Clinical outcomes of everolimus plus exemestane versus fulvestrant in patients with metastatic breast cancer after progression on cyclin-dependent kinase 4/6 inhibitors

For tens of thousands of women living with metastatic breast cancer, the arrival of cyclin-dependent kinase 4/6 inhibitors transformed a once-grim prognosis into something that could be managed for years. Drugs such as palbociclib, ribociclib, and abemaciclib, when paired with hormone therapy, have become the backbone of first-line treatment for tumors that carry hormone receptors but lack the HER2 growth receptor. Yet these targeted pills, for all their power, do not cure the disease. Sooner or later, the cancer finds a way around them, and clinicians are left facing a question that modern oncology has never fully answered: what should come next? A new retrospective study from South Korea, published in BMC Cancer, offers one of the clearest real-world comparisons to date of the two most common second-line options, and its findings are already sparking debate among breast cancer specialists.

The research team, led by Byeong-Chan Oh, Seong-Jung Kim, and Eui-Kyung Lee of the School of Pharmacy at Sungkyunkwan University, turned to a vast national claims database maintained by the Korean Health Insurance Review and Assessment Service. Because South Korea runs a single-payer health system covering virtually the entire population, the database captures nearly every cancer patient in the country, along with every drug they receive. The investigators combed through records from 2016 to 2023, identifying patients with hormone receptor–positive, HER2-negative metastatic breast cancer whose disease had progressed after treatment with a CDK4/6 inhibitor. They then split these patients into two cohorts based on the second-line therapy their physicians chose: 314 patients received everolimus plus exemestane, while 316 received fulvestrant.

The two treatments represent fundamentally different strategies for attacking a tumor that has learned to resist hormone blockade. Everolimus is an mTOR inhibitor, a targeted drug that shuts down a key intracellular signaling pathway, the mammalian target of rapamycin, which many resistant breast cancers rely on to keep growing despite estrogen deprivation. Exemestane, meanwhile, is an aromatase inhibitor that strips the body of estrogen production. The combination, approved after the landmark BOLERO-2 trial, is designed to hit the cancer from two directions at once. Fulvestrant takes a different tack entirely: it is a selective estrogen receptor degrader, an injected antibody-like drug that binds to and destroys the estrogen receptor itself, eliminating the molecular switch the tumor depends on. Which of these approaches works better after CDK4/6 inhibitor failure has been a matter of educated guesswork rather than established evidence, since no large randomized trial has definitively settled the question.

Because this was not a randomized trial but a retrospective cohort study, the researchers had to confront a stubborn statistical problem: patients who receive everolimus plus exemestane may differ systematically from those who receive fulvestrant, and those differences, not the drugs themselves, could drive any apparent survival gap. To address this, the team used inverse probability of treatment weighting, a sophisticated statistical technique that assigns each patient a probability of receiving one treatment or the other based on measured characteristics, then reweights the cohorts so that they become statistically comparable. The variables adjusted for included age, sex, comorbidity burden measured by the Charlson comorbidity index, cancer stage, and prior treatments. After weighting, the two cohorts were balanced on these measured confounders, allowing the investigators to estimate the independent effect of each drug strategy.

The headline finding concerns time to next treatment, or TTNT, a real-world surrogate for how long a therapy keeps the disease in check before patients and physicians must move on to something else. In the unadjusted analysis, patients on everolimus plus exemestane stayed on their treatment significantly longer than those on fulvestrant, with a median TTNT of 6.7 months versus 4.6 months, a difference that reached statistical significance with a p-value of 0.02. After adjustment for measured confounders, the gap persisted: the hazard ratio for moving to a next treatment was 1.25, with a 95 percent confidence interval of 1.06 to 1.47, meaning fulvestrant patients faced a 25 percent higher risk of needing to switch therapies at any given time. In practical terms, the mTOR-based combination appeared to buy patients roughly two additional months before the next treatment change.

