Pathological Complete Response to Nab-Paclitaxel–Containing Neoadjuvant Regimens in Early Triple-Negative Breast Cancer: A Systematic Review and Meta-Analysis Rasha Aziz Attia Salama, Mohamed El-Tanani, Syed Arman Rabbani, Mohamedanas Patni, Bhoomendra Bhongade, Rasha Babiker, Shakta Mani Satyam, Imran Rashid Rangraze, Adil Farooq Wali, Sirajunisa Talath, Omer Eladil Abdalla, Lorenzo Scardina Abstract Background: Triple-negative breast cancer (TNBC) is an aggressive breast cancer subtype with limited targeted treatment options. This systematic review and meta-analysis evaluated pathological complete response (pCR) outcomes associated with nab-paclitaxel-containing neoadjuvant regimens incorporating platinum chemotherapy and/or immune checkpoint inhibitors in early-stage TNBC. Methods: PubMed/MEDLINE, Scopus, Google Scholar, CENTRAL, and Embase were searched for eligible publications within the review window (January 2005 through 31 March 2026). Eligible studies were randomized or prospective clinical studies of non-metastatic TNBC receiving nab-paclitaxel-containing neoadjuvant regimens with platinum and/or immune checkpoint inhibition. Randomized checkpoint-inhibitor comparisons were synthesized as risk ratios; eligible prospective non-randomized pCR datasets were synthesized using a logit random-effects model with REML variance estimation and Hartung-Knapp-Sidik-Jonkman inference. Results: The final evidence base comprised 21 unique eligible prospective studies. Four clinically comparable randomized checkpoint-inhibitor trials (n = 1228) yielded a pooled pCR risk ratio of 1.25 (95% CI 1.12–1.41; I2 = 0%). Fourteen prospective non-randomized datasets (n = 539; 320 pCR events) yielded a pooled pCR proportion of 59.8% (HKSJ 95% CI 53.8–65.6%; I2 = 32.7%; τ2 = 0.0559). Conclusions: Within nab-paclitaxel-containing neoadjuvant backbones, checkpoint-inhibitor addition was associated with higher pCR in randomized comparative evidence. The approximately 60% pooled pCR proportion from prospective non-randomized studies is descriptive and should not be interpreted as a cross-regimen comparison because of clinical and methodological heterogeneity. 1. Introduction Triple-negative breast cancer (TNBC) is a biologically aggressive subtype of breast cancer characterized by the absence of estrogen receptor (ER) expression, progesterone receptor (PR) expression, and human epidermal growth factor 2 (HER2) overexpression or amplification []. Accounting for approximately 15–20% of all breast cancers, TNBC is associated with earlier recurrence, higher metastatic potential, and poorer overall survival compared with other molecular breast cancer subtypes [,]. Unlike hormone receptor-positive or HER2-positive breast cancers, TNBC lacks well-established targeted therapeutic options in many treatment settings, making cytotoxic chemotherapy the principal therapeutic backbone for early-stage disease []. Neoadjuvant chemotherapy has become a standard treatment approach in early-stage and locally advanced TNBC, offering several advantages, including tumor downstaging, improved surgical operability, and the opportunity to assess treatment responsiveness in vivo []. Among the available efficacy endpoints, pathological complete response (pCR) is an important early measure of treatment response and a well-established individual-level prognostic marker in TNBC. Patients who achieve pCR generally experience more favorable event-free and overall survival than those with residual disease []. However, this prognostic association does not establish pCR as a validated trial-level surrogate for treatment effects on event-free or overall survival. Accordingly, differences in pCR rates between treatment strategies should not be interpreted as evidence of a survival benefit without direct time-to-event evidence. Taxane-containing chemotherapy regimens remain central to neoadjuvant TNBC management. Nab-paclitaxel (nanoparticle albumin-bound paclitaxel) represents a solvent-free taxane formulation designed to improve drug delivery while reducing hypersensitivity reactions associated with solvent-based paclitaxel []. Its pharmacologic profile permits steroid-free administration and may enhance intratumoral drug penetration through albumin-mediated transport mechanisms [,]. Several prospective clinical studies have investigated nab-paclitaxel-containing neoadjuvant regimens in TNBC, demonstrating clinically relevant pathological response rates across randomized and uncontrolled settings [,,]. Treatment-intensification strategies in early TNBC have included platinum agents and immune checkpoint inhibitors. Platinum agents have been investigated as treatment-intensification strategies in TNBC, including prospective nab-paclitaxel-containing regimens and broader