Joyce O’Shaughnessy, MD:
We will start by discussing new analyses of the phase III KEYNOTE-756 and CheckMate 7FL trials. These studies are both evaluating perioperative ICI-based regimens in high-risk ER-positive/HER2-negative EBC.

The first analysis assessed pCR rates across patient subgroups in KEYNOTE-756.¹ The first interim analysis of this trial was presented at ESMO 2023.² In brief, KEYNOTE-756 enrolled 1278 patients with grade 3, stage II/III ER-positive/HER2 negative invasive ductal breast carcinoma. During the neoadjuvant phase, patients received pembrolizumab or placebo for 4 cycles plus paclitaxel for 12 weeks, followed by pembrolizumab or placebo plus either doxorubicin and cyclophosphamide or epirubicin and cyclophosphamide for 4 cycles. After surgery, patients received adjuvant pembrolizumab or placebo for 6 months, with endocrine therapy given for up to 10 years.

The coprimary endpoints are the rates for pCR, defined as ypT0/Tis ypN0, and event-free survival (EFS).

Joyce O’Shaughnessy, MD:
At ESMO 2023, Cardoso and colleagues² presented the results of the interim analysis, in which KEYNOTE-756 met its pCR endpoint. The pCR rate was 24.3% in the pembrolizumab arm vs 15.6% in the placebo arm—a statistically significant absolute difference of 8.5% (95% CI: 4.2%-12.8%; P = .00005).¹

Joyce O’Shaughnessy, MD:
On behalf of Cardoso and colleagues,¹ I presented data on pCR rates across patient subgroups at SABCS 2023. The pCR benefit with the addition of pembrolizumab increased as PD-L1 expression increased. We see an absolute difference of 9.8% in those with a PD-L1 CPS ≥1, which increases to 13.2% with CPS ≥10 and 17.4% with CPS ≥20.

This “stairstep” relationship between benefit and PD-L1 CPS in ER-positive/HER2-negative EBC is particularly notable because these results contrast with those reported in early-stage TNBC. In the phase III KEYNOTE-522 trial, the absolute increase in the pCR rates did not vary by PD-L1 CPS in patients with early TNBC when pembrolizumab was added to neoadjuvant chemotherapy and then continued as monotherapy in the adjuvant phase.³ Patients with early TNBC and a PD-L1 CPS <1 had an absolute improvement in pCR rate of 18.3% with pembrolizumab, whereas those with a PD-L1 CPS ≥20 had an absolute improvement of 18.5%.

Joyce O’Shaughnessy, MD:
One of the more important clinical findings in this subgroup analysis of KEYNOTE-756 was the greater benefit in patients with a PD-L1 CPS ≥1 and ER expression <10%.¹ In this subgroup, the addition of pembrolizumab was associated with an absolute difference in pCR rate of 24.2% (95% CI: 1%-45.1%) vs the addition of placebo. The pCR rate with pembrolizumab was 57.6%, which is closer to the pCR rate of 64.8% reported with pembrolizumab in early TNBC.³ By contrast, in KEYNOTE-756, the subgroup with a PD-L1 CPS <1 and ER expression ≥10% had an absolute difference in pCR rate of 4.6%.¹

Another interesting finding was that patients derived more absolute benefit from receiving pembrolizumab when they had “full exposure” to the chemotherapy regimen, defined as 10-12 doses of paclitaxel weekly and all 4 cycles of anthracycline and cyclophosphamide, vs less than full exposure.¹ Among those with full exposure, the pCR rate with placebo was 16.9% vs 26.2% with pembrolizumab, for an absolute difference of 9.3%. Among those with less than full exposure, the pCR rates were 6.4% vs 13.2%, respectively, for an absolute difference of 6.8%.

