule, and all costs and utilities were discounted at an annual rate of 4.5% in line with the 2025 version of the China Guidelines for Pharmacoeconomic Evaluations.

The base-case results quantified the economic gap precisely. The placebo group, receiving palbociclib and fulvestrant alone, accrued total costs of $29,741.84 and generated 2.38 quality-adjusted life-years. The inavolisib group incurred $53,444.43 in total costs while achieving 3.02 quality-adjusted life-years. The difference of $23,702.59 in incremental costs against 0.64 incremental quality-adjusted life-years produced the headline ICER of $36,932.91 per quality-adjusted life-year. Only direct medical costs were counted, including drug acquisition, routine laboratory tests, tumor assessments, best supportive care, terminal care, and management of grade 3 or higher adverse events occurring in more than 5% of either group. Drug prices were sourced from China’s government procurement platform, with palbociclib, fulvestrant, capecitabine, and paclitaxel priced from the most recent centralized procurement bids.

Scenario and sensitivity analyses confirmed the robustness of the conclusion. When the model used survival data from the overall INAVO120 population instead of the Asian cohort, the ICER was $35,315.68 per quality-adjusted life-year, still above the threshold. Assuming all patients used more expensive originator drugs rather than centrally procured generics pushed the ICER to $45,629.23 per quality-adjusted life-year. Alternative utility values, exclusion of adverse-event disutilities, and suboptimal survival curve distributions all left the ICER above the willingness-to-pay threshold. In the probabilistic sensitivity analysis, which ran 1,000 Monte Carlo simulations with all variables varying simultaneously, the probability that the inavolisib triple therapy was cost-effective at the national threshold was just 11.80%. The one-way sensitivity analysis identified progression-free survival utility, progressive disease utility, and the cost of inavolisib as the three most influential parameters.

China’s provincial economic disparities mean that local willingness-to-pay thresholds, calculated from provincial per capita gross domestic product, vary enormously. In wealthy regions such as Beijing, Shanghai, Jiangsu, Fujian, and Zhejiang, the local thresholds of $63,886.38, $60,798.84, $44,994.05, $38,617.37, and $37,957.97 per quality-adjusted life-year respectively all sit above the base-case ICER, meaning the triple therapy would be considered cost-effective in those jurisdictions. By contrast, provinces such as Gansu, Heilongjiang, and Guangxi have thresholds below $16,000 per quality-adjusted life-year, far beneath what the regimen could achieve. The authors argued that future reimbursement policy should incorporate these regional affordability differences rather than relying exclusively on national average estimates, though they cautioned that precise regional assessments would need to reflect actual provincial procurement prices and reimbursement rules. At the national threshold of twice per capita gross domestic product, a 27.02% price reduction beyond the negotiated level would bring the ICER down to the threshold. If the threshold were set at the stricter level of one times per capita gross domestic product, $13,953 per quality-adjusted life-year, the required price reduction would exceed 68.24%. Notably, the model showed that raising the threshold to three times per capita gross domestic product, $41,859 per quality-adjusted life-year, would render the therapy cost-effective, with a 69.7% probability of cost-effectiveness in the probabilistic analysis. The 2025 guidelines do permit higher thresholds for pharmaceuticals demonstrating exceptional value or substantial advances in equity, which leaves a formal pathway for such an interpretation.

The authors also acknowledged important limitations. Survival curve fitting and extrapolation introduce uncertainty, and standard parametric models may struggle to capture inflection points in curves with variable shapes and plateaued tails, though scenario testing with suboptimal distributions did not change the conclusions. Because Asian subgroup-specific data on adverse event incidence and dose intensity were unavailable, these parameters were drawn from the overall trial population, which may introduce bias. The Asian cohort itself has a limited sample size, and utility values for progression-free survival and progressive disease were derived from the UK general population using the standard gamble method, since reliable Chinese-specific utility data were lacking; cultural and demographic differences could bias these estimates. Grade 1 to 2 adverse events were excluded on the grounds that they are easily managed, but sensitivity analyses showed disutility variables had minimal impact on the results.

The study also flagged the safety signals that matter most for patients’ quality of life. Hyperglycemia and stomatitis occurred significantly more often with inavolisib than with placebo, at overall rates of 63.4% versus 13.5% and 55.3% versus 28.8% respectively, with grade 3 or higher occurrence rates of 6.8% versus 0% and 5.6% versus 0%. Clinical experts cited in the paper suggested that for obese patients or those with prediabetes, preventive metformin intervention might be considered one to two weeks before treatment, and that a dexamethasone-based, alcohol-free mouthwash could be used for the prevention or treatment of stomatitis; standardized management of these events is expected to improve quality of life for patients on the inavolisib regimen. Ultimately, the study concludes that while reimbursement has significantly eased the treatment burden for patients, inavolisib plus palbociclib/fulvestrant remains not cost-effective at the national level in China, and regionally differentiated insurance policies and further targeted price optimization may be needed to make the triple therapy economically sustainable and affordable across the country.

Subject of Research: Cost-effectiveness of inavolisib triple therapy for PIK3CA-mutated advanced breast cancer in China

Article Title: Cost-effectiveness of inavolisib plus palbociclib/fulvestrant versus palbociclib/fulvestrant for PIK3CA-mutated advanced breast cancer in China

Article References: Xiang, Z., Chen, X., & Ma, L. (2026). Cost-effectiveness of inavolisib plus palbociclib/fulvestrant versus palbociclib/fulvestrant for PIK3CA-mutated advanced breast cancer in China. Frontiers in Pharmacology, 17, 1992102. https://doi.org/10.3389/fphar.2026.1992102

Image Credits: AI Generated

DOI: 10.3389/fphar.2026.1992102

Keywords: inavolisib, PIK3CA, breast cancer, cost-effectiveness, palbociclib, fulvestrant, INAVO120, China, pharmacoeconomics, ICER, National Reimbursement Drug List, quality-adjusted life-year