De‐Escalation of Neoadjuvant Treatment for HER2‐Positive Breast Cancer Authors: Yiyin Ma, Dedian Chen, Sheng Huang Published: 2026 ABSTRACT ABSTRACTWith improved prognosis in early HER2‐positive breast cancer, safely de‐escalating treatment is increasingly investigated to minimize overtreatment. This article focuses on neoadjuvant de‐escalation strategies for HER2‐positive breast cancer. There have been initiatives to shorten the duration of anti‐HER2 therapy or to reduce the use of chemotherapy agents. Predictive or prognostic biomarkers and imaging techniques might help to identify subsets of patients within HER2‐positive cancer with the best chance of responding to treatment and benefiting from de‐escalated treatment. Abbreviations 18F‐fluorodeoxyglucose‐PET antibody‐drug conjugates adjuvant therapy breast cancer Chimeric Antigen Receptor T‐cell cell‐free DNA chemotherapy circulating tumor cells circulating tumor DNA dedicated breast positron emission tomography dynamic contrast‐enhanced disease‐free survival diffusion‐weighted imaging early‐stage breast cancer event‐free survival epidermal growth factor receptor estrogen receptor endocrine therapy Food and Drug Administration functional tumor volume Human Epidermal Growth Factor Receptor 2 Positive HER2‐enriched hormone receptor immune checkpoint inhibitors invasive disease‐free survival liquid biopsy lymphovascular invasion MicroRNAs minimal residual disease magnetic resonance imaging neoadjuvant treatment overall survival Prediction Analysis Of Microarray 50 pathological complete response Programmed Death Receptor‐1 Programmed Death Ligand‐1 positron emission tomography‐computed tomography progesterone receptor predicted residual cancer burden recurrence‐free survival receptor tyrosine kinase signal enhancement ratio standardized uptake values Trastuzumab Emtansine Trastuzumab Deruxtecan tumor‐infiltrating lymphocytes tyrosine kinase inhibitors triple negative breast cancer triple positive breast cancer Background Breast cancer (BC) has become the most commonly diagnosed cancer worldwide, accounting for 23.8% of new cases in 2022 [1], with over 2.3 million new cases and approximately 670,000 deaths reported that year alone [2]. Human epidermal growth factor receptor 2 (HER2)‐positive (HER2+) BC, an aggressive phenotype comprising 15%–20% of all BC cases, is characterized by high recurrence rates and inferior survival outcomes [3, 4]. Over the past two decades, HER2‐targeted therapy has significantly improved the prognosis of HER2+ BC patients in both early and advanced stages. The FDA has approved a range of HER2‐targeted agents for use in HER2+ BC in the (neo)adjuvant and/or metastatic settings, which are listed as follows: (1) monoclonal antibodies targeting the extracellular domain of HER2 include trastuzumab, pertuzumab, and margetuximab; (2) tyrosine kinase inhibitors (TKIs) which bind to the intracellular tyrosine kinase domains of HER2 and other HER family members, including lapatinib, neratinib, and tucatinib; (3) antibody‐drug conjugates (ADCs), for instance, trastuzumab emtansine (T‐DM1) and trastuzumab deruxtecan (T‐DXd). Ongoing research continues to explore the efficacy, safety, and incorporation of novel HER2‐targeted agents into current therapeutic strategies [5]. This review aims to elucidate the present therapeutic landscape of HER2+ BC, emphasizing approaches for treatment de‐escalation. Role of Neoadjuvant Treatment in HER2‐Positive BC Neoadjuvant treatment (NAT) has transformed from a purely surgical downstaging tool into a sophisticated platform guiding individualized treatment. Administered preoperatively, NAT now routinely incorporates HER2‐targeted or endocrine therapy (ET) alongside chemotherapy based on BC subtype. For high‐risk early‐stage HER2+ BC (tumor diameter ≥ 2 cm), the current standard combines chemotherapy with dual HER2 blockade (trastuzumab plus pertuzumab) [6]. Initially used for tumor reduction, NAT has increased breast‐conserving surgery rates to over 40% in patients previously requiring mastectomy [7, 8]. NAT accelerates therapy development [9, 10], allowing real‐time tumor response