Part XI · Hormone receptor-positive, HER2-negative disease · Chapter 61
Adjuvant CDK4/6 inhibition
Which patient gets which agent for how long, and why the two negative trials never asked the question the two positive trials answered.
1 · Scope, and why four trials of one drug class gave different answers
Four randomised trials added a CDK4/6 inhibitor to adjuvant endocrine therapy in hormone receptor-positive, HER2-negative early breast cancer. Two were positive and two were negative. This chapter is about why. It treats that question as answerable in part rather than as a standing mystery.
The unhelpful version of the question is whether the class works. Reading the four as a single class experiment is the most consequential error available here, and Cross-trial comparison and what actually differs between the populations states the practice consequence in brief. This chapter does the comparison in full, axis by axis.
The table below carries five rows and they do not all ask one question. Four of them share a structure. Each enrolled a woman with no measurable disease, gave her the endocrine therapy she would have received anyway, and randomised whether a CDK4/6 inhibitor was added on top. Each was powered on invasive disease-free survival. None required a multigene assay or a molecular residual disease signal to enter. That shared structure is what makes the four comparable at all.
The fifth row is WSG ADAPTcycle, and it is not an addition trial. Its control arm is chemotherapy rather than endocrine therapy alone, so it asks whether the inhibitor can replace cytotoxic treatment rather than supplement endocrine treatment1. It appears here because it is an adjuvant CDK4/6 trial in the same subtype. It belongs to a different argument, and Unresolved questions and what reads out next is where that argument is made.
Among the four addition trials the design axes all move together. Agent, dose relative to the advanced-disease dose, schedule, duration of the inhibitor, the endocrine backbone underneath it and the risk threshold for entry differ between every pair in the set. No two of these trials differ on one axis alone. That is the single fact a reader needs before comparing any two rows, and it is the reason a drug-level explanation of the split is an assertion rather than a conclusion.
Three neighbouring chapters carry what is absent from this one. The pharmacology of abemaciclib, ribociclib and palbociclib is in Targeted therapy. Metastatic use, combination and sequencing are in CDK4/6 inhibition, combinations and sequencing. The endocrine backbone on which every one of these trials was built is in Adjuvant endocrine therapy. It is not a detail. The four trials do not agree on what that backbone was.
HR-positive/HER2-negative disease with residual invasive disease after taxane-containing neoadjuvant chemotherapy and a CPS-EG score of 3 or more, or 2 with ypN-positive disease
palbociclib 125 mg on days 1-21 of a 28-day cycle for 13 cycles plus endocrine therapy vs placebo plus endocrine therapy
intermediate-risk HR-positive, HER2-negative early breast cancer selected by a 21-gene recurrence score and by endocrine response measured as post-endocrine Ki-67; 5292 screened and 1684 randomised at 85 German sites
ribociclib plus an aromatase inhibitor, with a gonadotropin-releasing hormone agonist in premenopausal women, vs chemotherapy followed by endocrine therapy
2 · monarchE, its population and the durability question
monarchE defined its eligible population entirely on clinicopathological features. Four or more positive nodes qualified on their own. One to three nodes qualified with a tumour of 5 cm or more, grade 3 histology, or a central Ki-67 of 20% or above5.
Two structural consequences follow from defining the population that way. Ki-67 was one of several routes into the trial rather than the axis of any comparison within it. The trial therefore cannot say whether Ki-67 identifies the women who benefit. The separately reported Ki-67 analysis describes enrolled subgroups rather than testing prediction6. The second consequence is that entry was decided by anatomy and grade rather than by a multigene assay, so a low recurrence score does not exclude a patient who meets the criteria Clinical and genomic risk assessment.
The durability question bears on whether two years is the right number. Abemaciclib was given for two years. The natural expectation for a cytostatic agent is that its effect fades once the drug stops. That is not what the follow-up has shown. The hazard ratio for invasive disease-free survival was 0.75 at the preplanned interim analysis and 0.734 at the end-of-study analysis5,7. Those two estimates are the argument. A carry-over effect is the only reading under which the later number does not drift towards one.
A competing explanation exists and it is not obviously the weaker of the two. The highest-risk patients recur early and are removed from both arms, which flattens the later hazard in a way that mimics carry-over. The trial cannot separate the two. Neither can any other trial in this chapter, because none of them randomised the duration of treatment.
NATALEE differs from monarchE on four axes at once. Each difference was deliberate and each was defended at the time.