platinum-based neoadjuvant evidence [,,]. However, within the evidence synthesized in this review, the effect of a platinum component cannot be interpreted independently of the specific chemotherapy backbone and comparator used in each trial. Accordingly, platinum-containing regimens are described according to their actual randomized or prospective treatment contrast rather than as evidence for a stand-alone platinum effect. Researchers have also investigated immune checkpoint blockades in early TNBC in the context of an immunogenic tumor microenvironment characterized by prominent tumor-infiltrating lymphocytes (TILs) [,]. Higher levels of TILs have also been associated with favorable treatment response and prognosis in TNBC [,]. In the randomized evidence considered here, checkpoint inhibitors were administered as additions to multi-agent neoadjuvant chemotherapy; therefore, the review evaluates the effect of adding checkpoint inhibition to a defined chemotherapy backbone and does not infer efficacy of immunotherapy as an isolated treatment component [,,,,,,,,]. Therefore, this systematic review aimed to evaluate the available evidence regarding pathological complete response outcomes associated with nab-paclitaxel-containing neoadjuvant regimens incorporating platinum agents and/or immune checkpoint inhibitors in patients with early-stage triple-negative breast cancer. By focusing specifically on nab-paclitaxel-based strategies, this review seeks to provide a clinically relevant synthesis of treatment intensification approaches in this evolving therapeutic landscape. 2. Methods 2.1. Study Design and Reporting Standards This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [] to evaluate pCR outcomes reported with platinum-containing and/or immune-checkpoint-inhibitor-containing nab-paclitaxel neoadjuvant regimens in patients with early-stage TNBC. The completed PRISMA 2020 checklist is provided as . The review protocol was registered in PROSPERO (Registration Number: CRD420261418350) on 8 June 2026, after completion of the literature search. 2.2. Search Strategy The electronic literature search was designed to maximize sensitivity and incorporated concepts related to triple-negative breast cancer (TNBC), neoadjuvant therapy, nab-paclitaxel, and platinum agents and/or immune checkpoint inhibitors. Pathological complete response (pCR) was not included as a mandatory search concept because eligible studies could report this outcome only within the full text rather than in indexed title, abstract, or keyword fields. Search terminology included alternative terms for nab-paclitaxel and relevant platinum agents, immune checkpoint inhibitor classes, and individual agents. Eligible publications were restricted to the period from January 2005 through 31 March 2026. Five sources were searched: PubMed/MEDLINE, Scopus, CENTRAL, Embase, and Google Scholar. The searches retrieved 103 records from PubMed/MEDLINE, 117 from Scopus, 140 from CENTRAL, 564 from Embase, and 300 from Google Scholar (30 relevance-ranked pages, 10 results per page), yielding a total of 1224 source records. Database-specific search strategies, platforms/providers, execution dates, search yields, publication-cutoff handling, and record-management procedures are provided in . 2.3. Eligibility Criteria Studies were considered eligible for inclusion if they met all of the following criteria: randomized controlled trials or prospective clinical studies involving adult patients with early-stage or locally advanced non-metastatic TNBC (stage I–III); evaluation of neoadjuvant treatment regimens containing nab-paclitaxel; inclusion of platinum-based chemotherapy and/or immune checkpoint inhibitors as part of the treatment strategy; reporting of pCR outcomes; and availability as full-text publications in the English language. Studies were excluded if they were retrospective observational studies, case reports, or case series; review articles, editorials, commentaries, or letters; conference abstracts lacking sufficient extractable data; studies involving metastatic breast cancer populations; studies not reporting TNBC-specific outcomes; studies not involving nab-paclitaxel-containing regimens; or duplicate publications and secondary analyses derived from previously included parent trials. 2.4. Study Selection After removal of 37 records published after the 31 March 2026 cutoff and 326 duplicate source records, 861 unique records remained for title/abstract screening. Of these, 748 were excluded, and 113 reports were sought for retrieval; all 113 reports were successfully retrieved. Accordingly, 113 full-text reports were assessed for eligibility, 92 were excluded with documented reasons, and 21 studies (21 primary reports) were included in the final review (). The final study-family manifest of the 21 included studies is provided in , and the complete report-level exclusion ledger for the 92 excluded reports is provided in . 