assessment to guide postoperative decisions [11]. Patients with residual invasive cancer should receive 14 cycles of T‐DM1, while those achieving pCR should complete 18 cycles of trastuzumab and pertuzumab [12, 13, 14]. The Relation Between pCR and Long‐Term Outcomes in HER2‐Positive BC The US FDA accepted pathologic complete response (pCR) as a surrogate endpoint for accelerated approval of new drugs in NAT for high‐risk early‐stage BC (EBC) [15]. Definitions of pCR varied across trials. The NeoSTEEP working group recommends pCR as the absence of residual invasive cancer in the resected breast specimen and all sampled regional lymph nodes (ypT0/Tis ypN0 per AJCC staging) [16]. The prognostic value of pCR is subtype‐dependent, with the strongest associations consistently demonstrated in early HER2+ BC and triple negative (TNBC) [17, 18]. In 12 international trials with nearly 12,000 patients, the CTNeoBC group confirmed pCR as a robust prognostic indicator at the individual level in these subtypes. Although it failed to validate pCR as a surrogate for event‐free survival (EFS) or overall survival (OS) at the trial level [19]. Nonetheless, the survival benefit conferred by pCR is substantial: a meta‐analysis of 4 randomized trials involving 1410 patients reported relative reductions of 55% in recurrence‐free survival (RFS) and 68% in OS among those achieving pCR [20]. De‐Escalation Strategies and Evidence A good therapeutic scheme balances the benefits versus risks of systemic therapy, aiming for pCR with de‐escalated treatments to reduce toxicity without compromising outcomes, especially for patients at lower risk of relapse or those intolerant to therapy‐related toxicity [21]. Chemotherapy‐Reducing Regimens Shorten the Chemotherapy Duration Abbreviating chemotherapy has emerged as a viable de‐escalation strategy. A large retrospective study compared the pCR status of HER2+ BC patients receiving four or six cycles of TCbH (taxane, carboplatin, and trastuzumab), sharing similar high pCR rates of 46.5% and 49.9% [22]. This suggests that shortening chemotherapy to four cycles in patients with excellent responses to anti‐HER2 therapies may be an effective de‐escalation strategy, as two additional cycles of TCbH may not significantly alter the final tumor response. Both NeoSphere [23, 24] and PEONY [25, 26] trials demonstrated comparable pCR rates (39.3%–45.8%) and favorable long‐term outcomes with four‐cycle THP regimens (trastuzumab, pertuzumab, and docetaxel). The incorporation of TKIs has broadened de‐escalation possibilities. Compared to monoclonal antibodies, TKIs offer potent antitumor activity, oral administration, broader pathway inhibition, and lower cardiotoxicity [27, 28]. Several phase II trials evaluated four‐cycle pyrotinib‐based regimens in HER2+ EBC, reporting promising pCR rates: 41.0% in PHEDRA with pyrotinib, trastuzumab, and docetaxel [29]; 57.1% with pyrotinib, trastuzumab, and albumin‐bound paclitaxel [30]; 69.8% in NeoATP with pyrotinib, trastuzumab, weekly paclitaxel, and cisplatin [31]; and 66.7% in an ongoing trial (NCT054444998) combining pyrotinib, inetetamab, and albumin‐bound paclitaxel, rising to 72.7% in the hormone receptor positive (HR+)/HER2+ subgroup [32]. Anthracycline‐Free Chemotherapy Anthracycline‐taxane‐based combinations have traditionally been standard for HER2+ NAT [12], but studies indicated that the benefits of anthracyclines are limited to a small subset of BC cases [33]. Anthracycline‐free, trastuzumab‐based adjuvant therapy (AT) has emerged as a viable de‐escalation strategy for selected patients with HER2+ EBC [34, 35]. The TRAIN‐2 trial provided further evidence for omitting anthracyclines in NAT for HER2+ EBC. Patients in the anthracycline group received three‐cycle FEC (5‐fluorouracil, epirubicin, and cyclophosphamide) plus trastuzumab and pertuzumab (HP), followed by six‐cycle TCbHP (paclitaxel/carboplatin/trastuzumab/pertuzumab). The non‐anthracycline group underwent nine cycles of weekly TCbHP. The comparison between the anthracycline group and the non‐anthracycline group shared equivalent pCR rates (67% vs. 68%, p = 