The population is broader at both ends of the risk range. Anatomical stage IIA with nodal involvement qualified, as did node-negative stage IIA with additional risk factors, along with stages IIB and III2. That admits patients monarchE would have excluded, and with them a substantially lower baseline event rate.
The dose is lower than the one licensed for the same drug in advanced disease. Ribociclib was given at 400 mg daily, on a three weeks on and one week off schedule. This is dose optimisation applied before registration rather than after it, which is the argument in Dose optimization and the move away from maximum tolerated dose. A well patient taking a drug for three years has a different tolerability requirement from a patient taking it until progression.
The duration is longer, at three years of ribociclib alongside 60 months of an aromatase inhibitor.
The endocrine backbone is the tightest of the four. A nonsteroidal aromatase inhibitor was mandated, with goserelin for premenopausal women and for men8. Every other trial in this chapter allowed the treating physician to choose. That matters when NATALEE is read against PALLAS, because the two differ in what the inhibitor was added to as well as in which inhibitor it was.
Exposure is the fact most often left out of the comparison. At the final preplanned invasive disease-free survival analysis, ribociclib had been stopped in 78.3% of the patients assigned to it2. Only 42.8% had completed the planned three years. The reported hazard ratio is an intention-to-treat estimate in a trial where fewer than half of the experimental arm finished the experimental treatment.
Overall survival remained immature at the five-year report9. The trial has now reported at several cut-offs with lengthening follow-up10. Why absolute rates cannot be compared across those cuts is explained in Absolute and relative benefit, and why the difference matters. That is why the hazard ratio and its comparison are the figures to quote here.
4 · PALLAS and PENELOPE-B, and why they were negative
The two negative trials probably failed for different reasons. One of them is a genuine counter-experiment to monarchE. The other is not a counter-experiment at all.
PALLAS is the closest thing to a direct contradiction of monarchE. It matched monarchE on duration at two years, on an addition design, and on a physician's-choice endocrine backbone. It enrolled stage II to III disease with no high-risk qualifier, gave palbociclib, and found nothing3. Its primary analysis came at a median follow-up of 31 months, so its four-year estimates rest on relatively little follow-up beyond the treatment period itself. The five-year report did not move it. The overall survival point estimate there sat on the control side of one without reaching significance11. That is worth knowing, because an underpowered null and a point estimate pointing the wrong way are different states.
PENELOPE-B is not a contradiction at all, because it studied a different disease state. Every patient had residual invasive disease after taxane-containing neoadjuvant chemotherapy, with a clinical-pathological staging score of 3 or more, or 2 with node-positive disease4. That is chemotherapy-refractory disease identifiable in the surgical specimen, a state described in Minimal residual disease as a measurable state. It is not the same population as a woman rendered disease-free by surgery and offered treatment against an invisible residue.
Its endocrine backbone was also the weakest of the four, because tamoxifen alone was a permitted partner4. A CDK4/6 inhibitor works with endocrine blockade rather than instead of it. A trial in which a share of patients took tamoxifen alone is testing a weaker combination than one in which every patient received an aromatase inhibitor. Its duration was also the shortest of the four, at one year.
The two negative trials used the same drug and their long-term survival estimates fall on opposite sides of one. PENELOPE-B's final survival report favoured palbociclib numerically and did not reach significance12. PALLAS pointed the other way and also did not reach significance. Neither estimate is a finding. Together they are a demonstration of what non-significant survival estimates in two different populations look like, and a reason not to read either as a statement about the molecule.
Caution
A negative trial in a population that differs on four axes cannot tell you which axis mattered. PENELOPE-B is routinely cited as evidence that CDK4/6 inhibition does not work in post-neoadjuvant residual disease. What it shows is that one year of palbociclib on a mixed endocrine backbone did not work in that population.
A woman with residual node-positive disease after neoadjuvant chemotherapy who meets the monarchE nodal criteria sits inside monarchE's enrolled population, because prior neoadjuvant chemotherapy was permitted there5. PENELOPE-B does not remove her from consideration for abemaciclib.
5 · Cross-trial comparison and what actually differs between the populations
Early disease · HR+/HER2- · adjuvant
Choosing the intensity and the duration of endocrine treatment
Does adding a CDK4/6 inhibitor to adjuvant endocrine therapy lower recurrence?