2.5. Data Extraction Data were extracted independently in duplicate from the primary study reports using a standardized extraction form. For each study included in the quantitative synthesis, the reviewers extracted the pCR event count, denominator, pCR definition, treatment arm, comparator where applicable, and analytic population (e.g., intention-to-treat, randomized efficacy population, or reported subgroup). The two extraction sheets were compared item by item, and discrepancies were resolved by review of the primary report and consensus with a third reviewer. The final agreed study-level values were used for all quantitative tables, figures, and analyses. We did not use secondary reviews as the source of event counts when a primary report was available. 2.6. Outcome Measure The primary outcome was pCR. The study-level pCR definition was extracted from each primary report because definitions were not completely uniform. For the descriptive prospective non-randomized synthesis, quantitative eligibility required: (1) an otherwise eligible prospective TNBC cohort receiving an eligible nab-paclitaxel-containing neoadjuvant regimen; (2) a pure pCR endpoint rather than a composite pathological response endpoint; (3) an extractable pCR numerator and denominator; and (4) a non-overlapping analytic cohort. Separately extractable TNBC subgroups were eligible when both numerator and denominator were reported. Multiple eligible non-overlapping cohorts from the same study were combined at study level where appropriate to avoid treating cohorts from one trial as independent. Randomized comparative trials were analyzed separately using risk ratios and were not incorporated into the descriptive non-randomized proportion meta-analysis. For denominator selection, the denominator used by the primary publication for its reported pCR analysis was retained (e.g., ITT/all enrolled, efficacy-evaluable, surgery-evaluable, or separately reported TNBC subgroup), without reclassifying participants whom the publication excluded from that analysis. Studies using efficacy- or surgery-evaluable denominators were then examined in a dedicated denominator sensitivity analysis. Treatment withdrawal, failure to undergo surgery, or discontinuation was counted as non-pCR only when the primary study itself included those participants in the pCR denominator as nonresponders. KEYNOTE-173 cohorts A-D were combined because they used nab-paclitaxel; cohorts E-F were excluded from quantitative synthesis because they used solvent-based paclitaxel. 2.7. Risk of Bias and Certainty of Evidence Methodological quality was assessed independently by two reviewers according to study design and specifically in relation to the pCR outcome. Randomized controlled trials were evaluated using the Cochrane Risk of Bias 2 (RoB 2) tool for the effect of assignment to intervention []. The five RoB 2 domains—bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome, and bias in selection of the reported result—were evaluated for each randomized comparison, and an overall judgment was assigned according to the RoB 2 framework. Prospective uncontrolled studies were assessed using the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Case Series as a design-appropriate framework for descriptive single-arm evidence []. The appraisal considered clarity of inclusion criteria, reliable ascertainment of the condition, participant recruitment and completeness, reporting of demographic and clinical characteristics, outcome reporting, study setting, and appropriateness of statistical analysis. These assessments were used to characterize methodological limitations of the uncontrolled evidence and were not interpreted as estimates of comparative treatment effect. Certainty of evidence for the principal randomized comparison of checkpoint inhibitor plus chemotherapy versus chemotherapy alone was evaluated using the GRADE framework across the domains of risk of bias, inconsistency, indirectness, imprecision, and publication bias. Certainty was assessed specifically for the pCR outcome. WSG-ADAPT-TN and the randomized early-stage cohort reported by Kuemmel et al. evaluated clinically distinct treatment contrasts and were therefore considered separately rather than incorporated into the principal checkpoint-inhibitor GRADE assessment. Single-arm pooled proportions were retained as descriptive evidence and were not used to infer comparative efficacy. 