0.95) [36]. At 3 years, EFS was also comparable, with 92.7% for the anthracycline group and 93.6% for the non‐anthracycline group [37]. These results support carboplatin and paclitaxel as valid alternatives to anthracyclines in HER2+ EBC. In the TRYPHAENA trial, a higher pCR rate (66.2%) was achieved with the non‐anthracycline regimen (docetaxel/carboplatin/trastuzumab/pertuzumab) compared to other anthracycline‐containing regimens [38]. Collectively, these trials confirm that anthracyclines offer no significant benefit when potent HER2‐targeted therapy is administered. Consequently, the 2021 NCCN Guidelines removed anthracycline‐based therapy from “preferred regimens” to an option for selective use [39]. Chemotherapy‐Free Regimens There is increasing attention on assessing neoadjuvant chemotherapy‐free regimens in HER2+ EBC. In the WSG‐ADAPT‐HER2+/HR‐ Trial, adding taxane to trastuzumab/pertuzumab significantly increased pCR rates (90.5% vs. 27%) [40]. Yet this dramatic improvement in pathologic response failed to translate into superior long‐term survival. The update of the 5‐year survival outcomes showed no significant differences between the chemotherapy (THP) and non‐chemotherapy (HP) arms in disease‐free survival (DFS, 98% vs. 92%, p = 0.36) or OS (98% vs. 94%, p = 0.43) [41]. This revealed that a modest yet notable number of HER2+ EBC patients could have excellent prognosis following an effective short‐term anti‐HER2 treatment, even without chemotherapy. Endocrine Therapy Combined With HER2‐Targeted Therapy The pCR rates vary by HR‐status and are higher in HR‐/HER2+ than in HR+/HER2+ EBC [19]. A subset of “triple positive” (Estrogen Receptor[ER]/Progesterone Receptor[PR]/HER2‐positive) tumors, or those with high HR expression, may represent a group with particularly favorable prognosis. For these patients, combining chemotherapy, HER2 blockade, and ET might be considered overtreatment [42, 43]. Yet, simply omitting chemotherapy without incorporating ET appears insufficient. In the chemotherapy‐free arm (trastuzumab plus pertuzumab) of the NeoSphere trial, where ET was excluded, the pCR in the HR+/HER2+ cohort was merely 6% [23]. The WSG‐TP‐II trial randomized HR+/HER2+ EBC patients to receive either 12 weeks of ET plus trastuzumab and pertuzumab, or paclitaxel plus trastuzumab and pertuzumab (ET + HP vs. CT + HP). Despite significantly lower pCR in the ET arm (23.7% vs. 56.4%) [44, 45], 5‐year EFS (92.0% vs. 94.8%, p = 0.65) and OS (100% vs. 97.9%, n.s.) were comparable [46]. For patients who achieve early pCR, it is possible to safely de‐escalate CT in early HR+/HER2+ BC with effective HER2‐targeted therapy. The phase II WSG‐ADAPT‐TP Trial evaluated T‐DM1‐based chemotherapy‐free regimens. Patients received 12 weeks of T‐DM1, T‐DM1/ET, or trastuzumab/ET. Both T‐DM1‐containing regimens achieved similar pCR rates (T‐DM1: 41.0% vs. T‐DM1/ET: 41.5%), far exceeding trastuzumab/ET (15.1%; p < 0.001) [47]. Ki67 molecular response at 3 weeks predicted pCR. Long‐term survival was comparable across all groups. The 5‐year survival analysis showed similar invasive disease‐free survival (iDFS) (88.9% of the T‐DM1 arm, 85.3% of the T‐DM1/ET arm, 84.6% of the trastuzumab/ET arm, p = 0.608) and OS (97.2%, 96.4%, 96.3%, p = 0.534). The trial indicated that pCR after 12 weeks of chemotherapy‐free NAT correlated with superior survival in HR+/HER2+ EBC without additional adjuvant chemotherapy [48]. TBCRC006 and TBCRC023 trials assessed the combination of ET with dual HER2‐targeted therapy in the neoadjuvant setting. TBCRC006 (12 weeks of trastuzumab/lapatinib, with letrozole for ER+) achieved an overall pCR of 27% (ER+: 21%, ER−: 36%) [49]. TBCRC023 compared 12 versus 24 weeks of the same regimen. Extended treatment nearly tripled pCR in ER+ patients (9% to 33%) but conferred no benefit in ER‐ patients (20% vs. 18%) 50, suggesting that prolonged HER2 blockade may compensate for chemotherapy omission in HR+ disease [50]. CDK4/6‐Associated Chemotherapy‐Free Strategies Building on the established role of CDK4/6 inhibitors in ER+ BC, several trials have investigated their incorporation into chemotherapy‐free