Cost-Effectiveness of Culmerciclib Plus Fulvestrant for Previously Treated HR-Positive/HER2-Negative Advanced Breast Cancer in China: An Integrated GBD 2023 Analysis

Abstract

Background

Breast cancer (BC) imposes a substantial and evolving health and economic burden in China. We integrated population-level disease burden with treatment-specific economic evaluation to assess the cost-effectiveness of culmerciclib plus fulvestrant for previously treated hormone receptor-positive (HR+), human epidermal growth factor receptor 2-negative (HER2−) advanced breast cancer (ABC).

Methods

Using Global Burden of Disease Study 2023 data, we assessed BC incidence, prevalence, mortality, disability-adjusted life-years (DALYs), and temporal trends in China. GBD estimates provided population-level context and were not direct inputs into the economic model. We then conducted a lifetime cost-effectiveness analysis from the Chinese healthcare perspective, comparing culmerciclib plus fulvestrant with placebo plus fulvestrant for previously treated HR+/HER2− ABC based on randomized Phase III trial efficacy. Outcomes included costs, life-years, quality-adjusted life-years (QALYs), and incremental cost-effectiveness ratios (ICERs). The willingness-to-pay (WTP) threshold was $26,896/QALY, equivalent to twice China’s per-capita gross domestic product.

Results

In 2023, BC caused 342,725 incident cases, 3,496,338 prevalent cases, 76,589 deaths, and 2,458,481 DALYs in China. From 1990 to 2023, the age-standardized rates (ASRs) of incidence and prevalence increased, whereas the ASRs of mortality and DALY decreased. All four ASRs rose during 2020–2023. In the base-case analysis, culmerciclib plus fulvestrant provided an additional 0.90 QALYs at an incremental cost of $34,150.43 compared with placebo plus fulvestrant, resulting in an ICER of $37,886.61/QALY. At the prespecified WTP threshold, the probability of culmerciclib plus fulvestrant being cost-effective was 2.7%. Culmerciclib price was the main uncertainty driver. A 25% reduction in the price of culmerciclib lowered the ICER to $26,677.03/QALY.

Conclusions

Against the background of a substantial and evolving BC burden in China, culmerciclib plus fulvestrant provides additional health benefits but is unlikely to be cost-effective compared with placebo plus fulvestrant for previously treated HR+/HER2− ABC at its current price. Integrating population-level disease burden with treatment-specific economic evidence provides complementary information for healthcare resource allocation.

Introduction

Breast cancer (BC) remains a leading driver of cancer morbidity and mortality among women worldwide. In 2022, an estimated 2.3 million new cases and approximately 666,000 deaths from female BC were reported globally, and the burden is projected to continue increasing over the coming decades.^1,2 China has become a central contributor to the global BC burden due to its large population and rapidly evolving demographic and risk-factor landscape, with approximately 357,200 incident cases in 2022.³ Although advances in screening and treatment have improved outcomes for many patients, the increasing number of women living with BC continues to generate substantial clinical and economic pressure on the healthcare system.

Hormone receptor-positive (HR+)/human epidermal growth factor receptor 2-negative (HER2−) disease is the predominant molecular subtype of BC, accounting for approximately 60–70% of cases.^4,5 Given the large overall BC population in China, this subtype represents a clinically important component of the national disease burden. Advanced breast cancer (ABC), including inoperable locally advanced and metastatic disease, remains largely incurable, and treatment is primarily aimed at prolonging survival, delaying disease progression, and maintaining quality of life. International guidelines from the National Comprehensive Cancer Network (NCCN), American Society of Clinical Oncology (ASCO), and the European Society for Medical Oncology (ESMO) consistently recommend endocrine therapy as the therapeutic backbone for most patients without visceral crisis, with early integration of cyclin-dependent kinase 4 and 6 (CDK4/6) inhibitors to enhance disease control.^6,7 Fulvestrant is a relevant partner after progression on previous endocrine therapy.⁸ However, resistance to endocrine therapy and to CDK4/6 blockade ultimately emerges in many patients,⁹ creating a persistent need for novel agents capable of extending benefit in previously treated populations.

Several CDK4/6 inhibitor-based endocrine regimens are currently available for HR+/HER2− ABC, and their clinical benefits have changed the treatment landscape in China. Culmerciclib (TQB3616) is a novel cyclin-dependent kinase 2/4/6 (CDK2/4/6) inhibitor designed to address cell-cycle escape pathways that may limit the durability of CDK4/6 inhibition.^10–12 In a randomized, double-blind Phase III trial involving patients with previously treated HR+/HER2− ABC, culmerciclib plus fulvestrant significantly prolonged progression-free survival (PFS) compared with placebo plus fulvestrant; median investigator-assessed PFS was 16.6 vs 7.5 months (hazard ratio [HR] 0.36), respectively.¹² However, improvements in clinical outcomes do not necessarily translate into good economic value, particularly when novel targeted therapies substantially increase treatment costs. Within the Chinese treatment landscape, palbociclib-, ribociclib- and abemaciclib-based regimens provide relevant CDK4/6 comparators, and Chinese economic evaluations have shown that their cost-effectiveness is sensitive to acquisition costs and reimbursement conditions.^13–17 These findings highlight the importance of evaluating newly introduced agents within the specific Chinese healthcare and pricing environment. Nevertheless, treatment-specific economic evaluations alone do not describe the broader population-level context in which resource-allocation decisions are made. Conversely, assessments of national disease burden quantify the magnitude and temporal evolution of health loss but do not determine whether a particular therapy represents good value for money. A framework incorporating both perspectives may provide more comprehensive evidence for healthcare decision-making.