2.8. Statistical Analysis Comparative randomized studies were grouped according to clinical treatment contrast before synthesis. Four trials evaluating checkpoint-inhibitor addition to a corresponding chemotherapy control (IMpassion031, GeparNuevo, NeoTRIPaPDL1, and CamRelief) were pooled as risk ratios (RRs) with 95% confidence intervals from study-level pCR events and randomized denominators. The primary four-trial checkpoint-inhibitor synthesis used a DerSimonian–Laird random-effects model on the log-risk-ratio scale. WSG-ADAPT-TN and Kuemmel/CO42177 were reported separately because their randomized contrasts addressed chemotherapy/platinum strategy rather than checkpoint-inhibitor addition versus matched non-ICI control. For the descriptive prospective non-randomized synthesis, a study entered the pool when an eligible TNBC cohort receiving a nab-paclitaxel-containing regimen had an extractable pure-pCR numerator and denominator. When a publication reported a defined efficacy/surgery population, the denominator used for the study-reported primary pCR analysis was extracted and examined in sensitivity analysis. Eligible non-overlapping cohorts from the same study were combined at study-family level to avoid double counting within-study. Study proportions were logit-transformed and pooled by inverse-variance random effects using restricted maximum likelihood (REML) for τ2; Hartung-Knapp-Sidik-Jonkman inference was used for the primary 95% confidence interval. Heterogeneity was summarized using Cochran’s Q, I2, and τ2. Sensitivity analyses excluded the very small Huang et al. TNBC subgroup (n = 2; 1/2 pCR), the Chen et al. radiotherapy-containing cohort (13 recruited; 10 modified-ITT efficacy-evaluable; 9/10 pCR), both together, and studies whose reported pCR denominator was an efficacy/surgery-evaluable population rather than essentially the full enrolled/treated population (TREND, Chen et al., and Zheng et al.). The same REML logit random-effects framework with Hartung-Knapp-Sidik-Jonkman inference was used for these sensitivity analyses. A DerSimonian-Laird model was used only as an estimator-sensitivity analysis for the non-randomized synthesis. Complete leave-one-out analysis was performed for the four randomized checkpoint-inhibitor trials. Because only four randomized trials contributed to that comparison, funnel-plot asymmetry tests were not used. The primary comparative meta-analysis was restricted to the clinically coherent checkpoint-inhibitor contrast. The platinum-versus-gemcitabine comparison from WSG-ADAPT-TN was analyzed separately because it addressed a different clinical question. No omnibus meta-analysis combining these distinct randomized treatment contrasts was performed. Because only four randomized trials contributed to the checkpoint-inhibitor meta-analysis, formal statistical assessment of small-study effects or publication bias was not performed. Robustness of the primary checkpoint-inhibitor meta-analysis was evaluated using a complete leave-one-out sensitivity analysis. Each of the four contributing randomized trials was omitted sequentially, and the pooled risk ratio (RR), 95% confidence interval (CI), and I2 statistic were recalculated using the same random-effects model as the primary analysis. Meta-analytic calculations were generated from the extracted study-level event and denominator data. Randomized checkpoint-inhibitor comparisons were pooled on the log-risk-ratio scale using random-effects models, with estimator/inference robustness analyses as specified above. Prospective non-randomized pCR proportions were logit-transformed, pooled by inverse-variance random effects with REML estimation, and back-transformed. Heterogeneity was summarized using Cochran’s Q, I2, and τ2. Forest plots were generated from the final extraction dataset. Complete leave-one-out sensitivity analysis was performed for the randomized checkpoint-inhibitor synthesis, and design/denominator sensitivity analyses were performed for the prospective non-randomized synthesis. 3. Results 3.1. Study Selection The search identified 1224 source records: PubMed/MEDLINE (n = 103), Scopus (n = 117), CENTRAL (n = 140), Embase (n = 564), and Google Scholar (n = 300). Before screening, 37 records published after the stated 31 March 2026 cutoff and 326 duplicate source records were removed, leaving 861 unique records for title/abstract screening. Of these, 748 were excluded. The remaining records yielded 113 reports sought for retrieval; all were retrieved and assessed for eligibility. Ninety-two reports were excluded at full text/report level, and 21 unique eligible studies were included. 