The research team, led by Xinwei Wu, Danni Wang, Yangbin Shi and senior authors Baohui Han and Wei Zhang, analyzed the records of 512 patients with previously untreated stage III or IV EGFR-mutant non-small cell lung cancer who began therapy between January 2017 and December 2023. Because patients were not randomly assigned to treatment groups, the investigators used propensity-score matching, a statistical technique that pairs patients across groups according to similar baseline characteristics such as age, sex, smoking history, disease stage and mutation type. After this 1:2 matching, 364 patients formed the balanced comparison cohort, allowing the team to estimate treatment effects while minimizing the confounding that plagues observational data.

The biological logic behind the two competing strategies is worth unpacking. First-generation inhibitors such as gefitinib and erlotinib reversibly block the mutated EGFR signaling hub that drives tumor growth, while second-generation drugs like afatinib bind irreversibly and target a broader family of related receptors. Third-generation inhibitors, exemplified by osimertinib, were engineered to selectively inhibit the most common sensitizing EGFR mutations while sparing the normal receptor, and to remain active against the T790M resistance mutation that frequently emerges under pressure from earlier drugs. In clinical guidelines, third-generation monotherapy is the recommended first-line standard, but in practice — particularly in health systems where the newer drug is expensive or not reimbursed — many physicians still prescribe older inhibitors in combination regimens, for example paired with chemotherapy or with a second targeted agent, in an attempt to squeeze out additional efficacy.

The access problem is far from marginal. As the authors note, as of October 2024 first-line osimertinib was not routinely reimbursed in five Asian countries, meaning that a substantial global population of patients faces exactly the choice this study examined. Direct head-to-head real-world comparisons between the two strategies had been scarce, which is what makes the new analysis a meaningful addition to the evidence base: it reflects the messy realities of clinical practice rather than the carefully curated conditions of a randomized trial.

On the primary endpoint of progression-free survival — the time a patient lives without the cancer growing or spreading — the third-generation monotherapy group clearly came out ahead. Median progression-free survival reached 20.5 months compared with 16.0 months for the combination approach, a difference that translated into a 37 percent reduction in the risk of progression or death (hazard ratio 0.63, 95 percent confidence interval 0.50 to 0.80, P less than 0.001). Importantly, the benefit was not confined to a narrow slice of patients. It held across both sexes, among smokers and never-smokers, in stage III and stage IV disease alike, in older patients, in those with good performance status, in patients with adenocarcinoma histology, in those whose tumors carried the exon 19 deletion — the most common sensitizing EGFR alteration — and in patients without brain metastases at baseline.

Tumor shrinkage, by contrast, told a different story. The objective response rate, which measures the proportion of patients whose tumors shrink measurably, was 63.1 percent with third-generation monotherapy versus 59.6 percent with combination therapy, a difference that was not statistically significant. The disease control rate — encompassing responses plus stable disease — was essentially identical at 99.2 percent and 99.5 percent. In other words, both strategies could suppress the cancer effectively in the short term; what separated them was durability, the length of time that control lasted before resistance emerged and the disease resumed its advance.

The most provocative finding concerned overall survival. Median overall survival was 46.1 months with third-generation monotherapy and 41.6 months with the combination strategy, a gap of four and a half months that nonetheless failed to reach statistical significance (hazard ratio 0.95, 95 percent confidence interval 0.72 to 1.26, P equal to 0.721). The explanation, the authors argue, lies in the sequencing patterns captured in the data: among patients who started on the older combination, 70.5 percent went on to receive a third-generation inhibitor in a later line of treatment after progression. Those who made that switch lived significantly longer than those who did not. In effect, the newer drug delivered much of its survival benefit even when it arrived second — a phenomenon that blurs the difference between the two initial strategies when the endpoint is measured over the entire course of care.

Tolerability data added a further consideration. Adverse-event-related discontinuation and dose reduction were both more frequent in the combination therapy group, suggesting that stacking an older EGFR inhibitor with additional agents extracts a price in side effects and treatment interruptions. For a therapy intended to be taken continuously for many months, the ability of patients to stay on the drug at full intensity is not a trivial matter; it shapes both quality of life and the effective dose delivered to the tumor over time.

The authors’ conclusion is carefully calibrated. Their findings support third-generation EGFR inhibitors as the preferred first-line option where they are accessible, reinforcing existing guideline recommendations with real-world evidence. At the same time, they acknowledge that for selected patients — particularly in settings with limited reimbursement or high out-of-pocket costs — first- or second-generation combination therapy remains a reasonable alternative, especially given that later switching to a third-generation drug can still be associated with extended survival. The retrospective design and the reliance on propensity-score matching mean residual confounding cannot be excluded, and the sequencing effect on overall survival complicates any simple interpretation of the survival endpoint. Still, for clinicians and health policymakers weighing the cost of expanding access to newer targeted agents, the study offers a quantified answer: the newer drug buys more time before progression, with fewer treatment-limiting side effects, and its benefit persists even when it is deployed later in the treatment sequence.

Beyond the immediate clinical question, the study speaks to a broader tension in modern oncology: the gap between what trials demonstrate in ideal conditions and what patients actually receive across diverse health systems. As targeted therapies proliferate, sequencing has become as consequential as selection, and evidence like this — drawn from hundreds of consecutive real patients over seven years — helps illuminate how the order of the queue shapes outcomes. For the growing number of countries still negotiating reimbursement for third-generation EGFR inhibitors, the message from Shanghai is that the wait carries a measurable cost in progression-free time, even if clever sequencing can partially, but not entirely, recover the ground later.

Subject of Research: Comparative effectiveness of first-/second-generation versus third-generation EGFR-TKI therapy in untreated advanced EGFR-mutant non-small cell lung cancer

Article Title: Comparative effectiveness of first-/second-generation EGFR-TKI combination therapy versus third-generation EGFR-TKI monotherapy in untreated locally advanced or metastatic EGFR-mutant NSCLC: a real-world, retrospective study in China

Article References: Wu, X., Wang, D., Shi, Y., Liu, B., Qian, F., Zhao, T., Lu, J., Zhang, Y., Lou, Y., Han, B., & Zhang, W. (2026). Comparative effectiveness of first-/second-generation EGFR-TKI combination therapy versus third-generation EGFR-TKI monotherapy in untreated locally advanced or metastatic EGFR-mutant NSCLC: a real-world, retrospective study in China. BMC Medicine. https://doi.org/10.1186/s12916-026-05340-7

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05340-7

Keywords: EGFR-mutant NSCLC, EGFR tyrosine kinase inhibitors, osimertinib, third-generation TKI monotherapy, combination therapy, progression-free survival, overall survival, propensity-score matching, real-world evidence, targeted therapy sequencing, lung cancer, drug reimbursement