Oncology · Medical Literature
Toxicity profile and tolerability of immune checkpoint inhibitors for the treatment of early-stage breast cancer: a systematic review and meta-analysis.
Abstract
Immune checkpoint inhibitors (ICIs) targeting the PD-1/PD-L1 axis have demonstrated clinical benefit in early-stage breast cancer (eBC), enhancing antitumor immune responses. However, their incorporation into perioperative polychemotherapy has introduced new concerns regarding safety and immune-related toxicities (irAEs). We conducted a systematic review and meta-analysis to comprehensively evaluate the safety profile and tolerability of ICIs in patients with eBC. A systematic review was conducted in PubMed, Embase, Cochrane Library, and relevant conference databases (ASCO, ESMO) to identify clinical trials reporting safety outcomes of ICIs in eBC. Data were analyzed using R (v4.2.2), applying random-effects models for both single-arm and comparative (pairwise) analyses. Thirteen studies were included, comprising 3977 patients across all studies, including 3138 treated with ICI-containing regimens and 839 controls. For comparative pairwise analyses, 2715 patients received ICI-based therapy and 2399 received chemotherapy alone. The most frequently reported all-grade AEs in ICI arms were alopecia (AR 86.49%, 95% CI: 59.25-96.57) and leukopenia (AR 52.09%, 95% CI: 15.14-86.89). For grade ≥3 events, decreased neutrophil count (AR 17.88%, 95% CI: 10.65-28.44) and neutropenia (AR 18.19%, 95% CI: 10.81-28.96) were most common. The most common irAE was hypothyroidism (AR 12.70%, 95% CI: 9.27-17.16), while the most frequent severe irAE (grade ≥3) was adrenal insufficiency (AR 1.47%, 95% CI: 0.68-3.12). This meta-analysis provides a comprehensive evaluation of ICI safety in eBC. While most AEs were low-grade and manageable, significant risks of endocrine irAEs-particularly adrenal insufficiency and thyroid dysfunction-were identified. These findings underscore the need for proactive monitoring and highlight the importance of future research to better understand the mechanisms of ICI-related toxicity and to develop strategies for its prevention.
Clinical relevance
In patients with early-stage breast cancer (eBC) receiving perioperative ICI-containing polychemotherapy regimens, what is the comprehensive safety and tolerability profile—including all-grade and high-grade adverse events, immune-related toxicities (irAEs), and treatment discontinuation rates—compared with chemotherapy alone? __SNAPSHOT__ {"studyDesign":"Systematic review and meta-analysis (single-arm pooled analyses and pairwise comparative analyses; random-effects models)","setting":"Multi-database search (PubMed, Embase, Cochrane Library, ASCO and ESMO conference proceedings); clinical trial data only; specific countries or institution types not reported","sampleSize":"n=3,977 total across all studies (ICI-containing arms: n=3,138; control arms: n=839). For pairwise comparative analyses specifically: ICI-based n=2,715; chemotherapy alone n=2,399","population":"Patients with early-stage breast cancer (eBC) enrolled in clinical trials evaluating ICI-containing perioperative (neoadjuvant and/or adjuvant) regimens; specific breast cancer subtypes, stage distribution, PD-L1 status, and age range not fully reported in the abstract","intervention":"Immune checkpoint inhibitors (ICIs) targeting the PD-1/PD-L1 axis, administered in combination with perioperative polychemotherapy","comparator":"Chemotherapy alone (perioperative; used in pairwise comparative analyses)","primaryOutcome":"Comprehensive safety and tolerability profile of ICIs in eBC, including all-grade and high-grade adverse events (AEs) and immune-related adverse events (irAEs); specific a priori primary endpoint designation not reported in the abstract","followUp":"Not reported","fundingSource":"Not reported"}
Key findings
Thirteen studies comprising 3,977 patients were included, with 3,138 treated with ICI-containing regimens and 839 controls in all-study analyses; pairwise comparative analyses included 2,715 ICI-treated and 2,399 chemotherapy-alone patients. In single-arm pooled analyses of ICI arms, the most frequently reported all-grade adverse events were alopecia (absolute rate [AR] 86.49%, 95% CI: 59.25–96.57) and leukopenia (AR 52.09%, 95% CI: 15.14–86.89). Results for grade ≥3 AEs, immune-related adverse events (irAEs) by category, treatment discontinuation rates, and pairwise comparative risk ratios or odds ratios versus chemotherapy alone were not fully reported in the available abstract. P-values, NNT, and NNH are not calculable from the data provided. Relative effect sizes (RR, OR, or HR) for individual AEs from comparative analyses, and data on serious or fatal irAEs, are not reported in the abstract.