regimens for HR+/HER2+ disease [9]. The phase II PALTAN trial evaluated a four‐cycle regimen of palbociclib, letrozole, and trastuzumab in stage II/III, ER+/HER2+ BC. The trial was terminated early due to a low observed pCR rate of only 7.7% [51]. However, in the phase II MUKDEN01 study, patients with stage II‐III triple‐positive BC (TPBC) were treated with five cycles of chemotherapy‐free NAT with pyrotinib, letrozole, and dalpiciclib, attaining a pCR of 29.5% [52]. The difference in the number of NAT cycles is unlikely to account for the observed disparity in pCR rates between the studies. In the phase II NA‐PHER2 study, patients with HR+/HER2+ BC received six cycles of trastuzumab and pertuzumab combined with five cycles of palbociclib and fulvestrant, resulting in a pCR of 27%. Triple targeting of ER, HER2, and RB1 may offer an effective chemotherapy‐free strategy for ER+/HER2+ BC [53]. The phase II TOUCH trial enrolled postmenopausal patients with HR+/HER2+ EBC, randomized to receive either weekly paclitaxel for four cycles (Arm A) or palbociclib for four cycles plus letrozole for 16 weeks (Arm B). All patients received five cycles of trastuzumab and pertuzumab. pCR rates were similar between arms (32.9% in Arm A and 33.3% in Arm B), demonstrating that palbociclib plus letrozole with dual HER2 blockade can achieve comparable pCR rates to weekly paclitaxel [54]. ADCs‐Associated Chemotherapy‐Free Strategies Trials evaluating T‐DM1‐based regimens have shown modestly lower pCR and comparable outcomes compared to conventional chemotherapy, but with improved tolerability. In the phase III KRISTINE trial, patients with stage II/III HER2+ EBC received six cycles of either T‐DM1/pertuzumab (T‐DM1 + P) or docetaxel/carboplatin/trastuzumab/pertuzumab (TCbHP). The TCbHP arm showed a higher pCR rate compared with the T‐DM1 + P arm (55.7% vs. 44.4%), with increased adverse events [55]. The TCbHP group showed a 3‐year EFS of 94.2%, compared to 85.3% for T‐DM1 + P group. Both arms shared equivalent 3‐year iDFS and OS [56]. Similarly, the phase II PREDIX HER2 trial compared neoadjuvant docetaxel/trastuzumab/pertuzumab (THP) with T‐DM1, with no significant difference in pCR rates (45.5% vs. 43.9%, p = 0.82) [57]. In the I‐SPY2 trial, T‐DM1 + P and THP (paclitaxel/trastuzumab/pertuzumab) arm were compared to a control arm (paclitaxel/trastuzumab alone). The predicted pCR rates were 63% for T‐DM1 + P, 72% for THP, and 33% for the control, with 3‐year EFS rates of 88%, 92%, and 87%, respectively [58]. Next‐generation ADCs are now entering the neoadjuvant arena. The ongoing phase III DESTINY‐Breast 11 trial is the first to evaluate T‐DXd in the neoadjuvant setting for high‐risk HER2+ EBC. Patients are randomized into three arms: eight‐cycle T‐DXd monotherapy, four‐cycle T‐DXd followed by four‐cycle paclitaxel/trastuzumab/pertuzumab (THP), or standard four‐cycle doxorubicin and cyclophosphamide followed by four‐cycle THP (ddAC‐THP) [59]. SHR‐A1811 is a third‐generation HER2‐targeted ADC. In the phase II FASCINATE‐N trial, patients were randomized to receive 24‐week NAT with either SHR‐A1811 monotherapy, SHR‐A1811 plus pyrotinib, or standard regimen (nab‐paclitaxel/carboplatin/trastuzumab/pertuzumab [TCbHP]). SHR‐A1811 monotherapy arm achieved a pCR rate of 63.2%, comparable to the 64.4% with standard treatment. Adding pyrotinib to SHR‐A1811 resulted in a pCR of 62.5% but increased toxicity without improved efficacy [60]. Long‐term efficacy and safety remain to be confirmed. Immunotherapy Combined With HER2‐Targeted Therapy Immunotherapy for HER2+ BC can be classified into four categories based on their mechanisms: (1) passive immunotherapy, such as monoclonal antibodies; (2) immune checkpoint inhibitors (ICIs), including programmed death receptor‐1 (PD‐1)/programmed death ligand 1 (PD‐L1) inhibitors and HER2‐directed monoclonal antibodies; (3) adoptive cell immunotherapy, exemplified by chimeric antigen receptor (CAR)‐T cell therapy; and (4) active immunotherapy, involving vaccines [61]. HER2+ BC is distinguished by higher PD‐L1 expression and tumor‐infiltrating lymphocytes (TILs) [62]. Evidence