KEYNOTE-522: Pembrolizumab Plus Chemotherapy in Early Triple-Negative Breast Cancer

Co-primary 1 (pCR [ypT0/Tis ypN0], first 602 patients ITT): 64.8% vs 51.2% — difference 13.6 percentage points (95% CI, 5.4 to 21.8; P<0.001) Co-primary 2 (EFS, all 1174 patients ITT): 91.3% vs 85.3% at 18 months — HR 0.63 (95% CI, 0.43 to 0.93; p=not reported) Key secondary (pCR [ypT0 ypN0]): 59.9% vs 45.3% — difference 14.5 percentage points (95% CI, 6.2 to 22.7) Safety signal: Immune-related adverse events of interest occurred in 38.9% of pembrolizumab-treated patients (grade ≥3: 12.9%) vs 18.3% with placebo-chemotherapy (grade ≥3: 1.8%)

📱 Clinic quick reference: Mobile version — save for point-of-care use

Clinical Bottom Line

KEYNOTE-522 demonstrated that adding pembrolizumab to neoadjuvant carboplatin-taxane followed by anthracycline-cyclophosphamide chemotherapy, with continued adjuvant pembrolizumab, significantly improved pathological complete response rates in patients with previously untreated, early-stage triple-negative breast cancer. The benefit in pathological complete response was observed regardless of PD-L1 expression status, and an early signal for improved event-free survival was also observed, though the EFS data were immature at this interim analysis and had not crossed the prespecified statistical boundary.

This trial fundamentally changed the treatment paradigm for early triple-negative breast cancer. Before KEYNOTE-522, the standard neoadjuvant approach was chemotherapy alone — typically an anthracycline-taxane backbone with or without carboplatin. The addition of perioperative pembrolizumab established immunotherapy as a standard component of curative-intent treatment for stage II–III TNBC. As of 2026, pembrolizumab plus chemotherapy followed by adjuvant pembrolizumab is a category 1 NCCN-recommended regimen for this population, and subsequent publications with longer follow-up confirmed the EFS benefit.

The key practical concern is the substantially higher rate of immune-mediated adverse events — particularly adrenal insufficiency, thyroid dysfunction, and severe skin reactions — on top of an already intensive chemotherapy backbone. Treatment-related discontinuations were nearly double with pembrolizumab, and clinicians must be prepared to manage endocrine and dermatologic toxicities that may require long-term hormone replacement or treatment interruption.

Trial Overview

Study Design

• Trial name and registration: KEYNOTE-522 (NCT03036488)
• Design: Phase 3, randomized, double-blind, placebo-controlled; prespecified interim analysis
• Randomization: 2:1 (pembrolizumab–chemotherapy : placebo–chemotherapy), stratified by nodal status (positive or negative), tumor size (T1 to T2 or T3 to T4), and schedule of carboplatin administration (once weekly or every 3 weeks)
• Setting: Previously untreated, nonmetastatic, stage II or stage III triple-negative breast cancer
• Enrollment period and sites: 1174 patients from 181 sites (plus 2 satellite sites) in 21 countries

Mechanism of Action

Pembrolizumab is a humanized monoclonal antibody that binds to the programmed death-1 (PD-1) receptor on T cells, blocking its interaction with PD-L1 and PD-L2 ligands expressed on tumor cells and tumor-infiltrating immune cells. This blockade releases PD-1 pathway–mediated inhibition of the immune response, including the antitumor immune response [2].

Patient Population

• Key eligibility: At least 18 years of age; centrally confirmed triple-negative breast cancer in all foci; newly diagnosed, previously untreated, nonmetastatic disease (T1c N1-2 or T2-4 N0-2, AJCC 7th edition); ECOG performance-status score of 0 or 1; adequate organ function
• Biomarker selection: Patients were eligible regardless of PD-L1 status. PD-L1 expression was assessed centrally using the PD-L1 IHC 22C3 pharmDx assay; combined positive score (CPS) ≥1 was considered PD-L1–positive.
• Sample size flow: 1174 randomized (784 pembrolizumab–chemotherapy, 390 placebo–chemotherapy) → safety population 1170 (781 pembrolizumab–chemotherapy, 389 placebo–chemotherapy). Number screened was not reported in this publication. The pCR primary analysis included the first 602 patients randomized (401 pembrolizumab–chemotherapy, 201 placebo–chemotherapy).