Practice implications
This systematic review and meta-analysis provides a consolidated safety and tolerability dataset for PD-1/PD-L1 inhibitors used perioperatively in early-stage breast cancer, a clinically important but heterogeneous population receiving potentially curative treatment. The reported all-grade alopecia rate (AR ~86%) is largely attributable to concomitant cytotoxic chemotherapy rather than ICI activity per se, and its very wide 95% CI (59–97%) reflects substantial between-study heterogeneity. The leukopenia rate (AR ~52%) is similarly confounded by the chemotherapy backbone. These figures do not allow isolation of ICI-specific toxicity contributions without full comparative data from the pairwise analyses. Because only a truncated abstract was available, the full spectrum of irAEs (e.g., thyroid dysfunction, hepatitis, pneumonitis, adrenal insufficiency), high-grade (grade ≥3) AE rates, ICI discontinuation rates, and comparative risk estimates relative to chemotherapy alone cannot be evaluated here, limiting definitive clinical interpretation. The analysis applies to patients enrolled in clinical trials, who are typically younger, have fewer comorbidities, and are more closely monitored than typical real-world patients—potentially underestimating real-world irAE burden. Findings appear most directly generalizable to high-risk early TNBC and, to a lesser extent, HR+/HER2− breast cancer, as these subtypes have been the focus of key eBC ICI trials (e.g., KEYNOTE-522, KEYNOTE-756). Whether results support practice change, particularly for non-TNBC subtypes, cannot be determined from the abstract alone. Note: this summary is based only on the abstract; full-text review is strongly recommended before applying these findings clinically. __LIMITATIONS__ Only the truncated abstract was available; grade ≥3 AE rates, irAE-specific rates, pairwise comparative effect estimates, and discontinuation data are not reported here and cannot be assessed.,Substantial between-study heterogeneity is implied by very wide confidence intervals (e.g., alopecia 95% CI: 59–97%; leukopenia 95% CI: 15–87%), which may limit reliability of pooled estimates.,Thirteen included studies represent a small evidence base; individual study sample sizes, design quality, and risk of bias are not described in the abstract.,Inclusion of both single-arm and comparative study designs in the same analysis introduces methodological heterogeneity.,All data derive from clinical trials with selected patient populations; irAE rates in real-world practice are consistently higher than those observed in pivotal trials (e.g., real-world pembrolizumab irAE rates of 61–71% versus ~39% in KEYNOTE-522).,The chemotherapy backbone varied across included studies, making it difficult to attribute AEs specifically to ICI versus chemotherapy.,Lack of long-term follow-up data means delayed or chronic irAEs (e.g., adrenal insufficiency, late-onset thyroiditis occurring months to years after treatment) are likely underrepresented.,PD-L1 status, breast cancer subtype (TNBC vs. HR+/HER2−), and treatment phase (neoadjuvant vs. adjuvant) subgroup data are not reported in the abstract, limiting granularity for subgroup-specific clinical decisions.,Funding source is not reported, precluding assessment of potential industry sponsorship bias.,Absence of patient-reported outcome (PRO) or quality-of-life data means the subjective tolerability burden cannot be assessed. __BOTTOM_LINE_CATEGORY__ potentially_practice_informing __BOTTOM_LINE__ This systematic review and meta-analysis synthesizes safety data from 13 clinical trials (n=3,977) of perioperative PD-1/PD-L1 inhibitors in early-stage breast cancer, finding high all-grade AE rates (alopecia ~86%, leukopenia ~52%) that predominantly reflect the cytotoxic chemotherapy backbone rather than ICI-specific toxicity. Full irAE-specific, grade ≥3, and comparative risk data are not available in the abstract, precluding definitive conclusions about the incremental toxicity burden of ICIs over chemotherapy alone. Clinicians should obtain and review the full publication before drawing practice-level conclusions, remain vigilant for immune-mediated toxicities in this curative-intent population, and recognize that real-world irAE rates may substantially exceed those reported in clinical trial settings.