indicated that trastuzumab's immune‐modulating effect could be enhanced through structural refinement or synergistic combinations with other agents [63]. In EBC, where the immune system remains intact, integrating HER2‐targeted treatments with ICIs offers a strong rationale for improving efficacy [62]. The phase II Neo‐PATH sought to replace carboplatin in the TCbHP regimen with an ICI to improve safety while maintaining efficacy. Patients with stage II/III HER2+ BC received atezolizumab, docetaxel, trastuzumab, and pertuzumab, demonstrating a promising pCR of 61% with moderate toxicity [64]. The WSG‐KEYRICHED‐1 trial is the first to combine chemotherapy‐free NAT with immunotherapy. Patients with HER2‐enriched (HER2‐E) EBC underwent four cycles of pembrolizumab, trastuzumab, and pertuzumab. This 12‐week chemotherapy‐free regimen achieved a pCR of 46% [65]. In the DTP trial, patients with HER2‐E/ER‐/PR‐ received six cycles of trastuzumab, pertuzumab, and Durvalumab, resulting in a pCR of 48.6% [66]. These regimens offer a potentially effective, low‐toxicity alternative to conventional chemotherapy in this HER2‐E subtype. The KEYRICHED‐1 trial showed a higher pCR rate in PD‐L1‐positive tumors, consistent with the Neo‐PATH and DTP trials [67]. Future Perspectives: Rationale for Precision Medicine The evolution toward de‐escalation in HER2+ BC is predicated on the ability to precisely identify which patients can safely receive less intensive therapy. Precision medicine aims to improve cancer diagnosis and treatment by utilizing molecular information to identify predictive markers that inform treatment decisions, classify molecular subtypes, monitor treatment responses, and detect resistance and recurrence [68]. Intrinsic Subtyping The Prediction analysis of microarray 50 (PAM50), a clinically applicable gene expression test classifies BC into five intrinsic subtypes based on the expression of 50 genes, with 93% accuracy [69, 70]. These subtypes include luminal A, luminal B, HER2‐E, basal‐like, and normal‐like. Among these, the HER2‐E subtype, accounting for 10%–15% of all cases, is distinguished by high ERBB2 and EGFR/HER2‐regulated gene expression and is more likely to respond to HER2‐targeted therapies, such as trastuzumab, pertuzumab, and TKIs like lapatinib and pyrotinib [70, 71]. Gene expression profiling has shown that approximately 30% of HR+/HER2+ BC fall under the HER2‐E subtype, exhibiting robust activation of the ERBB2 pathway and relative suppression of ER signaling [70, 71]. In contrast, approximately 70% of HR+/HER2+ cancers exhibit a luminal‐like phenotype [70, 71], for which the effectiveness of standard anti‐HER2 therapies and chemotherapy is limited, highlighting a significant unmet clinical need [72]. The PAMELA trial revealed that HER2‐E is a strong predictor of dual HER2 blockade sensitivity, with a fourfold higher likelihood of achieving pCR compared to non‐HER2‐E subtypes, irrespective of HR status [73]. PAM50 intrinsic subtyping may help identify patients eligible for chemotherapy‐sparing. Beyond the PAM50 classification, a recent multiomics profiling analysis refined HER2+ BC into four transcriptome‐based subtypes with distinct molecular and therapeutic implications. Approximately 28.3% of tumors fall into the classical (HER2‐CLA) subtype, characterized by pervasive ERBB2 signaling and exceptional responsiveness to anti‐HER2 therapies [74]. Another 20% constitute the immunomodulatory (HER2‐IM) subtype, distinguished by a robust tumor immune infiltrate [74]. This phenotype predicts favorable outcomes with standard therapy, implicating potential for de‐escalation, and simultaneously identifies candidates for immunotherapy. The luminal‐like (HER2‐LUM, 30.6%) subgroup shares the molecular landscape of hormone‐driven, HER2‐negative cancers, suggesting that treatment should integrate endocrine agents and CDK4/6 inhibitors [74]. Conversely, the basal/mesenchymal‐like (HER2‐BM, 21.1%) subtype is largely refractory to conventional dual HER2 blockade but displays activation of compensatory receptor tyrosine kinase (RTK) pathways, particularly involving