The Global Burden of Disease Study 2023 (GBD 2023) provides updated estimates of incidence, prevalence, mortality, and disability-adjusted life-years (DALYs) for 375 diseases and injuries across 204 countries and territories from 1990 to 2023, alongside comparative risk assessment for modifiable risk factors.^18–20 In the present study, the GBD analysis and cost-effectiveness analysis address distinct but complementary questions. The GBD component characterizes the magnitude, temporal trends, demographic drivers, and modifiable risk factors of the overall BC burden in China, whereas the economic model evaluates the value of culmerciclib plus fulvestrant in the specific population of previously treated HR+/HER2− ABC. Accordingly, we integrated GBD 2023 evidence with a model-based cost-effectiveness analysis of culmerciclib plus fulvestrant vs placebo plus fulvestrant from the Chinese healthcare perspective to provide complementary evidence for the adoption and prioritization of emerging BC therapies in China.

Materials and Methods

Overview

This study comprised two complementary components (Figure 1): (i) a descriptive and analytical assessment of BC burden in China using the GBD 2023,^18–20 and (ii) a model-based cost-effectiveness analysis comparing culmerciclib plus fulvestrant vs placebo plus fulvestrant for previously treated HR-positive, HER2-negative ABC in China.¹² The population-level GBD estimates are not used to estimate the number of patients eligible for culmerciclib or as direct inputs into the economic model; rather, they provide the broader health-system context in which treatment value, affordability, and resource allocation should be interpreted. The economic evaluation followed the Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) guideline (Table S1).²¹

Figure 1.

Design framework of the study. (A) Flowchart of the Global burden of disease analysis for breast cancer in China. (B) Flowchart of three-state partitioned survival model.

Abbreviations: PSM, partitioned survival model.

Global Burden of Disease Study Analysis

Data Source and Measures

We extracted BC estimates for China from the GBD results database (https://vizhub.healthdata.org/gbd-results/) (Figure 1A). Outcomes extracted included incidence, prevalence, mortality, and DALYs, together with corresponding age-standardized rates (ASRs) using the GBD standard population: age-standardized incidence rate (ASIR), age-standardized prevalence rate (ASPR), age-standardized mortality rate (ASMR), and age-standardized DALY rate (ASDR). Temporal analyses were conducted for 1990–2023 overall and by age group (15–19, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–79, 80–84, 85–89, 90–94, and 95+ years). BC was identified using International Classification of Diseases (ICD) codes, including ICD-10: C50–C50.9, D05–D05.9, D24–D24.9, D48.6, and D49.3; and ICD-9: 174–175.9, 217–217.8, 233.0, 238.3, 239.3, and 610–610.9.¹⁹

Joinpoint Analysis

To characterize temporal trends and identify potential changes in trend slopes, we conducted joinpoint regression analysis for ASRs from 1990 to 2023.²² The maximum number of joinpoints was set to 6,²³ and model selection was evaluated using Monte Carlo permutation tests. For each final model, we estimated the annual percent change (APC) for each segment and the average annual percent change (AAPC) across specified intervals. To describe trends before and during the coronavirus disease 2019 (COVID-19) pandemic period, we calculated AAPC for three intervals: 1990–2019, 2020–2023, and 1990–2023.

Decomposition Analysis

To quantify drivers of change in BC burden, we performed decomposition analysis of changes in absolute numbers of incidence, prevalence, mortality, and DALYs between 1990 and 2023 based on the method proposed by Das Gupta.²⁴ Following established decomposition frameworks, the total change was partitioned into three components: (i) population growth, (ii) population aging (changes in age structure), and (iii) changes in age-specific rates (epidemiologic change). The decomposition was implemented by holding two components constant while varying the third and averaging across permutations to reduce path-dependence. Component contributions were expressed as percentages of the corresponding 1990 burdens.

Risk Factor Attribution Analysis

Following decomposition, we extracted and assessed risk factors associated with BC mortality and DALYs within the GBD 2023 Comparative Risk Assessment framework,¹⁸ including (i) all risk factors; (ii) behavioral risks at level 1, and tobacco, alcohol use, dietary risks, and low physical activity at level 2; (iii) metabolic risks at level 1, and high fasting plasma glucose (FPG) and high body-mass index (BMI) at level 2.

Cost-Effectiveness Analysis

The cost-effectiveness analysis evaluated a cohort of patients with previously treated HR+/HER2− ABC, reflecting the population enrolled in a randomized, double-blind, Phase 3 clinical trial.¹² Two strategies were compared: (i) culmerciclib 180 mg orally once daily in a 4-week cycle, combined with fulvestrant 500 mg intramuscular injection on day 1 and day 15 of cycle 1 and day 1 of subsequent cycles; and (ii) placebo administered according to the trial schedule, combined with fulvestrant dosed identically as above.

Model Construction

We developed a partitioned survival model (PSM) with three mutually exclusive health states: (i) PFS, (ii) progressed disease (PD), and (iii) death. All patients entered the model in the PFS health state (Figure 1B). The proportion of the cohort in PFS was obtained directly from the PFS curve, the proportion dead was derived as 1 − overall survival (OS), and the proportion in PD was calculated as OS − PFS. A lifetime time horizon was adopted to capture all relevant long-term costs and health outcomes, with the model continuing until more than 99% of the simulated cohort had died. The cycle length was 28 days, consistent with the treatment cycle in the clinical trial.¹² Both costs and health outcomes were discounted at an annual rate of 4.5%, in accordance with the “China Guidelines for Pharmacoeconomic Evaluations (2025)”.²⁵

Clinical Effectiveness Inputs and Survival Extrapolation

The survival data were derived from the published Kaplan–Meier curves for PFS and OS from the Phase 3 trial.¹² We digitized the PFS and OS curves using WebPlotDigitizer (version 4.8; available at: https://apps.automeris.io/wpd4/index.zh_CN.html) and extracted coordinates for each treatment group.²⁶ The digitized data were then processed and reconstructed in R statistical software to obtain reconstructed individual patient data suitable for parametric survival modeling.²⁷ Based on the reconstructed individual patient data, PFS and OS curves were fitted independently for each treatment arm using a range of candidate parametric distributions, including the generalized gamma, Weibull, Gompertz, exponential, log-logistic, gamma, and log-normal distributions. Model selection was guided by statistical goodness-of-fit criteria, namely the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), together with visual inspection to ensure clinically plausible extrapolation beyond the observed trial period.²⁸ Based on these criteria, the log-normal distribution was selected for both OS and PFS in the culmerciclib plus fulvestrant arm. In the placebo plus fulvestrant arm, the log-logistic distribution was selected for OS and the generalized gamma distribution for PFS (Table S2). The fitted curves are presented in Figure S1.