Baseline Characteristics

Treatment Protocol

Experimental Arm: Pembrolizumab–Chemotherapy (n=784 randomized; safety population n=781)

Neoadjuvant: 4 cycles pembrolizumab (200 mg IV Q3W) + paclitaxel + carboplatin, followed by 4 cycles pembrolizumab + doxorubicin or epirubicin + cyclophosphamide. Adjuvant: pembrolizumab Q3W for up to 9 cycles.

• Dose and schedule: Pembrolizumab 200 mg IV Q3W; paclitaxel 80 mg/m² weekly; carboplatin AUC 5 mg/mL/min Q3W or AUC 1.5 mg/mL/min weekly (first 12 weeks); then pembrolizumab + doxorubicin 60 mg/m² or epirubicin 90 mg/m² + cyclophosphamide 600 mg/m² Q3W (subsequent 12 weeks); adjuvant pembrolizumab Q3W up to 9 cycles
• Treatment duration: Not reported in this publication
• Median treatment duration: 51.1 weeks (range, 0.1 to 88.4)
• Median relative dose intensity: Not reported in this publication

Control Arm: Placebo–Chemotherapy (n=390 randomized; safety population n=389)

Neoadjuvant: 4 cycles placebo Q3W + paclitaxel + carboplatin, followed by 4 cycles placebo + doxorubicin or epirubicin + cyclophosphamide. Adjuvant: placebo Q3W for up to 9 cycles.

• Dose and schedule: Placebo IV Q3W; paclitaxel 80 mg/m² weekly; carboplatin AUC 5 mg/mL/min Q3W or AUC 1.5 mg/mL/min weekly (first 12 weeks); then placebo + doxorubicin 60 mg/m² or epirubicin 90 mg/m² + cyclophosphamide 600 mg/m² Q3W (subsequent 12 weeks); adjuvant placebo Q3W up to 9 cycles
• Treatment duration: Not reported in this publication
• Median treatment duration: 54.1 weeks (range, 0.1 to 79.3)

No crossover was permitted between the treatment phases.

Efficacy Outcomes

Co-Primary Endpoint 1: Pathological Complete Response (ypT0/Tis ypN0)

Definition: Pathological stage ypT0/Tis ypN0 at the time of definitive surgery — no residual invasive cancer in the complete resected breast specimen and all sampled regional lymph nodes.

Analysis population: First 602 patients who underwent randomization (intention-to-treat)

Statistical method: Stratified method of Miettinen and Nurminen, with weights proportional to the stratum size

Alpha allocation: 0.003 one-sided at this interim analysis (prespecified criterion); overall one-sided alpha of 0.005 for pCR

Pembrolizumab–Chemotherapy: 64.8% (260 of 401 patients; 95% CI, 59.9 to 69.5) Placebo–Chemotherapy: 51.2% (103 of 201 patients; 95% CI, 44.1 to 58.3) Comparison: Difference 13.6 percentage points (95% CI, 5.4 to 21.8; P<0.001, one-sided)

Co-Primary Endpoint 2: Event-Free Survival

Definition: Time from randomization to disease progression that precludes definitive surgery, local or distant recurrence, a second primary cancer, or death from any cause, whichever occurred first.

Analysis population: Intention-to-treat (all 1174 patients)

Statistical method: Stratified log-rank test; stratified Cox proportional-hazards model with Efron's method

Median follow-up: 15.5 months (range, 2.7 to 25.0)

Alpha allocation: 0.02 one-sided at the final analysis

Pembrolizumab–Chemotherapy: 18-month EFS 91.3% (95% CI, 88.8 to 93.3); 58 of 784 patients (7.4%) had an event Placebo–Chemotherapy: 18-month EFS 85.3% (95% CI, 80.3 to 89.1); 46 of 390 patients (11.8%) had an event Comparison: HR 0.63 (95% CI, 0.43 to 0.93)

Important note on EFS maturity: With 104 events of 327 expected at the final analysis, the EFS data were immature at this second interim analysis. No p-value was reported for EFS; the prespecified boundary for statistical significance had not been crossed at this interim analysis. The 95% CI for the hazard ratio was not adjusted for multiplicity.