EGFR, indicating potential susceptibility to specific TKIs [74]. Such a classification is pivotal for personalizing NAT intensity and drug selection. Prognostic Biomarkers Liquid Biopsy (LB) LB is a non‐invasive, accessible, and repeatable approach for real‐time monitoring of tumor‐derived components released into the bloodstream or other body fluids from primary or metastatic lesions [75]. These components include cell‐free DNA (cfDNA), circulating tumor DNA (ctDNA), microRNAs (miRNAs), non‐coding RNA, and extracellular vesicles such as exosomes. These substances may originate from single circulating tumor cells (CTCs) or tumor cell clusters [76]. CTCs provide valuable insights into tumor biology and disease progression. In the BEVERLY‐2 trial, HER2+ inflammatory BC patients received neoadjuvant bevacizumab and trastuzumab. Sequential CTCs assessments, combined with pCR status stratified relapse risk: persistent CTCs without pCR indicated high recurrence risk, while early CTCs clearance with pCR predicted favorable outcomes [77]. CTCs‐positive patients may benefit from intensified or novel therapies, such as ICIs, requiring further validation. ctDNA, a tumor‐derived component of cfDNA, is a promising neoadjuvant biomarker [76]. The presence of ctDNA at baseline and persistence after NAT predicts higher relapse risk and worse outcomes (RFS and OS), independent of pCR [78]. This may reflect elevated ctDNA shedding in aggressive subtypes (e.g., HER2+, TNBC) [79, 80]. Alternatively, ctDNA clearance reflects minimal residual disease (MRD), while pCR indicates the absence of detectable invasive cancer, offering complementary prognostic value [81]. Detectable ctDNA within 2–3 weeks of NAT initiation correlates with lower pCR rates and increased recurrence risk [79, 80, 82]. During NAT, ctDNA enables noninvasive detection of genomic alterations, such as PIK3CA mutations which are linked to poor pCR [83, 84]. In HER2+ BC, tracking PIK3CA through ctDNA may help identify patients who could benefit from combined anti‐HER2 and PI3K inhibition. Nevertheless, the optimal timing for ctDNA assessment remains uncertain. The ongoing HARMONY study investigates ctDNA/ctRNA‐based MRD evaluation to stratify patients. By analyzing serial samples before and after NAT, it aims to spare low‐risk patients from overtreatment while intensifying therapy for high‐risk groups [85]. miRNAs participate in critical tumor‐related processes including proliferation, differentiation, invasion, angiogenesis, and apoptosis, and modulate both innate and adaptive immune responses [86, 87]. Plasma‐based miRNA profiling may serve as a promising tool to predict response to trastuzumab and guide de‐escalation strategies [88]. Other Prognostic Biomarkers Tumor specimens remain valuable for correlative biomarker analysis, although most available data are from retrospective studies with baseline samples and small exploratory cohorts. Lymphovascular invasion (LVI), tumor cells in lymphatic or blood vessels, is associated with increased risks of node involvement, distant metastasis, and mortality [89, 90]. Several neoadjuvant studies further link LVI to worse outcomes [91, 92]. TILs, mainly referring to αβ T cells (CD4+ and CD8+), represent all lymphocytic populations infiltrating tumor tissue [93]. In the shortHER trial, TILs showed independent prognostic value for MFS in both HR+ and HR‐ subgroups. Low TILs were linked to benefit from 1‐year trastuzumab, whereas high TILs indicated excellent outcomes regardless of treatment duration [94]. Patients with TILs ≥ 20% receiving de‐escalated chemotherapy and shorter trastuzumab showed no increased risk of relapse or death [95]. A retrospective study found TILs ≥ 10% independently predicted pCR and better outcomes in HER2+ BC [96]. The prognostic role of TILs was further validated in NeoALTTO, PAMELA, and PREDIX HER2 trials, where high TILs or HER2‐E subtype correlated with strong responses to anti‐HER2 therapy [57, 97, 98]. Predictive Biomarkers The phase II PerELISA study assessed whether a chemotherapy‐free regimen with dual anti‐HER2 blockade could benefit HR+/HER2+ patients selected by