Rationale
CDK4/6 inhibition roughly doubles progression-free survival in metastatic hormone receptor-positive disease, and the cells that seed a late recurrence are the same lineage. If the drug works on macroscopic disease it should work on residual disease.
Experiment
Four trials added a CDK4/6 inhibitor to adjuvant endocrine therapy, differing in the drug, the duration, and how high the enrolled risk was.
Finding
Two showed a durable improvement in invasive disease-free survival and two did not, and the split follows the specific drug, schedule and risk population rather than the class.
Limitation
These are not head-to-head trials. The populations, the treatment durations and the endpoints differ enough that the divergence cannot be attributed to the molecule alone. Overall survival has separated in one of the four. It is negative in one, immature in another and unreported in the last. The survival record therefore follows the same split the disease-free record does, and adds no independent evidence about which feature of these trials produced it.
Next question
If escalation only pays at high risk, what is the risk threshold, and does it have to be measured before treatment or can it be read from the response to treatment?
monarchENATALEEPALLASPENELOPE-B
The story above carries no figures, and it is one node of the [Chapter Stories](stories.html) map. The table below carries the figures and nothing else.
HR-positive/HER2-negative disease with residual invasive disease after taxane-containing neoadjuvant chemotherapy and a CPS-EG score of 3 or more, or 2 with ypN-positive disease
palbociclib 125 mg on days 1-21 of a 28-day cycle for 13 cycles plus endocrine therapy vs placebo plus endocrine therapy
invasive disease-free survival
stratified HR 0.93 (95% repeated CI 0.74-1.17), p=0.525, with 308 events at a median follow-up of 42.8 months
7-year overall survival 86.8% vs 85.0%, HR 0.842 (0.722-0.981), p=0.027
One thing in that table is agreement rather than division. Where a benefit appeared it was of similar relative size. The two positive trials returned hazard ratios for invasive disease-free survival within a few hundredths of each other. They did so across different drugs, doses, schedules and durations, in populations with very different event rates. Whatever separates the two positive trials from the two negative ones, it is not the size of the effect a CDK4/6 inhibitor produces when it produces one.
Saying that the trials differed is not an analysis. What follows is what differed, axis by axis, with the trials named on each.
Eligibility risk: monarchE required node-positive disease with an additional high-risk feature. NATALEE admitted stage IIA including selected node-negative tumours. PALLAS took stage II to III with no high-risk qualifier. PENELOPE-B took residual disease after neoadjuvant chemotherapy. The baseline event rate therefore differs substantially across the set, and so does the absolute gain that any given hazard ratio can produce.
Agent: abemaciclib in the trial with mature survival data, ribociclib in the other positive trial, palbociclib in both negative trials. The three differ in relative CDK4 and CDK6 selectivity and in which toxicity limits the dose.
Dose relative to the advanced-disease dose: abemaciclib and palbociclib were given at their full advanced-disease doses. Ribociclib was given below its advanced-disease dose.
Schedule: abemaciclib is dosed continuously and twice daily. Ribociclib ran three weeks on and one week off. Palbociclib ran 21 days of every 28 in both of the trials that used it. Continuous dosing means uninterrupted target engagement. The cell cycle resumes during a break, and no trial has tested whether that matters in this setting.
Duration: one year in PENELOPE-B, two in monarchE and PALLAS, three in NATALEE.
Endocrine backbone: NATALEE mandated a nonsteroidal aromatase inhibitor, with goserelin for premenopausal women and for men. PENELOPE-B permitted tamoxifen alone. monarchE and PALLAS both allowed physician's choice of endocrine agent.
Exposure when the events accrued: this is the axis nobody tabulates. NATALEE's final analysis arrived with most of its ribociclib arm already off the drug. PALLAS reported at 31 months of median follow-up against two years of planned treatment.
Disease state: PENELOPE-B's population had been selected by chemotherapy for survival through cytotoxic pressure. Whether that also selects against dependence on cell cycle machinery is unknown. It is biologically plausible and it has never been looked for.
The decision in clinic rests on monarchE and not on the set. Its control arm is exactly what the clinic would otherwise do for a high-risk node-positive woman, which is endocrine therapy alone. It is also the only one of the four to report a significant overall survival difference. NATALEE is load-bearing only for the patients monarchE refused. There it is the sole randomised evidence rather than the better of two.