3.2. Study Characteristics The final qualitative synthesis included 21 independent prospective clinical studies, comprising randomized controlled trials, prospective phase II studies, and early-phase interventional trials evaluating nab-paclitaxel-containing neoadjuvant regimens in patients with early-stage or locally advanced non-metastatic TNBC. The included studies originated from diverse geographic regions, including Europe, North America, and Asia, reflecting broad international clinical experience. Sample sizes ranged from small early-phase studies involving approximately 30 participants to large phase III randomized trials enrolling more than 400 patients. Studies were categorized according to treatment intensification strategy (). Platinum-containing nab-paclitaxel regimens without immune checkpoint inhibition were evaluated in WSG-ADAPT-TN, Yuan et al., and Mrózek et al. WSG-ADAPT-TN was randomized, whereas Yuan et al. and Mrózek et al. were prospective uncontrolled studies. Immunotherapy-containing nab-paclitaxel regimens were evaluated using different chemotherapy backbones and study designs, including randomized comparisons and prospective uncontrolled studies. These findings are not used to infer the efficacy of immunotherapy as an isolated component or to rank regimens across studies. Regimens containing both platinum and immune checkpoint inhibition were evaluated in randomized and prospective uncontrolled studies, including NeoTRIPaPDL1, CamRelief, cTRIO, and NeoSAC [,,,]. These studies differed in chemotherapy backbones, concomitant agents, populations, and comparators. Their pCR findings are therefore reported according to the actual study design and treatment contrast and are not interpreted as evidence that combined platinum–immunotherapy treatment is superior to other regimen categories or that the two components have additive effects. NeoPanDa03 additionally incorporated the VEGFR-2 inhibitor apatinib and therefore represents a multimodal regimen []. NeoSAC [] evaluated sintilimab, apatinib, carboplatin, and nab-paclitaxel in a prospective single-arm design and is retained as descriptive study-specific evidence rather than comparative evidence. 3.3. Risk of Bias and Certainty of Evidence Outcome-specific RoB 2 assessment was performed for the pCR outcome in all randomized studies. IMpassion031, GeparNuevo, NeoTRIPaPDL1, and CamRelief were judged to have low overall risk of bias for pCR. WSG-ADAPT-TN was judged to have some concerns because 336 patients were randomized whereas 324 constituted the reported pCR-evaluable population; this was reflected in the missing-outcome-data domain rather than assumed to be low risk. The exploratory randomized Kuemmel/CO42177 cohort was judged to have some concerns overall because of its phase Ib exploratory design and reporting context (A). The JBI appraisal was used only to describe methodological features of prospective uncontrolled studies; it was not used to determine quantitative denominator eligibility, which was adjudicated separately according to the rule in (B). The GRADE assessment of the certainty of evidence for the principal randomized checkpoint-inhibitor comparison is presented in C. 3.4. Pathological Complete Response Outcomes Pathological complete response outcomes are summarized according to the clinical contrast and study design (). WSG-ADAPT-TN compared nab-paclitaxel plus carboplatin with nab-paclitaxel plus gemcitabine and reported pCR rates of 45.9% and 28.7%, respectively. This finding is interpreted only as evidence for that specific chemotherapy-partner comparison and not as an estimate of an isolated platinum effect. In the randomized checkpoint-inhibitor trials, the relevant contrast was the addition of an immune checkpoint inhibitor to a trial-specific chemotherapy control. For example, IMpassion031 reported pCR of 58% with atezolizumab plus chemotherapy versus 41% with placebo plus chemotherapy. These results support the effect of checkpoint-inhibitor addition to the studied chemotherapy backbone; they do not represent immunotherapy used independently. Studies incorporating both platinum and checkpoint inhibition, including NeoTRIPaPDL1, CamRelief, and several prospective phase II studies, were summarized according to their individual regimen and study design. Because treatment backbones, concomitant agents, populations, and comparators differed, these findings do not permit attribution of an independent or additive effect to either platinum or immunotherapy alone. Study-specific pCR values are presented descriptively without cross-category ranking. No formal comparison of pCR proportions across platinum-containing, immunotherapy-containing, and combined-regimen categories was performed because the underlying studies were clinically and methodologically heterogeneous; consequently, a between-category p-value would not provide a valid estimate of differences attributable to individual treatment components. 