Ki67 response after 2 weeks of letrozole. HER2‐E patients with molecular response achieved a pCR rate of 45%, indicating that combining intrinsic subtype and Ki67 inhibition enhances personalized treatment. A correlation between Ki67 change, HR expression, and PAM50 subtyping was observed, confirming its potential for tumor differentiation. The PerELISA study also demonstrated that significant pCR is achievable without chemotherapy based on Ki67 reduction after short‐term letrozole [99]. Based on the PerELISA trial, the HER2DX genomic test was developed and validated as a 27‐gene expression and clinical feature‐based classifier. HER2DX delivers two independent scores predicting long‐term prognosis and the likelihood of achieving pCR after neoadjuvant anti‐HER2 treatment. Compared to the PAM50 HER2‐E subtype, HER2DX provides more predictive value [100]. Emerging evidence indicates that HER2DX may help identify a subgroup of HR+/HER2+ EBC patients who could avoid chemotherapy when treated with anti‐HER2 therapy and ET. Single biological markers may not fully assess treatment response or survival, nor guide early treatment decisions such as chemotherapy necessity. Integrating additional biomarkers could expand safe de‐escalation strategies. Neoadjuvant trials using pCR or biomarkers like Ki67 as primary endpoints are designed to develop drugs, but do not alter standard care until confirmatory survival trials are completed. Imaging Techniques Clinical evaluation of NAT efficacy through physical examination and imaging techniques is necessary before surgery. Conventional imaging aids early pCR prediction, helps avoid unnecessary toxicity for non‐responders, and guides alternative treatments. Clinical response to NAT occurs in around 80% of cases, with rare disease progression. After NAT, ultrasound and mammography are routinely performed, while MRI is recommended for surgical planning. Postsurgical pathologic evaluation remains the gold standard for assessing response [101]. MRI‐derived metrics of peak signal enhancement ratio (SER) and functional tumor volume (FTV) are superior to clinicopathologic evaluation in predicting NAT response and survival [102]. The TRAIN‐3 trial explored the potential for tailoring chemotherapy duration based on radiological response in stage II–III HER2+ BC receiving neoadjuvant paclitaxel/carboplatin/trastuzumab/pertuzumab. Breast MRI and lymph node biopsy evaluated response every three cycles. Patients with complete radiological response or completing nine cycles proceeded to surgery. Among HR‐/HER2+ patients, one‐third achieved pCR after three cycles, and 87% with complete MRI response achieved pCR. Complete radiological response could identify early complete responders in HR‐ tumors, potentially shortening treatment duration and reducing adverse effects, pending 3‐year EFS validation. However, breast MRI is insufficient for HR+ tumors, with only 53% achieving pCR despite complete radiological response, necessitating the development of precise predictive biomarkers [103]. The novel metabolism‐based PET‐CT offers advantages in differentiating fibrosis or histiocytosis in BC and metastatic lymph nodes after NAT, where mammography or ultrasound displays ambiguous or no morphologic changes [104]. Previous studies demonstrated that PET exhibited higher specificity in evaluating axillary lymph node response to NAT [105, 106, 107, 108]. 18F‐fluorodeoxyglucose‐PET (18F‐FDG‐PET) quantifies tumor glycolysis and may detect early therapeutic responses before observable morphological alterations on anatomical images. This early metabolic evaluation may identify HER2+ tumors with high anti‐HER2 sensitivity and a higher likelihood of pCR. In the NeoALTTO trial, 18F‐FDG‐PET predicted early response to neoadjuvant lapatinib, trastuzumab, and their combination [109]. The TBCRC026 study demonstrated that standardized uptake values (SULmax) on 18F‐FDG‐PET/CT predict pCR in HR‐/HER2+ EBC patients receiving neoadjuvant trastuzumab/pertuzumab [110]. In the phase II PHERGAIN study, patients with HER2+ stage I–IIIA EBC received NAT with either chemotherapy (docetaxel/carboplatin) plus