The Absolute Effect / Evidence / Breast Cancer
PALLAS (ABCSG-42/AFT-05/BIG-14-03)
Clinical question: Does adding two years of adjuvant palbociclib to standard endocrine therapy improve invasive disease-free survival in stage II–III HR-positive/HER2-negative early breast cancer?
Registry: NCT02513394 · One underlying study; aliases and follow-up publications consolidated.
Design, population and eligibility
Phase III, international, open-label randomized trial
5,796 enrolled; 5,761 in the intention-to-treat population. Final analysis: 2,884 palbociclib plus endocrine therapy and 2,877 endocrine therapy alone.
Stage II–III HR-positive/HER2-negative early breast cancer after definitive local treatment; patients received standard adjuvant endocrine therapy.
Follow-up and endpoint maturity are stated with the reported efficacy results; further long-term updates are noted where applicable.
Treatment and comparator
Palbociclib 125 mg orally daily, days 1–21 of each 28-day cycle for 2 years plus standard endocrine therapy versus endocrine therapy alone. Palbociclib was discontinued early for futility after interim review.
Primary and secondary endpoints
Primary endpoint: iDFS. Secondary endpoints included invasive breast cancer-free survival, distant recurrence-free survival, locoregional cancer-free survival and OS.
Key efficacy results
At final protocol-defined analysis (median follow-up 31 months), 4-year iDFS was 84.2% with palbociclib plus endocrine therapy versus 84.5% with endocrine therapy alone; HR 0.96 (95% CI 0.81–1.14; P=0.65). No significant differences were observed in secondary time-to-event endpoints or subgroups.
Four-year OS was 93.8% versus 95.2%; this numerical difference was not statistically significant and should not be interpreted as treatment harm.
Source-reported outcome visualization
Four-year invasive disease-free survival.
Negative superiority result: no significant iDFS benefit. The numerical difference is not evidence of equivalence or harm.
Absolute effect
At 4 years, iDFS was 84.2% versus 84.5% (absolute difference −0.3 percentage points); HR 0.96 (95% CI 0.81–1.14; P=0.65). No NNT or NNH for efficacy is appropriate because superiority was not demonstrated. This is not a formal equivalence trial.
Trial-specific toxicity and treatment burden
Grade 3–4 neutropenia occurred in 61.3% with palbociclib versus 0.3% with endocrine therapy alone; grade 3–4 leukopenia 30.2% versus 0.1%. Serious AEs: 12.4% versus 7.6%. 44.9% did not complete the planned 2 years of palbociclib, mainly because of protocol-defined reasons including neutropenia. No treatment-related deaths were reported.
Quality of life and patient-reported outcomes
No QoL advantage was established. The additional two-year oral treatment, laboratory monitoring and high rate of treatment interruption/discontinuation are important burden in the absence of demonstrated efficacy benefit.
Subgroups and applicability
Final subgroup analyses did not identify a subgroup with a convincing iDFS benefit. PALLAS differs from NATALEE in agent, dose, duration and trial population; results must not be generalized as a CDK4/6 class effect.
Current-practice context · reviewed 3 November 2026
Adjuvant palbociclib is not recommended on the basis of PALLAS. The trial is essential negative evidence when discussing adjuvant CDK4/6 inhibitors and contrasts with positive monarchE and NATALEE data.
Limitations and critical interpretation
Early discontinuation for futility reduced planned exposure; median follow-up 31 months at final protocol-defined analysis. The result is a negative superiority test, not proof of equivalence. Subsequent follow-up does not rescue a benefit in the primary endpoint.
Bottom line: PALLAS showed no iDFS benefit from adding palbociclib to adjuvant endocrine therapy and a substantial neutropenia and treatment-discontinuation burden. It does not support adjuvant palbociclib.
Primary sources and follow-up publications
Outcome values are sourced from the linked reports. Unreported values are not imputed. NNT/NNH is only stated when an explicit fixed-time absolute difference supports a defensible calculation; no NNT is derived from median survival.