Multiplicity Control

The graphical method of Maurer and Bretz was used to control the type I error rate at a one-sided alpha level of 0.025 across both co-primary endpoints and all interim and final analyses. The Lan–DeMets O'Brien–Fleming spending function was used to control type I error across interim and final analyses. At this prespecified interim analysis, pCR met its prespecified statistical criterion (one-sided alpha boundary of 0.003). The EFS co-primary endpoint had not crossed its prespecified boundary at this interim.

Key Secondary Endpoints

Pathological Complete Response (ypT0 ypN0)

Definition: No residual invasive and in situ cancer in the complete resected breast specimen and all sampled regional lymph nodes.

Analysis population: First 602 patients who underwent randomization.

Formally tested: Not reported in this publication.

Pembrolizumab–Chemotherapy: 59.9% (240 of 401 patients; 95% CI, 54.9 to 64.7) Placebo–Chemotherapy: 45.3% (91 of 201 patients; 95% CI, 38.3 to 52.4) Difference: 14.5 percentage points (95% CI, 6.2 to 22.7)

Pathological Complete Response (ypT0/Tis)

Definition: No invasive cancer in the breast, irrespective of ductal carcinoma in situ or nodal involvement.

Analysis population: First 602 patients who underwent randomization.

Formally tested: Not reported in this publication.

Pembrolizumab–Chemotherapy: 68.6% (275 of 401 patients; 95% CI, 63.8 to 73.1) Placebo–Chemotherapy: 53.7% (108 of 201 patients; 95% CI, 46.6 to 60.8) Difference: 14.8 percentage points (95% CI, 6.8 to 23.0)

Efficacy by PD-L1 Status

PD-L1–Positive Population (CPS ≥1)

Analysis population: PD-L1–positive patients among first 602 randomized.

Pembrolizumab–Chemotherapy: 68.9% (230 of 334 patients) Placebo–Chemotherapy: 54.9% (90 of 164 patients) Difference: 14.2 percentage points (95% CI, 5.3 to 23.1)

PD-L1–Negative Population (CPS <1)

Analysis population: PD-L1–negative patients among first 602 randomized.

Pembrolizumab–Chemotherapy: 45.3% (29 of 64 patients) Placebo–Chemotherapy: 30.3% (10 of 33 patients) Difference: 18.3 percentage points (95% CI, –3.3 to 36.8)

The pCR benefit of pembrolizumab was observed in both PD-L1–positive and PD-L1–negative subpopulations, though the difference in the PD-L1–negative subgroup did not reach statistical significance (95% CI crossed zero). Confidence intervals were not reported for the individual arm rates in the PD-L1 subpopulations.

Safety

Safety Population

Safety was assessed in the as-treated population during the neoadjuvant phase: 781 patients in the pembrolizumab–chemotherapy group and 389 patients in the placebo–chemotherapy group.

Safety Summary

Across all treatment phases (neoadjuvant and adjuvant combined), treatment-related adverse events of grade ≥3 occurred in 78.0% of the pembrolizumab–chemotherapy group and 73.0% of the placebo–chemotherapy group. In the adjuvant phase, treatment-related AEs occurred in 48.1% of 547 patients in the pembrolizumab–chemotherapy group and 43.0% of 314 patients in the placebo–chemotherapy group.

The most common serious treatment-related adverse events in the neoadjuvant phase included febrile neutropenia (14.6% vs 12.1%), anemia (2.6% vs 2.1%), and pyrexia (2.6% vs 0.3%).