Joyce O’Shaughnessy, MD:
No new safety signals were observed in this trial.¹

Joyce O’Shaughnessy, MD:
Next, we will discuss an exploratory biomarker analysis of the phase III CheckMate 7FL trial, which I consider to be a “sister” trial to KEYNOTE-756. CheckMate 7FL enrolled 521 patients with high-risk ER-positive/HER2-negative EBC that was clinical stage II//III and either grade 3 with ER expression ≥1% or grade 2 with ER expression 1% to 10%.⁴ During the neoadjuvant phase, patients received nivolumab or placebo plus paclitaxel for 4 cycles, followed by nivolumab or placebo plus an anthracycline and cyclophosphamide for 4 cycles. After surgery, patients received adjuvant nivolumab or placebo plus endocrine therapy.

After abemaciclib received approval in the adjuvant setting, the trial design for CheckMate 7FL was amended in April 2022 out of concern that a CDK4/6 inhibitor could not be given safely with an anti–PD-1 agent. The primary endpoint was changed to only include pCR, defined as ypT0/Tis ypN0, and no further patients were enrolled after 521 participants were randomized. The 11 patients enrolled at the Russian site—which was closed because of the ongoing conflict in that region—were excluded from analysis, and the modified ITT population consisted of 510 patients.

Joyce O’Shaughnessy, MD:
Loi and colleagues⁵ first presented results from CheckMate 7FL at ESMO 2023, where they reported that the addition of nivolumab significantly improved the pCR rate by 10.5% (95% CI: 4.0%-16.9%; P = .0021) vs placebo in the modified ITT population. This absolute benefit in the pCR rate is similar to that reported with the addition of pembrolizumab in KEYNOTE-756.¹

In the biomarker analysis presented at SABCS 2023, we saw again a “stairstep” relationship between increasing pCR benefit with an anti–PD-1 agent and increasing PD-L1 expression in the setting of ER-positive/HER2-negative EBC.⁴ Of interest, the absolute difference in pCR rate with nivolumab vs placebo was 52.3% in patients with PD-L1 CPS ≥20.

Joyce O’Shaughnessy, MD:
This analysis also reported greater benefit in patients with grade 3 disease and ER expression ≤50%, where the absolute difference in pCR rate was 29.3% vs 7.4% in those with ER expression >50%.⁴

Sara A. Hurvitz, MD:
Continuing our discussion of studies in the setting of hormone receptor–positive/HER2-negative EBC, Hortobagyi and colleagues⁷ presented the final invasive disease–free survival (iDFS) analysis of the open-label phase III NATALEE trial at SABCS 2023. NATALEE is comparing the addition of the CDK4/6 inhibitor ribociclib for 3 years to adjuvant endocrine therapy with a nonsteroidal aromatase inhibitor (NSAI) vs NSAI alone in 5101 patients with intermediate-/high-risk hormone receptor–positive/HER2-negative EBC. Of note, ribociclib was started at 400 mg daily in this trial—a lower dose than the 600 mg indicated for ABC.⁸

The final protocol-specified iDFS analysis was performed after approximately 500 events,⁷ updating the interim analysis of iDFS presented earlier in 2023.⁹

Sara A. Hurvitz, MD:
With a median follow-up of 33.3 months, the 3-year iDFS rate was significantly improved with the addition of ribociclib to adjuvant NSAI vs NSAI alone.⁷ The 3-year iDFS rate was 90.7% with ribociclib plus NSAI vs 87.6% with NSAI alone, yielding an absolute improvement of 3.1% and a statistically significant HR of 0.749 (95% CI: 0.628-0.892; nominal 1-sided P = .0006). This absolute benefit in the 3-year iDFS rate is greater than the 1.5% benefit reported at 2 years.

Sara A. Hurvitz, MD:
I found the subgroup analyses of iDFS benefit to be particularly important. Recall that the monarchE trial evaluating adjuvant abemaciclib plus endocrine therapy enrolled patients with high-risk disease in this setting, whereas NATALEE permitted patients with node-negative stage II disease to enroll.^7,10

The NATALEE investigators presented an informative breakdown of iDFS based on stage and nodal status.