Costs and Utilities

All key model inputs, including clinical parameters, health parameters (utilities and disutilities), costs, and adverse events (AEs), are summarized in Table 1. Drug acquisition costs were derived from average prices in the YAOZHI database.³⁵ Other direct medical costs, such as administration, routine management, best supportive care, and end-of-life care per patient, were incorporated according to the sources reported in Table 1. All costs were standardized to 2024 price levels using the consumer price index and were expressed in US dollars (USD) using the 2024 average exchange rate of USD 1 = CNY 7.12. Only severe AEs (grade ≥ 3) with an incidence greater than 5% were included, consistent with the objective of capturing clinically meaningful toxicity with potential economic and quality-of-life impacts (Table 1). AE-related costs were applied in the first cycle only, reflecting acute management at treatment initiation.

Table 1.

Key Model Inputs in the Partitioned Survival Model

Variables	Baseline Value	Range		Distribution	References
		Minimum	Maximum
Clinical parameters
Risk of main AEs in culmerciclib plus fulvestrant
Diarrhea	7.22%	5.77%	8.66%	Beta	[12]
Neutropenia	24.74%	19.79%	29.69%	Beta	[12]
Leukopenia	14.95%	11.96%	17.94%	Beta	[12]
Anemia	10.82%	8.66%	12.99%	Beta	[12]
Hypertriglyceridemia	7.22%	5.77%	8.66%	Beta	[12]
Hypokalemia	14.43%	11.55%	17.32%	Beta	[12]
Lymphocyte count decrease	7.22%	5.77%	8.66%	Beta	[12]
γ-glutamyl transferase increase	5.67%	4.54%	6.80%	Beta	[12]
Health parameters
Health utilities and disutilities
Utility of PFS	0.837	0.753	0.921	Beta	[17,29]
Utility of PD	0.443	0.399	0.487	Beta	[17,29]
Diarrhea	0.05	0.04	0.06	Beta	[17,30]
Neutropenia	0.09	0.072	0.108	Beta	[13]
Leukopenia	0.09	0.072	0.108	Beta	[13]
Lymphocyte count decrease	0.09	0.072	0.108	Beta	[13]
Anemia	0.07	0.07	0.08	Beta	[31]
γ-glutamyl transferase increase	0.16	0.11	0.20	Beta	[31]
Hypertriglyceridemia	0.0294	0.0203	0.0338	Beta	[32]
Hypokalemia	0.08	0.07	0.09	Beta	[33,34]
Costs, $
Culmerciclib per 180 mg	45.30	36.24	54.35	Gamma	[35]
Fulvestrant per 250mg	46.63	37.30	55.96	Gamma	[35]
Routine follow-up per cycle	1248.61	998.89	1498.33	Gamma	[16,36–38]
Administration per cycle	227.04	181.63	272.45	Gamma	[17,31]
Best supportive care per cycle	2579.22	2063.38	3095.06	Gamma	[17,39]
End-of-life care per patient	2616.44	2093.15	3139.73	Gamma	[17,40]
Expenditures on main AEs, $
Diarrhea	41.18	32.944	49.416	Gamma	[16]
Neutropenia	388.43	310.74	466.12	Gamma	[31,41]
Leukopenia	545.44	436.35	654.53	Gamma	[31,41]
Lymphocyte count decrease	0	0	0	NA
Anemia	1204.08	963.26	1444.90	Gamma	[31]
γ-glutamyl transferase increase	306.35	245.08	367.62	Gamma	[31,42]
Hypertriglyceridemia	122.29	97.83	146.75	Gamma	[32]
Hypokalemia	13.80	11.04	16.56	Gamma	[33]
Discount rate (%)	4.5	0	5	Beta	[25]

Note: All costs are expressed in US dollars.

Abbreviations: OS, overall survival; PFS, progression-free survival; AEs, adverse events; PD, progressed disease; NA, not available.

In the base case, the model estimated discounted outcomes and costs for each strategy over the lifetime horizon. Results were reported as total cost, life-years (LYs), and quality-adjusted life-years (QALYs). Incremental outcomes included incremental cost, incremental LYs, incremental QALYs, and the incremental cost-effectiveness ratio (ICER; USD per QALY gained). A willingness-to-pay (WTP) threshold of $26,896 per QALY gained, equivalent to twice China’s per-capita gross domestic product, was applied in accordance with the China Guidelines for Pharmacoeconomic Evaluations (2025).²⁵

Sensitivity and Scenario Analysis

To evaluate parameter influence and robustness, we conducted one-way sensitivity analysis using the parameter ranges reported in Table 1. For parameters without a specified range, we applied a ±20% variation around the base-case value, reflecting plausible uncertainty in real-world settings. Results were visualized with tornado diagrams to identify the most influential parameters on the ICER. In addition, we performed probabilistic sensitivity analysis (PSA) to propagate parameter uncertainty through the model. Based on parameter types and uncertainty structures, probability distributions were assigned (beta distributions for utilities and AE probabilities; gamma distributions for cost parameters), and 1000 Monte Carlo simulations were conducted. PSA outputs were summarized as scatterplots on the incremental cost-effectiveness plane and as cost-effectiveness acceptability curves (CEACs), reporting the probability of cost-effectiveness across a range of WTP thresholds. Furthermore, we conducted scenario analyses to assess the potential impact of pharmaceutical pricing policies on the economic value of culmerciclib in China. Hypothetical reductions of 25%, 50%, and 75% in the unit price of culmerciclib were evaluated while holding all other model inputs constant. These scenarios were intended as policy-relevant sensitivity analyses rather than predictions of the future negotiated price of culmerciclib.