Grade ≥3 Adverse Events of Clinical Significance

Adverse Events of Special Interest

Adverse events of interest (immune-related and infusion reactions) occurred in 304 patients (38.9%) in the pembrolizumab–chemotherapy group and 71 patients (18.3%) in the placebo–chemotherapy group; grade ≥3 events occurred in 101 (12.9%) and 7 (1.8%), respectively.

⚠️ Immune-Related Adverse Events of Interest: Critical Safety Signal

• Any grade: 304 (38.9%) vs 71 (18.3%)
• Grade ≥3: 101 (12.9%) vs 7 (1.8%)
• Clinical implication: The substantially higher rate of immune-mediated events with pembrolizumab requires proactive monitoring and multidisciplinary management. The most clinically significant immune-related events are detailed below.

Specific immune-related adverse events:

⚠️ Adrenal Insufficiency: Critical Safety Signal

• Any grade: 18 (2.3%) vs 0
• Grade ≥3: 10 (1.3%) vs 0
• Clinical implication: Adrenal insufficiency occurred exclusively in the pembrolizumab arm and frequently required systemic corticosteroid replacement. Clinicians should have a low threshold for cortisol testing in patients presenting with fatigue, hypotension, or electrolyte abnormalities during or after pembrolizumab treatment. Many patients will require lifelong hormonal supplementation.

⚠️ Severe Skin Reaction: Critical Safety Signal

• Any grade: 34 (4.4%) vs 4 (1.0%)
• Grade ≥3: 30 (3.8%) vs 1 (0.3%)
• Clinical implication: The majority of severe skin reactions with pembrolizumab were grade ≥3, indicating that most cases that developed were clinically significant. Early dermatology consultation and prompt treatment interruption should be considered.

Deaths

• Treatment-related deaths: 3 patients (0.4%) in the pembrolizumab–chemotherapy group and 1 patient (0.3%) in the placebo–chemotherapy group. The cause of death in the placebo–chemotherapy group was septic shock. Causes of the three treatment-related deaths in the pembrolizumab arm were not specified in the main text.
• Total deaths by arm: Not reported in this publication.

Subgroup Analyses

Subgroup analyses were performed for the co-primary endpoint of pCR (ypT0/Tis ypN0) among the first 602 randomized patients.

pp = percentage points

The treatment effect on pCR favored pembrolizumab–chemotherapy across the majority of prespecified subgroups. Notable observations include a larger treatment difference in node-positive patients compared with node-negative patients, and in patients receiving weekly carboplatin compared with every-3-week carboplatin. The ECOG 1 subgroup showed a numerically negative difference, though this subgroup was very small (73 vs 28 patients) and should not be overinterpreted. No interaction p-values were reported.

These subgroup analyses were not powered for formal statistical comparisons. Results should be interpreted as hypothesis-generating.

Key Comparator Trials

Cross-trial comparisons are limited by differences in patient populations, trial designs, and endpoints. These data are presented for context, not for direct statistical comparison.

Contextual Notes

KEYNOTE-522 was the largest and most definitive trial testing neoadjuvant checkpoint inhibition in early TNBC. The IMpassion031 trial [4] evaluated atezolizumab (anti–PD-L1) added to a nab-paclitaxel → anthracycline-cyclophosphamide backbone, without carboplatin, and reported a numerically larger treatment difference in pCR but in a smaller population (333 patients). The GeparNuevo trial [5] evaluated durvalumab in a window design with a smaller sample size (174 patients) and did not reach statistical significance for the primary pCR endpoint. Critically, KEYNOTE-522 is the only trial in this space that incorporated both neoadjuvant and adjuvant immunotherapy phases and demonstrated a statistically significant co-primary endpoint of event-free survival (confirmed in the subsequent final analysis publication). The inclusion of carboplatin in the KEYNOTE-522 chemotherapy backbone further differentiates it from IMpassion031, as carboplatin itself contributes to higher pCR rates in TNBC and complicates the attribution of the pCR benefit specifically to pembrolizumab.