Horizons are what make the four result cells incomparable. monarchE's difference in invasive disease-free survival is measured at seven years, in a population selected for high clinicopathological risk7. NATALEE's is measured at three years in a broader population with survival still immature2. A hazard ratio near 0.75 in a population with a high seven-year event rate and the same ratio in a lower-risk population describe experiences that differ entirely for the woman taking the drug. Absolute and relative benefit, and why the difference matters gives the general form of that translation.
Four things the table has no column for decide how it should be read. Nobody randomised two agents against each other. Nobody randomised duration. Nobody randomised the endocrine backbone, which means the backbone difference travels with the agent difference in every comparison. And nobody enrolled on a genomic assay or a molecular residual disease signal, so the table says nothing about whether the benefit concentrates in a subgroup that could be identified in advance.
Caution
No trial varied one axis while holding the others fixed. Nothing in this evidence identifies which variable explains the split, so any confident attribution is an assertion rather than an inference.
The specific assertion to resist is that palbociclib is pharmacologically inferior. That may be true. It is equally compatible with the data that two years of an intermittently dosed agent, on a physician's-choice backbone, analysed at 31 months, was simply a weaker experiment than two years of a continuously dosed agent analysed at seven years. Those two hypotheses have not been separated. No trial now running would separate them either.
6 · Patient selection in practice
The evidence supports a single rule, applied below to the patients who actually turn up.
In practice
Use the agent in the population in which it was tested, for the duration for which it was tested.
Node-positive high-risk disease meeting the monarchE criteria: two years of abemaciclib, with a mature overall survival benefit7. This is the only adjuvant CDK4/6 indication with a survival result behind it. The absolute figures are an overall survival difference of 1.8 percentage points at seven years alongside a 6.5-point recurrence difference. Quote both rather than the hazard ratios, which are what a patient will otherwise hear.
Stage IIA to III including node-negative disease with risk factors, not meeting monarchE criteria: three years of ribociclib at 400 mg, with a recurrence benefit and immature survival data2. Say that the survival question is open rather than implying it is answered.
A patient meeting both sets of criteria: there is no randomised comparison. Choose on tolerability, on duration and on what she will actually take, and tell her that this is the basis.
Residual node-positive disease after neoadjuvant chemotherapy meeting monarchE's nodal criteria: abemaciclib is supported, because that patient was eligible for monarchE. PENELOPE-B tested a different drug for a different duration.
Low genomic risk with qualifying anatomy: neither positive trial selected on a multigene assay, so a low recurrence score does not disqualify. Those assays estimate prognosis under endocrine therapy, not benefit from cell cycle inhibition Prognostic and predictive assays and somatic profiling.
Do not substitute palbociclib on grounds of access or familiarity. It has been tested twice after surgery in this subtype3,4. Neither trial met its primary endpoint.
Two points belong to the conversation rather than to the algorithm. Adding a CDK4/6 inhibitor does not shorten the endocrine course. NATALEE ran its aromatase inhibitor for 60 months alongside three years of ribociclib. The patient is agreeing to an addition rather than a substitution. The second point is that the decision is made in a woman who is well, has no measurable disease, and will experience only the toxicity with certainty. Communication and shared decision-making sets the register that conversation needs and the framing that makes it honest.
7 · Toxicity, discontinuation, and the cost of two or three years
No table in this chapter carries a toxicity column. In the adjuvant setting that omission sits at the centre of the decision rather than at its edge. The agent-level detail belongs to Targeted therapy and is not repeated here. What is specific to this setting is the arithmetic of discontinuation.
NATALEE documented that arithmetic more precisely than the other three trials did. Only 42.8% of patients allocated to ribociclib completed the planned three years2. That is the adherence under which the reported hazard ratio was generated. It is not a defect in the trial or a reason to discount it. It is the trial telling you what three years of a drug costs a patient who feels well.
Two consequences follow for what you say to a patient. The benefit you quote already incorporates real discontinuation, so it should not be presented as the benefit of completing the course. And a patient who stops at eighteen months has not forfeited the trial's result, because a large fraction of the trial population did something similar.
Dose reduction is designed into one of these regimens from the start. NATALEE's 400 mg is already below the advanced-disease dose, chosen because the tolerability requirement changes when the intent is cure and the horizon is three years Dose optimization and the move away from maximum tolerated dose. Abemaciclib's diarrhoea is early in the course and responds to dose reduction. Dose reduction rather than discontinuation is the intended response to it.