Statistical Software

Data processing, analysis, and visualization were executed using R (version 4.4.1; The R Foundation for Statistical Computing, Vienna, Austria), with the R packages “ggplot2” (version 3.5.1), “tidyverse” (version 2.0.0), “flexsurv” (version 2.3.2), and “survHE” (version 2.0.5). Joinpoint regression analysis was conducted using the Joinpoint Regression Program (version 5.1.0.0; the Surveillance Research Program at the National Cancer Institute, USA), while PSM was performed using Microsoft Excel.

Results

Global Burden of Disease Study Analysis

Descriptive and Joinpoint Analyses

In 2023, BC in China was estimated to cause 342,725 incident cases (95% uncertainty interval [UI]: 270,872–413,171) and 3,496,338 prevalent cases (95% UI: 2,970,629–4,035,256), with 76,589 deaths (95% UI: 64,031–90,580) and 2,458,481 DALYs (95% UI: 2,078,431–2,918,695) (Table 2). From 1990 to 2023, ASIR increased from 24.54 to 31.62 per 100,000 (AAPC: 0.83%, 95% confidence interval [CI]: 0.39 to 1.27) (Table 2, Figure 2A), and ASPR increased from 239.17 to 316.89 per 100,000 (AAPC: 0.94%, 95% CI: 0.65 to 1.22) (Table 2, Figure 2B). By contrast, ASMR decreased from 10.65 to 6.74 per 100,000 (AAPC: −1.33%, 95% CI: −1.66 to −0.99) (Table 2, Figure 2C) and the ASDR decreased from 372.92 to 228.43 per 100,000 (AAPC: −1.42%, 95% CI: −1.78 to −1.06) (Table 2, Figure 2D). All four ASRs increased during 2020–2023. During 1990–2019, ASIR and ASPR increased modestly (AAPC 0.43% and 0.66%, respectively), whereas ASMR and ASDR declined (AAPC −1.77% and −1.88%, respectively) (Table 2). However, during 2020–2023, all ASRs increased, including those previously declining: ASIR increased with an AAPC of 5.28% (95% CI: 3.62 to 6.96), ASPR with an AAPC of 3.68% (95% CI: 2.30 to 5.08), ASMR with an AAPC of 3.75% (95% CI: 2.40 to 5.11), and ASDR with an AAPC of 3.16% (95% CI: 1.42 to 4.94) (Table 2).

Table 2.

Breast Cancer Burden in 1990 and 2023 and Average Annual Percent Changes in Age-Standardized Rates

Measure	Number, 95% UI		Age-Standardized Rates, per 100,000, 95% UI		AAPC of Age-Standardized Rate, 95% CI
	In 1990	In 2023	In 1990	In 2023	1990–2019	2020–2023	1990–2023
Incidence	119,317 (95,337 to 149,295)	342,725 (270,872 to 413,171)	24.54 (19.70 to 30.79)	31.62 (25.40 to 38.22)	0.43 (−0.01 to 0.86)	5.28 (3.62 to 6.96)	0.83 (0.39 to 1.27)
Prevalence	1,138,680 (950,572 to 1,360,921)	3,496,338 (2,970,629 to 4,035,256)	239.17 (200.44 to 283.91)	316.89 (269.13 to 366.06)	0.66 (0.38 to 0.95)	3.68 (2.30 to 5.08)	0.94 (0.65 to 1.22)
Mortality	49,616 (40,599 to 62,432)	76,589 (64,031 to 90,580)	10.65 (8.76 to 13.51)	6.74 (5.71 to 8.03)	−1.77 (−2.10 to −1.45)	3.75 (2.40 to 5.11)	−1.33 (−1.66 to −0.99)
DALYs	1,844,963 (1,493,993 to 2,303,636)	2,458,481 (2,078,431 to 2,918,695)	372.92 (301.56 to 465.93)	228.43 (195.32 to 269.22)	−1.88 (−2.25 to −1.51)	3.16 (1.42 to 4.94)	−1.42 (−1.78 to −1.06)

Abbreviations: UI, uncertainty interval; AAPC, average annual percent change; CI, confidence interval; DALYs, disability-adjusted life-years.

Figure 2.

Temporal trends and decomposition of the breast cancer burden in China. (A) Joinpoint regression analysis of age-standardized incidence rates from 1990 to 2023. (B) Joinpoint regression analysis of age-standardized prevalence rates from 1990 to 2023. (C) Joinpoint regression analysis of age-standardized mortality rates from 1990 to 2023. (D) Joinpoint regression analysis of age-standardized DALYs rates from 1990 to 2023. Black squares represent annual GBD estimates. Colored line segments represent trends identified by joinpoint regression, and annotations indicate the corresponding APCs and joinpoint years. (E) Decomposition of changes in incidence, prevalence, mortality, and DALYs between 1990 and 2023 into contributions from population aging, population growth, and epidemiological change; black dots indicate the total percentage change. * indicates that the annual percent change is significantly different from zero at the alpha = 0.05 level.

Abbreviations: APC, annual percent change; DALYs, disability-adjusted life-years.

Joinpoint regression further characterized multiple turning points (Figure 2, Table S3). For incidence, ASIR increased during 1990–2000 (APC 1.32%), showed alternating decreases and increases through 2020, and then rose sharply during 2020–2023 (APC 5.28%) (Figure 2A, Table S3). Prevalence followed a similar pattern, with an early increase in 1990–2000 (APC 1.12%), a significant rise in 2013–2020 (APC 0.68%), and an accelerated increase in 2020–2023 (APC 3.68%) (Figure 2B, Table S3). In contrast, ASMR exhibited sustained reductions across most of 2000–2020, particularly 2000–2005 (APC −3.32%), 2010–2013 (APC −4.76%), and 2017–2020 (APC −3.11%), followed by a significant increase in 2020–2023 (APC 3.75%) (Figure 2C, Table S3). ASDR showed a comparable shift: after significant declines during 1999–2010 (APC −2.76%) and 2013–2020 (APC −1.53%), the ASDR increased in 2020–2023 (APC 3.16%) (Figure 2D, Table S3).