3.5. Quantitative Meta-Analysis All 21 eligible studies were assigned an explicit analytical disposition. Four randomized checkpoint-inhibitor trials (IMpassion031, GeparNuevo, NeoTRIPaPDL1, and CamRelief) contributed to the pooled comparative RR analysis. WSG-ADAPT-TN and Kuemmel et al. were retained as randomized evidence but reported separately because their randomized contrasts addressed chemotherapy/platinum strategy rather than ICI addition versus a matched non-ICI control. Fourteen prospective non-randomized datasets with extractable pCR numerators and denominators contributed to the descriptive pooled proportion analysis. Yam et al. (atezolizumab) remained in the qualitative synthesis because the primary publication emphasized a composite pCR/RCB-I endpoint and an exact pure-pCR event count was not used in the pooled dataset.The study-level data and analytical disposition of studies contributing to the quantitative meta-analyses are summarized in . The randomized checkpoint-inhibitor dataset comprised 1228 participants: IMpassion031 (95/165 vs. 69/168), GeparNuevo (47/88 vs. 38/86), NeoTRIPaPDL1 (67/138 vs. 63/142), and CamRelief (126/222 vs. 98/219). The pooled random-effects estimate showed a higher probability of pCR with checkpoint-inhibitor addition to the chemotherapy backbone (RR = 1.25, 95% CI 1.12–1.41; I2 = 0%; p < 0.001). WSG-ADAPT-TN was reported separately because it evaluated a chemotherapy-partner comparison rather than checkpoint-inhibitor addition. The comparative findings are summarized in and displayed in . WSG-ADAPT-TN was not included in the pooled checkpoint-inhibitor estimate because it addressed a different randomized treatment contrast. Complete leave-one-out sensitivity analysis showed that the pooled checkpoint-inhibitor effect remained statistically significant after exclusion of each trial (). When IMpassion031 was omitted, the pooled RR was 1.20 (95% CI 1.05–1.38; I2 = 0%); after omission of GeparNuevo, RR = 1.26 (95% CI 1.11–1.43; I2 = 4.0%); after omission of NeoTRIPaPDL1, RR = 1.30 (95% CI 1.14–1.48; I2 = 0%); and after omission of CamRelief, RR = 1.24 (95% CI 1.07–1.45; I2 = 5.9%). These estimates were consistent with the primary pooled RR of 1.25 (95% CI 1.12–1.41), indicating that no single randomized trial materially altered the direction or statistical significance of the pooled result. The prospective non-randomized pCR synthesis included 14 datasets, comprising 539 participants and 320 pCR events. Study-level pCR event counts and denominators were as follows: Yuan et al. (32/67), Mrózek et al. (6/12), NeoTENNIS (39/70), cTRIO (35/62), NeoSAC (24/34), NeoPanDa03 (23/34), KEYNOTE-173 nab-paclitaxel cohorts A–D combined (26/40), Foldi et al. (24/55), NeoImmunoboost (33/50), Wang et al. (25/39), TREND (30/44), Chen et al. (9/10), Huang et al. (1/2), and Zheng et al. (13/20). The corresponding analytic populations were ITT for NeoSAC; the phase II ITT population at the recommended durvalumab dose for Foldi et al.; efficacy-evaluable for TREND; modified ITT efficacy-evaluable for Chen et al. (10 of 13 recruited participants); the separately extractable TNBC subgroup for Huang et al.; and surgery/evaluable for Zheng et al. Study-specific denominator definitions are reported in . Using a logit-transformed inverse-variance random-effects model with restricted maximum likelihood (REML) estimation and Hartung–Knapp–Sidik–Jonkman (HKSJ) inference, the pooled pCR proportion was 59.8% (95% CI 53.8–65.6%). Between-study heterogeneity was moderate (I2 = 32.7%; τ2 = 0.0559 on the logit scale; Cochran’s Q = 19.312, df = 13, p = 0.114). A sensitivity analysis using the DerSimonian–Laird estimator produced a similar pooled estimate, supporting the robustness of the result to the choice of between-study variance estimator. The REML model with HKSJ inference was retained as the primary analysis. No treatment-category-specific pooled proportions were used to compare the prospective non-randomized studies. Individual study-level pCR proportions are displayed in and the corresponding event/denominator data are provided in . These values are not ranked or interpreted as evidence of superiority or incremental benefit between treatment strategies because the underlying studies are clinically and methodologically non-equivalent. Sensitivity analyses using the same REML logit random-effects model with Hartung-Knapp-Sidik-Jonkman inference supported the stability of the descriptive pooled estimate. Excluding the Huang et al. TNBC subgroup yielded 59.9% (95% CI 53.6–65.9%); excluding the Chen et al. cohort yielded 59.3% (95% CI 53.7–64.7%); excluding both yielded 59.4% (95% CI 53.5–65.0%). A denominator-sensitivity analysis excluding TREND, Chen et al., and Zheng et al. yielded 58.1% (95% CI 51.8–64.2%). These analyses did not materially alter the interpretation of the primary estimate. The randomized comparative findings and the descriptive overall single-arm estimate are summarized in . No cross-category pooled single-arm estimates are presented because the single-arm studies are not directly comparable. Formal assessment of publication bias was not undertaken for the checkpoint-inhibitor comparative meta-analysis because only four randomized trials were available; funnel-plot asymmetry and regression-based tests are not considered reliable with such a small number of studies. 