Caution
Monitoring burden is part of the cost and is usually left out of the consent conversation. Ribociclib carries scheduled electrocardiographic and hepatic monitoring. Abemaciclib requires counselling about diarrhoea before the first dose rather than after the first episode.
A well woman is being asked to accept a certain toxicity and a fixed monitoring schedule for a probabilistic benefit. Quote both in the same units and over the same horizon, as Absolute and relative benefit, and why the difference matters requires.
intermediate-risk HR-positive, HER2-negative early breast cancer selected by a 21-gene recurrence score and by endocrine response measured as post-endocrine Ki-67; 5292 screened and 1684 randomised at 85 German sites
ribociclib plus an aromatase inhibitor, with a gonadotropin-releasing hormone agonist in premenopausal women, vs chemotherapy followed by endocrine therapy
invasive disease-free survival
not yet reported; local against central agreement for post-endocrine Ki67 at or below 10% was 80.5%, 95% CI 79.0 to 82.1
oestrogen receptor-positive HER2-negative luminal A or B disease
letrozole plus ribociclib for 6 months, single group, with serial biopsy and single-cell RNA sequencing
response to the combination with serial intratumoural profiling
not yet reported
not yet reported
One open question has no trial at all behind it and is met weekly. A germline BRCA carrier with high-risk node-positive disease may be eligible for adjuvant olaparib and for adjuvant abemaciclib. OlympiA did not permit a CDK4/6 inhibitor and monarchE did not require germline testing, so neither trial observed the combination or the sequence14. Choosing one, choosing the other, or giving them in series are three strategies with no comparative evidence behind any of them. A patient offered a definite order should be told that the order is a judgement. The olaparib evidence itself is in Adjuvant PARP inhibition for germline carriers.
Six questions are open at the time of writing. The table above holds two trials. One of them speaks to the first question. The other speaks to none of them.
WSG ADAPTcycle asks about substitution rather than addition. It randomises ribociclib with an aromatase inhibitor against chemotherapy followed by endocrine therapy, in intermediate-risk disease selected by a 21-gene recurrence score and by endocrine response measured as post-endocrine Ki-671. A positive result would move the class from an add-on to a chemotherapy substitute. That would change the conversation in exactly the group where the genomic assays are least decisive Genomic assays and the decision to omit chemotherapy.
Nothing about its efficacy is on the record. What is on the record is a measurement problem that will decide whether the result can be used. The endocrine response call is a Ki-67 threshold at 10%, read after short preoperative endocrine therapy. The agreement between local and central scoring of that call is in the table above. What the table cannot show is the direction of the disagreement. Among patients called responders locally, 13.1% were non-responders centrally1. Selection by response is a different kind of rule from selection by anatomy, because it depends on a proliferation index scored twice in the same patient. A positive trial read on central scoring would need central scoring to be exported with it.
NEOLETRIB is in the table because it is an early-setting CDK4/6 trial still running. It is not an adjuvant trial and will settle no adjuvant question. It gives letrozole and ribociclib for six months before surgery, in a single group, with serial biopsies and single-cell RNA sequencing of the tumour13. Its value to this chapter is that it samples what the adjuvant trials structurally cannot. An adjuvant trial treats disease that by definition cannot be biopsied, so it can measure recurrence and nothing else about the mechanism. A single-group design cannot attribute anything it observes to ribociclib rather than to letrozole, and six months of visible luminal disease is not two years of microscopic residue.
Overall survival in NATALEE remains immature9. A null result would be harder to excuse here than in metastatic disease, because the adjuvant setting has no crossover to dilute it Neoadjuvant, adjuvant, and metastatic design compared. That makes the readout informative in both directions, which is unusual.
Duration has no trial anywhere in the current portfolio. Nothing compares two years against three years of the same agent. The gap between monarchE and NATALEE confounds duration with agent, dose, schedule, backbone and population at the same time.
Agent has no trial either, and a head-to-head comparison of the three drugs is unlikely ever to be run.
Selection by molecular residual disease is untested. Nobody has randomised adjuvant CDK4/6 inhibition on a molecular residual disease signal. The arguments for and against doing so are in Liquid biopsy and longitudinal monitoring.
Re-treatment is the question a relapsing patient raises first. A woman who recurs after adjuvant abemaciclib has no randomised evidence for first-line CDK4/6 inhibition in metastatic disease, and CDK4/6 inhibition, combinations and sequencing sets out what is and is not known.