Of interest, the HR—and therefore the iDFS benefit—with adjuvant ribociclib was fairly consistent for stage II vs III and node-positive vs node-negative disease. Of course, patients with more advanced disease had worse iDFS outcomes, but the absolute benefit from ribociclib was consistent across subtypes defined by stage and nodal status.

Sara A. Hurvitz, MD:
With a median follow-up of 35.9 months for OS, these data were still quite immature, with <4% of events occurring in each arm.⁷

Sara A. Hurvitz, MD:
The safety outcomes in this final iDFS analysis were similar to those reported in the interim analysis.^7,9 In terms of QT interval prolongation—an established part of the safety profile for ribociclib—it was reassuring to see that QT interval prolongation occurred at any grade in only 5.3% of patients and at grade ≥3 in 1.0% of patients who received adjuvant ribociclib. Liver-related adverse events (AEs) of any grade occurred in 26.4% of patients with ribociclib, with grade ≥3 liver-related AEs in 8.6%.

Discontinuations due to AEs occurred in 19.5% of patients who received ribociclib in the interim analysis. This rate increased by only 0.8% in the final iDFS analysis, with liver-related AEs being the most common reason.

Joyce O’Shaughnessy, MD:
Continuing our discussion of adjuvant CDK4/6 inhibition for EBC, we will next review 2 studies performed using data from the monarchE trial. In brief, monarchE is an ongoing phase III trial comparing the addition of abemaciclib for 2 years to adjuvant endocrine therapy vs endocrine therapy alone in 5637 patients with high-risk, node-positive hormone receptor–positive/HER2-negative EBC.¹⁰ This trial led to the FDA approval and adoption of adjuvant abemaciclib as the current standard of care in this setting.^11,12 More recently, we saw 5-year outcomes presented at ESMO 2023 and subsequently published in 2024.^13,14 These data showed that adding abemaciclib to adjuvant endocrine therapy has continued to benefit patients in this setting, with iDFS curves continuing to separate and an absolute benefit of 7.6% at 5 years.

At SABCS 2023, Turner and colleagues¹⁵ presented a molecular profiling study of primary tumor tissue, and Loi and colleagues¹⁶ presented results from a pilot study on circulating tumor DNA (ctDNA) detection in monarchE participants. We will first discuss the results of the molecular profiling study.

Joyce O’Shaughnessy, MD:
This analysis evaluated the treatment effect of abemaciclib in a biomarker population of 1400 monarchE participants consisting of, first, a stratified, random sample of the ITT population and, second, all patients with iDFS events at the primary analysis cutoff.¹⁵ Baseline primary tumors from this population were then analyzed using whole-exome sequencing (WES) and exome-capture RNA sequencing (RNAseq). The investigators then examined the association between iDFS and various biomarkers in participants who had received abemaciclib plus endocrine therapy or endocrine therapy alone.

We should note that the iDFS benefit with the addition of abemaciclib was consistent across the ITT population and the subpopulations analyzed using WES and RNAseq.

Joyce O’Shaughnessy, MD:
The WES data were used to assign each tumor to an intrinsic molecular subtype—luminal A, luminal B, HER2 enriched, or basal like.¹⁵ Regardless of the intrinsic molecular subtype, there was a similar reduction in the risk of recurrence with the addition of abemaciclib to adjuvant endocrine therapy, as you can see here. Although the number of patients in the HER2-enriched and basal-like subsets were small, the HR values still favor abemaciclib.

There had been some concern that the basal-like subtype might not benefit from adjuvant abemaciclib based on data suggesting limited benefit with CDK4/6 inhibition in the metastatic setting among patients with hormone receptor–positive basal-like cancer.¹⁷ Fortunately, we did not see that phenomenon in this analysis of EBC.