Decomposition Analysis Between 1990 and 2023

Between 1990 and 2023, the absolute BC burden increased substantially, with the largest relative rise observed for prevalence (207.05%), followed by incidence (187.24%), whereas increases were more modest for mortality (54.36%) and DALYs (33.25%) (Figure 2E, Table S4). For incidence, the overall increase was mainly attributable to population aging (83.97%), with additional contributions from population growth (57.56%) and epidemiological change (45.71%) (Figure 2E, Table S4). For prevalence, the overall increase was again primarily driven by population aging (93.92%), while population growth (60.18%) and epidemiological change (52.95%) contributed comparably (Figure 2E, Table S4). In contrast, mortality displayed a distinct pattern. Decomposition indicated that population aging (72.69%) and population growth (42.00%) exerted upward pressure, whereas epidemiological change contributed negatively (−60.33%) (Figure 2E, Table S4). A similar pattern was observed for DALYs, with population aging (52.59%) and population growth (38.68%) increased DALYs, but epidemiological change reduced DALYs by 58.02% (Figure 2E, Table S4).

Disease Burden by Age

Age-stratified trends in BC burden differed by outcome between 1990 and 2023. Incidence and prevalence rates generally increased among middle-aged and older populations (≥35 years), whereas mortality and DALY rates declined across many of these age groups. Incidence and prevalence rates declined among those aged 15–29 years and increases in mortality and DALY rates were concentrated in the oldest age groups (≥85 years) (Figure 3, Table S5). For incidence in 2023, the highest number of incident cases occurred in individuals aged 50–54 years, whereas the highest age-specific incidence rate was observed in the 60–64 years group (Figure 3A, Table S5). From 1990 to 2023, the largest AAPC in the age-specific incidence rate was in 85–89 years (2.14%; 95% CI: 1.38–2.92), while the smallest AAPC occurred in 20–24 years (−1.29%; 95% CI: −1.79 to −0.78) (Figure 3B, Table S5). For prevalence in 2023, the highest number of prevalent cases was in the 55–59 years group, whereas the highest age-specific prevalence rate was in 65–69 years (Figure 3A, Table S5). The largest AAPC in the age-specific prevalence rate was also observed in 65–69 years (1.18%; 95% CI: 0.96–1.39), and the smallest AAPC occurred in 20–24 years (−0.65%; 95% CI: −1.10 to −0.20) (Figure 3B, Table S5). For mortality in 2023, the highest number of deaths occurred in 55–59 years, while the highest age-specific mortality rate was in the ≥95 years group (Figure 3A, Table S5). The largest AAPC in the age-specific mortality rate was in ≥95 years (0.87%; 95% CI: 0.57–1.17), whereas the smallest AAPC was in 20–24 years (−3.90%; 95% CI: −4.36 to −3.45) (Figure 3B, Table S5). For DALYs in 2023, the highest DALY count occurred in 55–59 years, whereas the highest age-specific DALY rate was in 60–64 years (Figure 3A, Table S5). The largest AAPC in the age-specific DALY rate was observed in ≥95 years (0.80%; 95% CI: 0.49–1.10), and the smallest AAPC occurred in 20–24 years (−3.86%; 95% CI: −4.21 to −3.52) (Figure 3B, Table S5).

Figure 3.

Age-specific breast cancer burden and temporal trends by age group in China. (A) Age-specific rates of breast cancer incidence, prevalence, mortality, and DALYs in 2023. (B) AAPCs in age-specific rates of incidence, prevalence, mortality, and DALYs from 1990 to 2023. In panel B, shaded areas represent 95% confidence intervals, and the horizontal dashed line indicates an AAPC of 0%.

Abbreviation: AAPC, average annual percent change.

Risk Factor Attribution Analysis

From 1990 to 2023, the fraction of BC burden attributable to all the included risks remained close to one third for both mortality and DALYs, with only a modest increase over time (mortality: from 33.57% to 34.26%; DALYs: from 33.02% to 34.11%) (Figure 4A, Table S6). Notably, this apparent stability in total population-attributable fractions (PAFs) masked a clear shift in the underlying risk composition. Across both outcomes, attributable fractions for behavioral risks declined slightly (mortality: from 26.73% to 25.36%; DALYs: from 26.63% to 25.73%), whereas metabolic risks increased steadily (mortality: from 9.38% to 11.92%; DALYs: from 8.79% to 11.28%) (Figure 4A, Table S6). At level 2, behavioral and metabolic risks showed a consistent ordering across mortality and DALYs, while their trajectories diverged (Figure 4B, Table S6). In 1990, the leading contributors were tobacco and dietary risks (each around 13.00% for both outcomes), followed by high FPG (6.00–7.00%), with high BMI (2.00–3.00%), low physical activity (1.60–1.80%), and alcohol use (1.00%) ranking lower (Figure 4B, Table S6). By 2023, dietary risks remained persistently high and ranked first (13.50% for both mortality and DALYs), tobacco declined yet remained second (10.90–11.50%), high FPG was largely stable and remained third (6.50–6.80%), whereas high BMI, although still fourth, approximately doubled (5.20–5.60%) (Figure 4B, Table S6). Meanwhile, low physical activity increased modestly but stayed fifth (2.00–2.30%), and alcohol use declined further and remained the lowest-ranked contributor (0.80–1.00%) (Figure 4B, Table S6).

Figure 4.