4. Discussion This systematic review included 21 unique eligible prospective studies. Four randomized checkpoint-inhibitor trials involving 1228 participants yielded a pooled RR of 1.25 (95% CI 1.12–1.41; I2 = 0%). Separately, 14 prospective non-randomized datasets involving 539 analysis participants and 320 pCR events yielded a pooled pCR proportion of 59.8% (HKSJ 95% CI 53.8–65.6%; I2 = 32.7%). These analyses address different questions and should be interpreted separately; the non-randomized pooled proportion is descriptive and does not establish comparative efficacy across regimens. TNBC remains one of the most clinically challenging breast cancer subtypes because of its aggressive biology, early recurrence pattern, and lack of endocrine or HER2-targeted treatment options [,]. Chemotherapy, therefore, remains a cornerstone of treatment, particularly in the neoadjuvant setting where treatment response can be directly assessed []. Achieving pCR has important individual-level prognostic significance in TNBC and is consistently associated with more favorable long-term outcomes, including event-free and overall survival []. Additional meta-analytic evidence supports the prognostic association of pCR with favorable long-term outcomes, particularly in TNBC [,], while residual cancer burden provides further prognostic stratification across the spectrum of residual disease after neoadjuvant therapy []. These patient-level associations should not be interpreted as validation of pCR as a trial-level surrogate endpoint for survival or as evidence that a treatment-induced increase in pCR necessarily produces a corresponding improvement in event-free or overall survival. The biological rationale for platinum chemotherapy and immune checkpoint inhibition remains plausible; however, the available evidence does not permit the independent contribution of each component to be inferred across heterogeneous regimens. The randomized checkpoint-inhibitor analysis estimates the effect of adding checkpoint blockade to trial-specific chemotherapy controls, whereas WSG-ADAPT-TN addresses a carboplatin-versus-gemcitabine chemotherapy-partner comparison. The prospective single-arm evidence cannot establish comparative efficacy and is therefore interpreted only as descriptive evidence of pathological response within each study population. These findings should also be considered alongside broader platinum-based neoadjuvant evidence in TNBC, including meta-analytic data and randomized trials demonstrating increased pathological response with carboplatin-containing chemotherapy [,], as well as long-term follow-up from the BrighTNess trial demonstrating an event-free survival benefit with carboplatin-containing treatment []. Several prospective studies reported clinically notable pCR proportions with nab-paclitaxel-containing regimens, including cTRIO, NeoSAC, NeoPanDa03, and NeoTENNIS. These observations are presented as study-specific findings only. They do not demonstrate that platinum, immunotherapy, or their combination is independently effective, nor do they establish superiority or an incremental benefit attributable to one treatment component. Component-specific efficacy requires randomized designs in which the component of interest is the planned difference between otherwise comparable treatment arms. An important distinction of this review is its specific focus on nab-paclitaxel-containing regimens. Many broader TNBC reviews combine solvent-based paclitaxel and nab-paclitaxel despite important pharmacologic differences [,,]. Nab-paclitaxel avoids Cremophor-associated hypersensitivity reactions, permits steroid-free administration, and may enhance intratumoral drug delivery through albumin-mediated transport mechanisms [,]. These characteristics may be particularly relevant in chemoimmunotherapy combinations, where minimizing corticosteroid exposure could theoretically preserve immune activation. The principal comparative finding of this review is limited to pCR: across four randomized trials, adding immune checkpoint blockade to the trial-specific chemotherapy backbone was associated with a higher probability of pCR. WSG-ADAPT-TN provides a separate randomized comparison of a carboplatin-containing regimen with a gemcitabine-containing regimen on a