Interplay
Two or three years of cell cycle inhibition applied to undetectable disease is a selection experiment with no readout. The RB1 loss and CCNE1 amplification that define CDK4/6 resistance in metastatic disease CDK4/6 inhibition and the evolution of cell cycle escape are being selected for during the adjuvant years, in a population that cannot be sampled. The durability seen in monarchE after treatment stops is the one observation that argues against pure suppression, because a merely suppressed population should regrow. Whether these years instead change which resistance route is available at relapse is the question CDK4/6 inhibition, combinations and sequencing will eventually have to answer. NEOLETRIB is collecting serial tissue under the same drug class, but before surgery and in visible disease, which is the opposite of the setting where the selection happens.
Hamann M, Christgen M, Gluz O, et al. Endocrine response assessment in HR-positive HER2-negative early breast cancer: concordance of local versus central Ki67 measurements in the WSG ADAPTcycle trial (n = 5292). ESMO Open 2025 10:105552. PMID 40815984
Hortobagyi GN, Lacko A, Sohn J, et al. A phase III trial of adjuvant ribociclib plus endocrine therapy versus endocrine therapy alone in patients with HR-positive/HER2-negative early breast cancer: final invasive disease-free survival results from the NATALEE trial. Ann Oncol 2025 36:149-157. PMID 39442617
Gnant M, Dueck AC, Frantal S, et al. Adjuvant palbociclib for early breast cancer: the PALLAS trial results (ABCSG-42/AFT-05/BIG-14-03). J Clin Oncol 2022 40:282-293. PMID 34874182
Loibl S, Marmé F, Martin M, et al. Palbociclib for residual high-risk invasive HR-positive and HER2-negative early breast cancer, the Penelope-B trial. J Clin Oncol 2021 39:1518-1530. PMID 33793299
Johnston SRD, Harbeck N, Hegg R, et al. Abemaciclib combined with endocrine therapy for the adjuvant treatment of HR+, HER2-, node-positive, high-risk, early breast cancer (monarchE). J Clin Oncol 2020 38:3987-3998. PMID 32954927
Harbeck N, et al. Adjuvant abemaciclib combined with endocrine therapy for high-risk early breast cancer: updated efficacy and Ki-67 analysis from the monarchE study. Ann Oncol 2021 32:1571-1581. PMID 34656740
Johnston S, Martin M, O'Shaughnessy J, et al. Overall survival with abemaciclib in early breast cancer. Ann Oncol 2026 37:155-165. PMID 41110697
Slamon D, Lipatov O, Nowecki Z, et al. Ribociclib plus endocrine therapy in early breast cancer. N Engl J Med 2024 390:1080-1091. PMID 38507751
Crown J, Stroyakovskii D, Yardley DA, et al. Adjuvant ribociclib plus nonsteroidal aromatase inhibitor therapy in patients with HR-positive/HER2-negative early breast cancer: 5-year follow-up of NATALEE efficacy outcomes and updated overall survival. ESMO Open 2025 10:105858. PMID 41320342
Fasching PA, Stroyakovskiy D, Yardley DA, et al. Ribociclib plus endocrine therapy in hormone receptor-positive/ERBB2-negative early breast cancer: 4-year outcomes from the NATALEE randomized clinical trial. JAMA Oncol 2025 11:1364-1372. PMID 40996773
Mayer EL, et al. Palbociclib with adjuvant endocrine therapy in early breast cancer: 5-year follow-up analysis of the global multicenter, open-label, randomized phase III PALLAS trial (ABCSG-42/AFT-05/PrE0109/BIG-14-13). Ann Oncol 2025 37:271-277. PMID 41110701
Loibl S, et al. Final survival results from the PENELOPE-B trial investigating palbociclib versus placebo for patients with high-risk HR+/HER2- breast cancer and residual disease after neoadjuvant chemotherapy. Ann Oncol 2025 36:832-837. PMID 40139460
Fjermeros K, Ghannoum S, Geisler SB, et al. The NEOLETRIB trial: neoadjuvant treatment with letrozole and ribociclib in ER-positive, HER2-negative breast cancer. Future Oncol 2024 20:2457-2466. PMID 39073142
Tutt ANJ, Garber JE, Kaufman B, et al. Adjuvant olaparib for patients with BRCA1- or BRCA2-mutated breast cancer. N Engl J Med 2021 384:2394-2405. PMID 34081848