Temporal trends of population-attributable fractions for breast cancer-related mortality and DALYs from 1990 to 2023. (A) Population-attributable fractions for all risk factors combined and for the level 1 categories of behavioral and metabolic risks. (B) Population-attributable fractions for selected level 2 risk factors, including alcohol use, dietary risks, high body-mass index, high fasting plasma glucose, low physical activity, and tobacco.

Abbreviation: DALYs, disability-adjusted life-years.

Cost-Effectiveness Analysis

Base-Case Analysis Results

In the base-case analysis, culmerciclib plus fulvestrant yielded 3.68 LYs and 2.59 QALYs at a total cost of $117,996.33, whereas placebo plus fulvestrant yielded 2.84 LYs and 1.69 QALYs at a total cost of $83,845.90 (Table 3). Compared with placebo plus fulvestrant, culmerciclib plus fulvestrant increased health outcomes by 0.84 LYs and 0.90 QALYs, with an incremental cost of $34,150.43. The ICER for culmerciclib plus fulvestrant vs placebo plus fulvestrant was $37,886.61/QALY, exceeding the WTP threshold by $10,990.61 (40.86% above the threshold) (Table 3).

Table 3.

Results of the Base-Case and Scenario Analyses

Treatment	Total Cost ($)	LYs	QALYs	Incremental Cost ($)	Incremental LYs	Incremental QALYs	ICER ($/QALY)
Base-case analysis
Culmerciclib plus fulvestrant	117,996.33	3.68	2.59	34,150.43	0.84	0.90	37,886.61
Placebo plus fulvestrant	83,845.90	2.84	1.69
Scenario analysis 1: 25% reduction in the price of culmerciclib
Culmerciclib plus fulvestrant	107,892.18	3.68	2.59	24,046.28	0.84	0.90	26,677.03
Placebo plus fulvestrant	83,845.90	2.84	1.69
Scenario analysis 2: 50% reduction in the price of culmerciclib
Culmerciclib plus fulvestrant	97,788.03	3.68	2.59	13,942.13	0.84	0.90	15,467.45
Placebo plus fulvestrant	83,845.90	2.84	1.69
Scenario analysis 3: 75% reduction in the price of culmerciclib
Culmerciclib plus fulvestrant	87,683.88	3.68	2.59	3,837.98	0.84	0.90	4,257.87
Placebo plus fulvestrant	83,845.90	2.84	1.69

Abbreviations: LY, life-year; QALY, quality-adjusted life-years; ICER, incremental cost-effectiveness ratio.

Sensitivity Analysis Results

The one-way sensitivity analysis identified the cost of culmerciclib (per 180 mg) as the primary driver of model uncertainty, with the ICER ranging from $28,918.94/QALY to $46,854.27/QALY (Figure 5A). The utility of PFS and the cost of best supportive care per cycle were also influential, generating ICER ranges of $33,671.74–$43,307.65/QALY and $34,167.20–$41,606.01/QALY, respectively. Additional parameters with moderate impact included the cost of routine follow-up per cycle, utility of PD, the cost of administration per cycle, and cost of fulvestrant (per 250 mg). Across all parameters examined, the ICER remained above the WTP threshold, including under the most favorable deterministic assumption ($28,918.94/QALY). Figure 5B and 5C present the PSA results. The incremental cost-effectiveness plane and the CEAC analysis indicated that placebo plus fulvestrant had a 97.3% probability of being cost-effective at a WTP of $26,896/QALY, while culmerciclib plus fulvestrant had a 2.7% probability. At $30,000/QALY, the probability for culmerciclib plus fulvestrant increased to 9.8% (placebo: 90.2%), and at $35,000/QALY it increased to 31.4% (placebo: 68.6%) (Figure 5C). The probability exceeded 50% at $38,000/QALY (51.1% for culmerciclib vs 48.9% for placebo), reached 86.1% at $45,000/QALY, and 96.4% at $50,000/QALY. At thresholds of $70,000/QALY and above, the probability for culmerciclib plus fulvestrant was 100% (Figure 5C).

Figure 5.

Sensitivity analyses of culmerciclib plus fulvestrant vs placebo plus fulvestrant. (A) One-way sensitivity analysis presented as a tornado diagram. Horizontal bars show the range of ICERs obtained by varying each parameter over its prespecified range; the reference lines indicate the base-case ICER and the prespecified WTP threshold. (B) Cost-effectiveness plane based on 1000 probabilistic sensitivity analysis simulations. Each point represents one Monte Carlo simulation, the triangle indicates the base-case result, the ellipse represents the 95% probability region for the joint distribution of incremental costs and incremental QALYs, and the dashed line represents the WTP threshold. (C) Cost-effectiveness acceptability curves showing the probability that each treatment strategy is cost-effective across alternative WTP thresholds; the vertical dashed line indicates the base-case WTP threshold.

Abbreviations: ICER, incremental cost-effectiveness ratio; PFS, progression-free survival; PD, progressed disease; WTP, willingness-to-pay; QALY, quality-adjustedlife-years.

Scenario Analysis Results

The scenario analyses evaluating culmerciclib price reductions are summarized in Table 3. In scenario analysis 1, culmerciclib plus fulvestrant remained associated with a total cost of $107,892.18, with an incremental cost of $24,046.28 and an ICER of $26,677.03/QALY (Table 3). In scenario analysis 2, the total cost of culmerciclib plus fulvestrant decreased to $97,788.03, corresponding to an incremental cost of $13,942.13 relative to placebo plus fulvestrant. The resulting ICER was $15,467.45/QALY, indicating cost-effectiveness under this assumption (Table 3). In scenario analysis 3, the total cost of culmerciclib plus fulvestrant further declined to $87,683.88, reducing the incremental cost to $3,837.98 and yielding an ICER of $4,257.87/QALY (Table 3). The corresponding incremental cost-effectiveness planes and CEACs are presented in Figures S2–S4.