But the survival picture tells a more complicated story. When the researchers examined overall survival, the endpoint that ultimately matters most to patients, they found no statistically significant difference between the two cohorts. The hazard ratio for death was 1.09, with a 95 percent confidence interval of 0.87 to 1.36, a range that comfortably includes the possibility of no effect at all. Median overall survival estimates, calculated using the Kaplan–Meier method, were similar in both groups. In other words, while everolimus plus exemestane appeared to delay the next treatment change, it did not demonstrably extend how long patients lived. This disconnect between a surrogate endpoint and a hard clinical outcome is one of the most familiar tensions in oncology research, and the authors themselves flag it prominently in their conclusions.

The caution is warranted for several reasons. TTNT is a surrogate, not a direct measure of tumor control; a patient might switch treatments for reasons unrelated to disease progression, such as side effects, physician preference, or access issues. The study also relied on administrative claims data, which, while comprehensive, do not capture every clinical nuance that a physician would consider when choosing a therapy. Unmeasured confounders, such as performance status, tumor burden, or the precise biology of the resistant tumor, could still tilt the comparison. The authors used multivariate Cox regression to probe the robustness of their findings, but no statistical adjustment can fully compensate for variables that were never recorded in the database.

Despite these limitations, the study carries real weight for clinical practice, precisely because it draws on one of the largest unselected patient populations ever assembled for this question. Randomized trials in this setting tend to enroll highly selected patients who meet strict eligibility criteria, often excluding those with significant comorbidities or poor performance status. Real-world evidence, by contrast, reflects the messy diversity of actual clinical care, including elderly patients, those with multiple health problems, and those whose treatment choices were shaped by practical constraints. The Korean single-payer system adds another layer of value: because drug access is relatively uniform across the country, the comparison is less distorted by socioeconomic disparities than it might be in other health systems.

The findings land in a therapeutic landscape that is evolving rapidly. Since the study period, additional second-line options have emerged for patients whose disease progresses on CDK4/6 inhibitors, including PI3K inhibitors for tumors with PIK3CA mutations and, more recently, oral selective estrogen receptor degraders. Clinical guidelines from the National Comprehensive Cancer Network and other bodies have struggled to establish a clear standard of care for this population, in large part because head-to-head evidence has been so scarce. The Korean study, by directly comparing two of the most widely used strategies in a large unselected cohort, helps fill that evidence gap, even if it cannot deliver the definitive answer that only a randomized trial could provide.

For patients and their oncologists, the message is one of tempered optimism. Everolimus plus exemestane may offer a modest advantage in keeping patients on effective therapy longer, roughly two additional months before the next treatment change, which in the context of metastatic disease can be meaningful time. But the absence of a demonstrated overall survival benefit means the choice between these two strategies remains a judgment call, one that must weigh side effects, patient preferences, tumor biology, and the practical realities of each drug’s administration. The authors themselves urge cautious interpretation, noting that TTNT is a surrogate endpoint and that no statistically significant survival difference was observed. As the field awaits more definitive randomized evidence, this large real-world comparison adds a valuable piece to the puzzle of how best to treat metastatic breast cancer after CDK4/6 inhibitor failure, a question that touches the lives of patients around the world and that oncology has yet to fully resolve.

Subject of Research: Clinical outcomes of everolimus plus exemestane versus fulvestrant in patients with metastatic breast cancer after progression on cyclin-dependent kinase 4/6 inhibitors

Article Title: Clinical outcomes of everolimus plus exemestane versus fulvestrant in patients with metastatic breast cancer after progression on cyclin-dependent kinase 4/6 inhibitors

Article References: Oh, B.-C., Kim, S.-J., & Lee, E.-K. (2026). Clinical outcomes of everolimus plus exemestane versus fulvestrant in patients with metastatic breast cancer after progression on cyclin-dependent kinase 4/6 inhibitors. BMC Cancer. https://doi.org/10.1186/s12885-026-16981-5

Image Credits: AI Generated

DOI: 10.1186/s12885-026-16981-5

Keywords: Clinical, outcomes, everolimus, plus, exemestane, versus, fulvestrant, patients, metastatic, breast, cancer, progression