nab-paclitaxel backbone. The single-arm studies provide descriptive response estimates only and should not be used to infer superiority, treatment ranking, additive effects, or incremental benefit across regimen categories. Although pCR is prognostically important, these results should not be interpreted as demonstrating improvement in event-free survival, overall survival, toxicity, quality of life, or overall clinical benefit. Such outcomes require direct assessment from appropriately designed randomized trials and mature follow-up. Broader randomized and meta-analytic evidence in early TNBC provides relevant survival data for specific perioperative regimens [,]. In particular, mature follow-up from KEYNOTE-522 demonstrated an overall survival benefit with the pembrolizumab-containing perioperative regimen []. However, these survival data are contextual and do not represent treatment effects estimated by the present pCR meta-analysis. The broader effect of neoadjuvant checkpoint inhibition in early TNBC has been evaluated in larger syntheses; for example, Villacampa et al. included 2097 patients with TNBC and reported an increase in pCR from 46.6% to 59.9% with neoadjuvant checkpoint inhibition []. The specific contribution of the present review is therefore its restriction to nab-paclitaxel-containing neoadjuvant backbones and its study-level synthesis of randomized and prospective non-randomized evidence within that treatment context. This study has several strengths. It provides a focused evaluation of nab-paclitaxel-containing neoadjuvant regimens in TNBC and incorporates contemporary platinum- and immunotherapy-based studies. The randomized evidence was synthesized according to the clinical question addressed: checkpoint-inhibitor trials formed the comparative meta-analysis, whereas the WSG-ADAPT-TN platinum comparison was analyzed separately. The structured PRISMA-based methodology, formal risk-of-bias assessment, duplicate data extraction, and quantitative pooling of clinically comparable studies support the interpretability of the evidence synthesis. Several limitations should be acknowledged. First, only four randomized trials were suitable for the checkpoint-inhibitor comparative meta-analysis, limiting precision and prohibiting reliable formal assessment of publication bias. Second, even within this coherent comparison, treatment backbones, sequencing, and checkpoint inhibitors varied across studies. Third, randomized platinum/chemotherapy-partner evidence from WSG-ADAPT-TN and Kuemmel et al. addressed different treatment contrasts and therefore was not combined with the checkpoint-inhibitor trials. Fourth, the 14-dataset prospective non-randomized synthesis remains susceptible to selection bias, confounding, heterogeneous populations, regimens, sequencing, concomitant therapies, and pCR definitions. Some reports also used surgery- or efficacy-evaluable denominators rather than the full enrolled population; this was addressed in sensitivity analysis. Future randomized studies should prioritize event-free survival (EFS) and overall survival (OS) alongside pCR, with adequate follow-up to determine whether improvements in pathological response translate into durable clinical benefit. Future studies should also incorporate biomarker-guided patient selection, harmonized pCR definitions, toxicity and quality-of-life endpoints, and randomized component-separation designs when the independent contribution of platinum or checkpoint inhibition is the specific question. These priorities are consistent with contemporary randomized evidence in early TNBC, in which survival benefits may emerge with longer follow-up despite earlier differences in pathological response [,,]. 5. Conclusions The randomized evidence synthesized in this review indicates a higher probability of pCR when an immune checkpoint inhibitor is added to the chemotherapy backbones studied. WSG-ADAPT-TN separately reported higher pCR with a nab-paclitaxel plus carboplatin regimen than with a nab-paclitaxel plus gemcitabine regimen. These findings represent distinct regimen-level treatment contrasts and should not be interpreted as proving an isolated platinum effect, an isolated immunotherapy effect, additive benefit from combining treatment components, or superiority across regimen categories. The pooled proportion from prospective non-randomized studies is descriptive only. Because the present synthesis primarily evaluates pCR, no inference is made regarding event-free survival, overall survival, toxicity, quality of life, or overall clinical benefit; these outcomes require direct evidence from mature, randomized studies. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Funding Statement This research received no external funding. Footnotes