Discussion

This study combines two complementary perspectives on BC in China: the population-level burden characterized using GBD 2023 and the treatment-specific economic value of culmerciclib plus fulvestrant for previously treated HR+/HER2− ABC. Our GBD analyses indicate rising ASIR and ASPR from 1990 to 2023, alongside long-term declines in ASMR and ASDR. Importantly, we observed a reversal after 2020 with simultaneous increases across all four age-standardized metrics. In the economic evaluation, culmerciclib plus fulvestrant provided additional health benefits compared with placebo plus fulvestrant but was unlikely to be cost-effective at the current price and prespecified WTP threshold. By combining epidemiological burden analysis with a pharmacoeconomic assessment, this study offers a more comprehensive perspective for policymakers, supporting value-oriented decision-making to align new treatments with public health needs.

The observed long-term trends in BC likely reflect the combined effects of socioeconomic development, lifestyle transitions, population aging, and improved prevention and control measures over the past few decades. Accelerated economic growth and lifestyle westernization in China have plausibly induced a constellation of risk factors, including later age at first birth, reduced breastfeeding, rising obesity prevalence, and physical inactivity, ultimately heightening population-level BC risk.^43–45 At the same time, expanded screening programs and advances in diagnostic imaging, such as wider use of mammography and ultrasound, have led to more BC being detected, thus elevating incidence rates.⁴⁶ Crucially, many of these detected cases are caught earlier, and combined with improvements in treatment, including more effective surgeries, chemotherapies, endocrine and targeted therapies, and improved radiotherapy techniques, survival has improved.^47–49 These factors may help explain why more women are living with BC (higher ASPR) even as ASMR and ASDR declined before 2020. However, we identified a marked inflection after 2020, with simultaneous increases in all four age-standardized metrics. A plausible explanation is the impact of the COVID-19 pandemic. During the pandemic, many regions (including China) experienced delays or reductions in cancer screening, diagnosis, and treatment continuity.^50–56 Delayed detection might have produced a clustering of diagnoses from 2021 to 2023, while interrupted treatment could have temporarily worsened mortality and DALY trends.^54–56 Nevertheless, the available ecological data cannot distinguish a true epidemiological shift from an ascertainment or reporting artifact. The temporal coincidence with pandemic-related disruptions supports cautious interpretation but does not establish causation. Longer follow-up and linked registry data will be needed to determine whether the post-2020 increases persist.

The decomposition analysis provides further insight into what is driving the rising BC burden. Between 1990 and 2023, population aging was the dominant driver of the rising absolute number of cases (accounting for 84% of incidence increase and 94% of prevalence increase). Population growth adds additional upward pressure. World Bank data indicate China’s population increased from 1.135 billion in 1990 to a peak of 1.412 billion in 2021.⁵⁷ These findings align with previous analyses highlighting rapid population growth and aging as major reasons for the expanding cancer burden in China.^58–60 As case numbers rise, the affordability of care becomes increasingly vital, highlighting the critical importance of cost-effectiveness thresholds for new therapeutic interventions. Notably, substantial progress in oncology treatment and preventive strategies has yielded considerable decreases in mortality and DALYs (−60.33% and −58.02%, respectively), highlighting the critical need to consolidate and expand these achievements in BC control. These macro-level counts describe population need but do not define the population eligible for culmerciclib. Treatment decisions still require receptor, stage and prior-treatment information.

It is also noteworthy how the burden is shifting across age groups. We observed that age-specific incidence and prevalence rates have increased most among middle-aged and older women, whereas younger women (under 30) have seen stable or even declining rates. The largest relative increases were observed in the oldest age brackets (over 85 years). This may reflect generational differences in risk exposure and the cumulative effects of aging. It also possibly indicates successes in controlling BC among younger women, potentially attributable to heightened awareness, or reduced exposure to specific risk factors in recent birth cohorts. Conversely, the rising age-specific mortality and DALY rates in older women highlight a critical equity concern: current screening protocols and preventive interventions may insufficiently cover this expanding demographic, particularly given that national screening initiatives in China predominantly target younger cohorts.^43,61 Moreover, older patients often have comorbidities that can complicate treatment. This underscores the urgency of developing geriatric-tailored BC control strategies.

Our analysis of risk factor attribution reveals a shifting landscape of modifiable influences on BC. Over the past three decades, approximately one-third of BC-related deaths and DALYs in China were consistently attributable to the combination of known behavioral and metabolic risk factors. This enduring 34% preventable fraction underscores considerable potential for prevention, yet the relatively static total fraction masks a clear transition in risk profile. Tobacco use showed a decrease in population-attributable fraction in our study. This is likely due to successful tobacco control policies and the fact that smoking rates among women in China have remained low or declined.^62–64 High BMI roughly doubled its attributable fraction, in line with the accelerating rise of overweight and obesity in Chinese women.^65,66 However, these GBD risk estimates are not stratified jointly by receptor status, stage and treatment history. We cannot quantify what proportion of the metabolic-risk burden occurs specifically in previously treated HR+/HER2− ABC. These findings suggest that China’s BC prevention strategies should adapt to an epidemiological transition. Traditional efforts like tobacco control remain vital but should be coupled with aggressive measures to combat obesity and metabolic syndrome. Policy options could include community programs for weight management and exercise, nutritional guidelines, food industry regulation, and sustained public education on lifestyle risks.