Joyce O’Shaughnessy, MD:
Using the RNAseq data, the investigators were able to infer the 21-gene Oncotype recurrence score.¹⁵ Based on this inferred Oncotype risk score, 71% of patients were categorized as having a high-risk recurrence score. Abemaciclib was associated with a comparable reduction in the risk of recurrence in patients categorized as high risk and low risk.

Joyce O’Shaughnessy, MD:
The investigators also explored the relationship between iDFS benefit with abemaciclib and oncogenic alterations that had a prevalence ≥9%, such as mutations in PIK3CA or TP53 and amplification of CCND1.¹⁵

Of these alterations, only MYC amplification was associated with decreased benefit from the addition of abemaciclib. This is consistent with preclinical data suggesting that c-Myc upregulation mediates CDK4/6 resistance.¹⁸ However, I would consider this finding to be hypothesis generating only and requiring further exploration and confirmation in other studies.

Joyce O’Shaughnessy, MD:
Turning now to the second biomarker analysis of monarchE, Loi and colleagues¹⁶ presented results from a pilot study on the technical feasibility of ctDNA detection and dynamics to predict recurrence in a subset of monarchE participants.

The investigators preselected 178 patients with a range of tumor mutation burden based on existing WES data from primary tumors. This subset was limited to those who had received adjuvant chemotherapy and initiated endocrine therapy before randomization and excluded patients who experienced an iDFS event within the 2-year abemaciclib treatment period.

Archival primary breast cancer tissue was collected at study entry, and plasma samples were collected at Months 0 and 24—bracketing the 2-year abemaciclib treatment period. To detect ctDNA, the investigators used the Signatera ctDNA assay, which was individualized to each patient using baseline WES data.

Joyce O’Shaughnessy, MD:
After completing 2 years of adjuvant abemaciclib, 42 (23.6%) patients were ctDNA positive, and all patients experienced recurrence.¹⁶ Thus, the positive predictive value for being ctDNA positive at Month 24 was 100%.

However, 136 (76%) patients were ctDNA negative after completing 2 years of abemaciclib and, of these, 28 (21%) experienced recurrence. This yielded a negative predictive value of ctDNA negativity predicting for lack of recurrence at Month 24 of 79%.

By contrast, the positive and negative predictive values for recurrence were both lower using ctDNA status at Month 0. Many patients who were ctDNA positive at baseline became ctDNA negative over the course of 2 years of endocrine therapy with or without abemaciclib.

Sara A. Hurvitz, MD:
We will next discuss a trial assessing the potential benefit of “priming” with a lead-in ICI before adding neoadjuvant chemotherapy in the setting of early TNBC. The noncomparative phase II “Neo-N” (BCT1902/IBCSG 61-20 Neo-N) trial enrolled 110 patients with stage I-IIA/B TNBC and randomized them to 2 parallel treatment arms.¹⁹ The “lead-in” arm received nivolumab alone for 2 weeks, followed by nivolumab plus carboplatin/paclitaxel for 4 cycles. The “concurrent” arm started with nivolumab plus carboplatin/paclitaxel for 4 cycles followed by nivolumab alone. Of note, these were anthracycline-free regimens.

The primary endpoint was pCR, defined as ypT0/is ypN0.

Sara A. Hurvitz, MD:
A similar number of patients were randomized to each noncomparative treatment arm.¹⁹ Overall, two thirds had stage II/III disease. One third had high levels of tumor-infiltrating lymphocytes, defined as ≥30%, and approximately one half were considered PD-L1 positive, defined as ≥1% per the SP142 assay. The median Ki-67 expression was 70%. Similar data have been reported in other studies of the TNBC population.²⁰

Sara A. Hurvitz, MD:
The pCR rates were fairly similar across arms, with pCR being observed in 51% (90% CI: 39%-63%) of patients in the lead-in arm and 55% (90% CI: 43%-66%) of patients in the concurrent arm.¹⁹

Of note, a high level of tumor-infiltrating lymphocytes was the sole predictor of pCR (67% vs 46%; odds ratio: 2.47) in a multivariate logistic regression analysis.