trastuzumab/pertuzumab (Group A) or trastuzumab/pertuzumab alone, with ET added for HR+ tumors (Group B). After two cycles, 18F‐FDG‐PET scans were performed. Group A patients completed six cycles regardless of PET results. In Group B, PET responders continued the same regimen, while non‐responders switched to chemotherapy plus trastuzumab/pertuzumab. Among PET responders in Group B, 37.9% achieved pCR, potentially sparing chemotherapy [111]. Once optimized, this 18F‐FDG‐PET‐based pathological response‐adapted technique could tailor treatment by identifying patients who may not need chemotherapy and predicting the timing of treatment failure. PET/CT is limited by sensitivity for micrometastases and high costs, restricting its widespread use. To address these limitations, dedicated breast PET (dbPET) scanners have been developed, offering more sensitive measurement and treatment response stratification [112]. A pilot study highlighted dbPET's potential to complement MRI during early NAT, supporting precision treatment strategies [102]. Combined Biomarkers Analysis Recognizing that single biomarkers are insufficient, recent initiatives integrate multiple data modalities to refine de‐escalation. The I‐SPY 2.2 trial was a sequential adaptive trial designed to optimize tailored treatment, maximize survival, and minimize unnecessary toxicity. Utilizing a sequential multiple assignment randomized trial (SMART) design, it allows multiple serial randomizations based on a treatment response with a platform approach. The subtype‐specific algorithm, “predicted residual cancer burden (pre‐RCB)”, integrates MRI FTV and tumor biopsy data to predict the probability of pCR, advancing personalized treatment [113]. The BMMR2 challenge aimed to develop multiparametric MRI‐based predictive models for early pCR prediction in women undergoing NAT for invasive BC. The dataset included imaging modalities (diffusion‐weighted imaging [DWI], dynamic contrast‐enhanced [DCE] imaging, and T2‐weighted imaging) and clinical variables (age, race, HR and HER2 status, tumor size, lesion type, and grade). The challenge produced imaging‐based models with superior performance in predicting NAT responses, supporting functional breast MRI as an early treatment response indicator for critical treatment decisions [114]. Conssclusion Over the past two decades, remarkable progress has transformed the therapeutic landscape for HER2+ BC. This evolution emphasizes tailoring treatment to optimize the risk–benefit ratio of NAT, with increasing focus on response‐oriented strategies, including escalation and de‐escalation. Patients achieving pCR may be eligible for de‐escalated regimens to reduce toxicity, while those with residual disease may require additional or experimental therapies, aligning with the broader shift toward precision medicine. De‐escalation in HER2+ EBC includes shortened treatment duration, reduced chemotherapy intensity, limited targeted therapy, or adding endocrine and immune therapies. Advances in biomarker research have deepened understanding of molecular characteristics, prognosis, and therapeutic sensitivity, making biomarker analysis essential for personalizing therapies. Real‐time biomarkers monitoring in the neoadjuvant setting help identify patients who may safely omit chemotherapy. Liquid biopsies and imaging modalities may also alter pCR interpretation and open new possibilities for trial design. Agile, biomarker‐driven, and patient‐centered clinical trials are essential to evaluate de‐escalated strategies, ensuring tailored regimens while conserving resources and enhancing clinical outcomes. Author Contributions Yiyin Ma: writing – original draft. Dedian Chen: writing – review and editing. Sheng Huang: writing – review and editing. Funding This work was supported by First‐Class Discipline Team of Kunming Medical University (Kunming Medical University Breast Cancer Precision & Translational Medicine Research Team 2024XKTDYS08) and the National Natural Scientific Foundation of China (82403353 to S.H. and 82160532 to S.H.). Conflicts of Interest The authors declare no conflicts of interest.