Our pharmacoeconomic analysis indicates that despite its clinical efficacy, the culmerciclib plus fulvestrant combination is unlikely to be an economically efficient therapy in China at current prices. These results align with prior cost-effectiveness analyses of similar agents in both Western and Chinese settings, which consistently reported unfavorable ICERs for CDK4/6 inhibitors at their original prices. For example, a US analysis by Mamiya et al found that adding palbociclib to fulvestrant yielded an ICER of about $918,166/QALY, with essentially 0% chance of cost-effectiveness at $100,000/QALY.⁶⁷ Even in China, a 2019 study estimated an ICER of $182,779/QALY for palbociclib plus fulvestrant, at that time an untenable cost-effectiveness ratio.³⁶ Likewise, a recent study reported an ICER of over $1,070,526/QALY for ribociclib plus fulvestrant vs fulvestrant, indicating it was not cost-effective in China’s context.⁶⁸ Probabilistic uncertainty analysis, given current pricing, also result in a very low probability of cost-effectiveness (2.7% at the threshold). The one-way sensitivity analyses further underscore that the price of culmerciclib was by far the most influential parameter. This finding is consonant with other pharmacoeconomic evaluations of targeted cancer therapies, where the acquisition cost often determines overall cost-effectiveness.^36,68 These results echo the consensus that substantial price reductions or discounts are usually required for expensive new oncology drugs to achieve acceptable cost-effectiveness.⁶⁹ Following substantial improvements in patient access through healthcare reimbursement negotiations, Jia et al compared palbociclib, ribociclib, and abemaciclib plus fulvestrant for second-line HR+/HER2− advanced or metastatic breast cancer in China, using palbociclib plus fulvestrant as the reference regimen.⁷⁰ At a WTP threshold of $37,738/QALY, ribociclib plus fulvestrant had the highest probability of being cost-effective (99.8%). Notably, this contrasts with our approach, which employed a threshold of twice per-capita GDP. Encouragingly, our scenario analyses demonstrated that price concessions could dramatically improve the value profile. This implies that price negotiation could pivot the drug into a cost-effective range.

The epidemiological and economic components of this study should be viewed as complementary rather than directly linked analytical models. The GBD analysis identifies the scale, trajectory, demographic drivers, and modifiable risk factors of the national BC burden, thereby characterizing the broader demand placed on the Chinese healthcare system. The cost-effectiveness analysis addresses a different question: whether the additional health gains of a specific intervention in a defined clinical population justify its additional costs. A high or increasing disease burden does not itself make an intervention cost-effective; instead, it increases the importance of considering affordability, opportunity costs, and efficient allocation of limited healthcare resources. As the number of patients requiring BC care increases, the cumulative budgetary consequences of adopting high-cost targeted therapies may become increasingly important even when the eligible population represents only a subset of all patients with BC. Cost-effectiveness evidence can complement epidemiological burden estimates by informing value-based pricing and reimbursement, while GBD evidence provides the broader context for prioritizing cancer prevention, treatment, and survivorship resources. This integrated interpretation may support more balanced healthcare decision-making by considering population health needs, treatment benefits, affordability, and opportunity costs simultaneously.

We acknowledge several limitations in our study. First, the GBD 2023 estimates we used, while comprehensive and up-to-date, are derived from modeling and existing data sources that carry inherent uncertainties. There may be underestimation of BC burden due to historical under-reporting or lags in reflecting recent interventions. Second, our integrated approach inherently compares population-level burden with a subtype-specific intervention. The GBD data encompass all BC cases across all subtypes and stages, whereas our cost-effectiveness evaluation pertains specifically to HR+/HER2− ABC after prior therapy. Because GBD does not provide joint receptor, stage and treatment-history strata, the percentage of the 3,496,338 prevalent cases that falls into the modeled subgroup cannot be estimated, and the metabolic-risk estimates cannot be assigned specifically to that subgroup. The modeled subgroup is nested within the broader BC population, so caution is needed when extrapolating from overall trends to this particular subgroup. Future studies might refine this by extracting GBD data specific to advanced disease and molecular subtype if available. Third, the cost-effectiveness model was primarily informed by efficacy data from a single randomized Phase III trial. Because original individual patient-level survival data were unavailable, PFS and OS data were reconstructed from published Kaplan–Meier curves and subsequently extrapolated beyond the observed trial follow-up using parametric survival models. Long-term estimates of survival, costs, QALYs, and the resulting ICER are dependent on the selected extrapolation functions. Although model selection was informed by AIC, BIC, visual fit, and clinical plausibility, longer follow-up and mature survival data might alter the projected long-term outcomes. Furthermore, trial-derived efficacy, health-state utilities, and costs obtained largely from published literature may not fully reflect routine clinical practice in China. Future real-world studies of culmerciclib incorporating longer-term survival, quality-of-life outcomes, treatment patterns, and actual healthcare costs would help validate and refine the present economic estimates.

Conclusion

In conclusion, the population-level GBD analysis and treatment-specific economic evaluation provide distinct but complementary evidence. The former highlights the substantial and evolving healthcare demand associated with BC in China, whereas the latter indicates that culmerciclib plus fulvestrant provides additional health benefits but is unlikely to be cost-effective at its current price under the present model assumptions. These findings may inform price negotiation and resource-allocation discussions, although the economic value of culmerciclib should be reassessed as prices change and longer-term clinical and real-world evidence becomes available.

Funding Statement

This work was funded by the Science and Technology Planning Project of Medicine and Health in Zhuji City, Zhejiang Province, China (2023YW014). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Ethical Approval and Informed Consent Statements

This study was a secondary analysis of publicly available, anonymized, and aggregate-level estimates from the Global Burden of Disease Study 2023. No individual-level data, personal identifiers, or patient contact information were accessed. According to Article 32, Items 1 and 2, of the Measures for Ethical Review of Life Science and Medical Research Involving Human Subjects issued by the National Health Commission of the People’s Republic of China on February 18, 2023, research using legally obtained public data without involving identifiable personal information is exempt from institutional ethical review. Therefore, ethics committee approval was not required for this study. Written informed consent was not applicable because no identifiable individual information was included in the analyses.

Data Sharing Statement

The datasets used in this study are publicly available at: http://ghdx.healthdata.org/gbd-results-tool. Additional questions may be addressed to the corresponding author.

Author Contributions

All authors made a significant contribution to the work reported, whether in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising, or critically reviewing the article; gave final approval of the version to be published; agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no conflicts of interest.