Grey Zones and Unanswered Questions

• Optimal chemotherapy backbone: Because the control arm included carboplatin + taxane → anthracycline-cyclophosphamide, the trial cannot isolate the benefit of pembrolizumab from the benefit of this specific backbone. Whether pembrolizumab would show a similar magnitude of benefit when added to a non-carboplatin-containing backbone remains uncertain.

• pCR-guided de-escalation of adjuvant pembrolizumab: The trial required adjuvant pembrolizumab regardless of pCR status. Whether patients who achieve pCR can safely omit adjuvant pembrolizumab — thereby avoiding additional immune-mediated toxicity and treatment burden — is a critical unanswered question. Similarly, whether patients who do not achieve pCR should receive alternative adjuvant strategies (e.g., capecitabine, olaparib in BRCA-mutated disease) instead of or in addition to pembrolizumab is not addressed.

• BRCA-mutated TNBC: Patients with germline BRCA1/2 mutations were not excluded but were not analyzed as a prespecified subgroup in this publication. Whether this population, which may have inherently higher chemosensitivity and pCR rates, derives the same magnitude of benefit from pembrolizumab is unclear.

• T1cN0 population and the lower boundary of eligibility: The trial enrolled patients with T1c N1-2 or T2-4 N0-2 disease, representing moderate to high-risk early TNBC. Whether patients with small, node-negative tumors (e.g., T1c N0) benefit from the addition of immunotherapy to neoadjuvant chemotherapy — where prognosis is already relatively favorable — is unknown and clinically relevant, as this group represents a meaningful proportion of newly diagnosed TNBC patients.

• Long-term immune-mediated toxicity: With a median follow-up of only 15.5 months, the long-term consequences of immune-mediated endocrinopathies (particularly adrenal insufficiency and thyroid dysfunction requiring permanent replacement) and the late-onset immune events that can emerge months to years after completion of immunotherapy are not captured in this interim analysis.

Clinical Implications

Rationale and Clinical Context

Before KEYNOTE-522, the treatment of early-stage TNBC relied on cytotoxic chemotherapy as the sole systemic modality. Neoadjuvant chemotherapy with anthracycline-taxane-based regimens achieved pCR rates of approximately 30%–45% depending on the backbone, and residual disease after neoadjuvant therapy was associated with substantially worse outcomes. The biological rationale for adding PD-1 blockade to TNBC treatment rests on the higher tumor mutational burden, greater immune infiltration, and more frequent PD-L1 expression observed in TNBC compared with other breast cancer subtypes. KEYNOTE-522 tested the hypothesis that pembrolizumab added to an intensive neoadjuvant chemotherapy platform could meaningfully increase pCR rates and, critically, improve event-free survival — bridging the gap between pathological response and long-term clinical outcomes.

Monitoring and Long-Term Follow-Up

Given the immune-mediated toxicity profile, patients treated with perioperative pembrolizumab require:

• Thyroid function testing before each cycle during treatment and periodically for at least 12 months after completion, given the 13.7% incidence of hypothyroidism
• Morning cortisol or ACTH stimulation testing at the first sign of fatigue, hypotension, hyperkalemia, or hyponatremia, given the 2.3% incidence of adrenal insufficiency (exclusively in the pembrolizumab arm)
• Hepatic function monitoring throughout treatment, with elevated ALT (grade ≥3: 5.2%) occurring at more than double the rate of the control arm
• Skin surveillance with low threshold for dermatology referral, given the 3.8% grade ≥3 severe skin reaction rate

The median follow-up of 15.5 months in this publication is insufficient to assess long-term EFS and OS outcomes. Subsequent analyses with longer follow-up are essential for confirming the durability of the survival benefit and characterizing late immune-mediated toxicity.

Unanswered Questions

The two most practice-relevant open questions from this trial are (1) whether adjuvant pembrolizumab can be safely omitted in patients who achieve pCR, thereby reducing toxicity and cost without compromising outcomes, and (2) how to sequence or combine pembrolizumab with capecitabine or olaparib in patients with residual disease after neoadjuvant treatment — populations currently addressed by CREATE-X and OlympiA data, respectively, but not in the context of prior pembrolizumab exposure.