Sara A. Hurvitz, MD:
Continuing with the theme of ICI-based regimens in early TNBC, the open-label phase III ALEXANDRA/IMpassion030 trial is comparing adjuvant chemotherapy with or without atezolizumab followed by atezolizumab maintenance vs monitoring only in patients with early TNBC.²² This study addresses the important question of whether it is necessary to start ICI in the neoadjuvant phase, when an the invasive cancer is present to train the immune system, or if the ICI can be given only in the adjuvant phase.

Earlier, the phase III IMpassion031 trial demonstrated that adding atezolizumab to neoadjuvant chemotherapy in patients with early TNBC significantly improved the pCR rate vs chemotherapy alone (58% vs 41%, respectively; 1-sided P = .0044).²³

ALEXANDRA/IMpassion030 randomized 2199 patients with resected stage II-III TNBC to receive induction with atezolizumab plus chemotherapy vs chemotherapy alone, followed by maintenance atezolizumab to complete 1 year, vs monitoring visits only.²² Of note, the chemotherapy included a taxane, anthracycline, and cyclophosphamide, but no platinum.

At SABCS 2023, Ignatiadis and colleagues²² presented an interim analysis of the primary endpoint, iDFS in the ITT population.

Sara A. Hurvitz, MD:
I was very interested to see that there was no iDFS benefit from adding atezolizumab to adjuvant chemotherapy in this setting.²² In the ITT population, the proportion of patients experiencing an iDFS event was essentially the same for each arm (11.5% with atezolizumab plus chemotherapy vs 10.2% with chemotherapy alone), yielding an HR of 1.12 (95% CI: 0.87-1.45; P = .37). Thus, futility was declared for the primary endpoint of this trial; we should note that the study design was changed midway to add a futility analysis.

No benefit in iDFS was observed among subgroups, and there was no OS benefit in the ITT population.

Sara A. Hurvitz, MD:
No new safety signals were observed in this trial, and atezolizumab did not compromise the administration of chemotherapy.²²

Joyce O’Shaughnessy, MD:
We will conclude our discussion of studies in EBC with a very important, long-awaited trial on locoregional disease control. The phase III NRG Oncology/NSABP B-51/RTOG 1304 trial enrolled 1641 patients with biopsy-proven axillary node–positive disease across hormone receptor–positive/HER2-negative, HER2-positive, and triple-negative breast cancer subtypes (clinical T1-3, N1, M0 disease).²⁴ Patients had received standard neoadjuvant therapy and, at definitive surgery, were found to have become axillary node negative (ypN0) on sentinel lymph node biopsy (≥2 nodes excised), axillary lymph node dissection, or both. Patients could have undergone mastectomy or lumpectomy.

Patients who underwent lumpectomy received whole-breast radiation therapy with or without regional nodal irradiation (RNI). Patients who underwent mastectomy were randomized to receive chest wall radiation with RNI vs no radiation.

The primary endpoint was invasive breast cancer recurrence-free interval (IBCRFI), defined as the time from randomization to invasive local, regional, or distant recurrence, or death from breast cancer. Secondary endpoints include OS and safety.

Joyce O’Shaughnessy, MD:
The bottom line was that there was no significant difference in IBCRFI between treatment arms.²⁴ The 5-year estimated IBCRFI rate was 91.8% among patients who received no RNI vs 92.7% among those who received RNI, yielding a nonsignificant HR of 0.88 (95% CI: 0.60-1.29; P = .51).

With the caveat that the 5 years of follow-up is early in the natural history of EBC, there seemed to be no difference in OS or distant recurrences. There was a very minor numerical difference in the 5-year estimated rate of isolated locoregional recurrences, which was 98.4% without RNI vs 99.3% with RNI.

Joyce O’Shaughnessy, MD:
As expected, there was less radiation-related toxicity among the patients who received no RNI.²⁴