A phase III randomized clinical trial (RCT) of 1,050 patients with hormone receptor–positive (HR-positive), ERBB2-negative advanced breast cancer (ABC) showed that first-line (1L) use of cyclin-dependent kinase (CDK) 4 and 6 inhibitors did not improve overall survival (OS) compared with second-line (2L) use. At the same time, a post hoc subgroup analysis suggested an OS benefit with 1L use of CDK4/6 inhibitors in premenopausal patients.
CDK4/6 inhibitors effective in 1L and 2L
The addition of palbociclib, ribociclib, or abemaciclib to 1L (aromatase inhibitor) and 2L (fulvestrant) endocrine therapy has demonstrated a consistent improvement in progression-free survival (PFS), with similar results across all three CDK4/6 inhibitors. [N Engl J Med 2016;375:1925-1936; J Clin Oncol 2017;35:3638-3646; Lancet Oncol 2018;19:904-915; Lancet Oncol 2016;17:425-439; J Clin Oncol 2018;36:2465-2472; J Clin Oncol 2017;35:2875-2884]
In addition, a significant OS improvement was reported for ribociclib in the 1L and 2L settings and for abemaciclib in the 2L. [N Engl J Med 2022;386:942-950; N Engl J Med 2020;382:514-524; JAMA Oncol 2020;6:116-124]
Tendency to administer effective therapies early
Pivotal 1L trials did not include crossover to CDK4/6 inhibitor on progression in the placebo arm, despite evidence of efficacy in 2L setting. [Nat Rev Clin Oncol 2023;20:815-816]
“The 1L trials … did not evaluate whether 1L treatment meaningfully improves outcomes beyond the already established benefit in later lines. Nonetheless, CDK4/6 inhibitors have been adopted as 1L therapy in most international guidelines … [reflecting] a broader oncologic tendency to administer effective therapies early, to avoid the possibility of patients not receiving proven treatments later,” wrote the researchers. [Ann Oncol 2021;32:1475-1495; J Natl Compr Canc Netw 2023;21:594-608]
Phase III SONIA RCT: 1L vs 2L use of CDK4/6 inhibitors
The phase III SONIA RCT directly compared the strategy of using a CDK4/6 inhibitor in the 1L vs 2L. In the primary analysis, 1L CDK4/6 inhibitor use did not improve PFS after two lines of therapy and was associated with greater toxic effects, higher drug costs, and comparable health-related quality of life (HRQoL). [Nature 2024;636:474-480; J Natl Cancer Inst 2025;djaf334; Eur J Cancer 2025;231:116051]
“Given that OS and HRQoL are considered the most meaningful measures of patient benefit, the objective of this prespecified analysis was to evaluate whether 1L CDK4/6 inhibitor use improves OS compared with 2L in patients with HR-positive, ERBB2-negative ABC,” explained the researchers. [JAMA Oncol 2026;12:375-383]
Efficacy and toxicity
Between November 2017 and September 2021, 1,050 patients were randomized to the 1L CDK4/6 inhibitor group (n=524) or 2L CDK4/6 inhibitor group (n=526). Baseline characteristics were well balanced between treatment arms. Palbociclib was the CDK4/6 inhibitor of physicians’ choice in the treatment of most patients (91.2 percent), followed by ribociclib (8.2 percent), and abemaciclib (0.6 percent).
At a median follow-up of 58.5 months, the median OS was 47.9 months in the 1L group and 48.1 months in the 2L group (hazard ratio, 0.91; 95 percent CI, 0.77–1.07; p=0.24). OS results were consistent across all predefined subgroups, including CDK4/6 inhibitor type (palbociclib or ribociclib), with no test for interaction reaching statistical significance.
A post hoc subgroup analysis suggested an OS benefit with 1L use in premenopausal patients, but not in postmenopausal patients. In premenopausal patients (n=145), the median OS was not reached (NR) in the 1L group vs 35.3 months in the 2L group (hazard ratio, 0.53; 95 percent CI, 0.32–0.87).
Previous studies have shown that tumours in premenopausal patients often exhibit more aggressive biology, including a higher prevalence of the luminal B subtype, elevated Ki-67 values, and lower oestrogen receptor (ER) expression compared with postmenopausal patients. [Breast Cancer Res 2015;17:104; Clin Cancer Res 2018;24:5206-5218]
“This reduced ER dependence, along with enrichment of oncogenic drivers, such as BRCA1, BRCA2, and TP53, may explain a reduced response to endocrine monotherapy and greater relative benefit from early CDK4/6 inhibitor use in this population,” suggested the researchers. “Nonetheless, this subgroup analysis should be interpreted with caution given its post hoc nature, the small patient numbers, and the imbalance in baseline characteristics, including the proportion with visceral disease in premenopausal patients [1L group: 46.4 percent; 2L group: 65.8 percent].”
After a median follow-up of 58.5 months, the median PFS after two lines of therapy was 31.9 months in the 1L group and 26.7 months in the 2L group (hazard ratio, 0.82; 95 percent CI, 0.71–0.94; p=0.01).PFS results after two lines of therapy were consistent across predefined subgroups, with no test for interaction reaching statistical significance. In line with OS results, post hoc subgroup analyses showed PFS benefit after two lines of therapy with 1L vs 2L CDK4/6 inhibitor use in premenopausal patients (31.5 vs 19.4 months; hazard ratio, 0.56; 95 percent CI, 0.37–0.86), but not in postmenopausal patients.
“While surrogate endpoints like PFS and PFS after two lines of therapy may allow for earlier trial readouts, OS and HRQoL remain the most meaningful measures of patient benefit,” commented the researchers. [Lancet Oncol 2025;26:e80-e89; JAMA Oncol 2025;11:718-724]
Of patients who discontinued 2L study treatment, 84.8 percent in the 1L group and 84.2 percent in the 2L group received subsequent anticancer therapy, with comparable treatment patterns beyond 2L treatment in the two groups. The median chemotherapy-free survival was 34.4 months in the 1L group and 32.4 months in the 2L group (HR, 0.88; 95 percent CI, 0.76–1.02).
In total, 86.4 percent of patients in the 1L group and 75.7 percent of patients in the 2L group experienced at least one grade ≥3 adverse event (AE). Neutropenia was the most frequent adverse event reported.
Summary and suggestions
In this phase III RCT, 1L CDK4/6 inhibitor use did not improve OS compared with 2L use, but did increase treatment-related toxic effects.
“Post hoc analysis suggests that the impact of CDK4/6 inhibitor treatment sequencing may differ according to menopausal status, but these results should be interpreted with caution due to the inherent bias of post hoc analyses,” wrote the researchers.
“To better inform treatment sequencing, future trials evaluating drugs with established OS benefits in later-line settings should incorporate crossover and use OS or PFS after two lines of therapy as primary endpoints along with HRQoL,” they suggested.