Regulatory and Guideline Status

Regulatory

• FDA: Pembrolizumab was granted accelerated approval by the FDA on July 26, 2021, for high-risk, early-stage triple-negative breast cancer in combination with chemotherapy as neoadjuvant treatment and then continued as a single agent as adjuvant treatment after surgery, based on pCR data from KEYNOTE-522. Regular approval was subsequently granted based on the final EFS analysis. Regulatory status should be verified with current FDA/EMA labeling.
• EMA: Approved for the same indication under conditional marketing authorization, subsequently converted to standard authorization. Regulatory status should be verified with current EMA labeling.

⚠️ Regulatory status verified as of March 2026. Confirm current approval status and labeled indications before clinical use at FDA.gov.

Guidelines

• NCCN: Pembrolizumab plus chemotherapy (neoadjuvant) followed by adjuvant pembrolizumab is a Category 1 recommended regimen for patients with T1c N1-2 or T2-4 N0-2 triple-negative breast cancer in the NCCN Clinical Practice Guidelines in Oncology: Breast Cancer [3].

Companion Diagnostics

The KEYNOTE-522 trial enrolled patients regardless of PD-L1 status, and the FDA-approved indication does not require PD-L1 testing for patient selection. PD-L1 expression was assessed centrally using the PD-L1 IHC 22C3 pharmDx assay for exploratory subgroup analyses but is not a companion diagnostic requirement for this indication [2].

About the Author

Andrew Stevenson is the founder and systems architect of kill-cancer.com, a clinical intelligence platform delivering structured, source-traced oncology trial analysis to practicing clinicians. He holds 17 Google technical certifications in data systems, automation, and applied AI — the engineering foundation behind the extraction and verification pipeline that produces every article on this platform. Every number traces directly to its source publication. Zero calculation. Zero editorializing. Zero hallucination.

Five siblings lost to cancer built the urgency. The engineering builds the trust.

📧 andrew@kill-cancer.com 🌐 kill-cancer.com 💬 kill-cancer.com/forum

Disclaimer

This article is intended for healthcare professionals only. It is not medical advice and should not be used as a substitute for professional clinical judgment. Treatment decisions should be made in consultation with qualified healthcare providers based on individual patient circumstances.

All trial data presented in this article are sourced directly from the published clinical trial report and its supplementary materials. Numbers are reproduced exactly as reported; no calculations, derivations, or estimates have been performed.

Trial results are presented as reported in the source publication. Updated data, label changes, or guideline revisions published after the source article may alter clinical applicability.

Comparator trial data presented in Section 8 are sourced from their respective published reports and are provided for contextual purposes only. Cross-trial comparisons have inherent limitations and should not be interpreted as direct statistical comparisons.

References

1. Schmid P, Cortes J, Pusztai L, et al. Pembrolizumab for Early Triple-Negative Breast Cancer. N Engl J Med. 2020;382(9):810-821. doi:10.1056/NEJMoa1910549
2. Keytruda (pembrolizumab) prescribing information. U.S. Food and Drug Administration. Accessed March 2026.
3. NCCN Clinical Practice Guidelines in Oncology: Breast Cancer. Version 1.2026. Accessed March 2026.
4. Mittendorf EA, Zhang H, Barrios CH, et al. Neoadjuvant atezolizumab in combination with sequential nab-paclitaxel and anthracycline-based chemotherapy versus placebo and chemotherapy in patients with early-stage triple-negative breast cancer (IMpassion031): a randomised, double-blind, phase 3 trial. Lancet. 2020;396(10257):1090-1100. doi:10.1016/S0140-6736(20)31953-X
5. Loibl S, Untch M, Burchardi N, et al. A randomised phase II study investigating durvalumab in addition to an anthracycline taxane-based neoadjuvant therapy in early triple-negative breast cancer: clinical results and biomarker analysis of GeparNuevo study. Ann Oncol. 2019;30(8):1279-1288. doi:10.1093/annonc/mdz158