Subtype-guided treatment was reported with a higher rate of pathologic complete response, or pCR, in a 230-patient cohort of HR+/HER2− breast cancer, but the overall pCR rate remained modest. The study reported a pCR rate of 11.3% for subtype-directed precision treatment, compared with 4.5% for traditional PCb treatment, the control regimen of nab-paclitaxel plus carboplatin. The comparison came from FASCINATE-N, a randomized, open-label trial that assigned patients to subtype-specific therapy or standard treatment. The analysis then asked what separated unusually good responses from resistance across the four subtypes.
How the study sorted response
The analysis drew on 230 patients who had data in at least one molecular or tissue dimension. The subtype groups contained 42 SNF1, 48 SNF2, 80 SNF3 and 60 SNF4 patients. The resource combined bulk RNA sequencing, targeted sequencing, whole-genome sequencing and 202 histopathology whole-slide images, so patients did not all contribute the same kind of data. The trial used an adaptive design that allowed independent analysis of subtype arms after preset success or termination rules.
The primary endpoint was pCR, defined as ypT0/Tis ypN0 at surgery. Exceptional responders were divided into three categories: well responders who achieved pCR, poor responders with a Miller-Payne grade of 2 or below, and extreme poor responders with that grade plus stable or progressive disease during neoadjuvant treatment. This gave the researchers a way to compare both unusually strong responses and clear treatment resistance.
Different subtypes pointed to different biology
In the SNF1 precision arm, 10 patients achieved a partial response and 15 were classified as exceptional responders, including 3 poor responders and 12 extreme poor responders. In the poor-efficacy cases, neuron-related pathways, especially GABA-receptor pathways, were upregulated. The authors interpreted that pattern as a possible nerve-related, PNI-linked signal of endocrine resistance. The finding was associational, so it does not show that nerve-related biology caused treatment failure.
An animal experiment offered a separate test of that idea. In BALB/c nude MCF7 xenografts, tumor growth was reported as more strongly suppressed when RTX was added to tamoxifen and a CDK4/6 inhibitor than with tamoxifen and a CDK4/6 inhibitor alone. That result is evidence from a mouse model, not a human treatment result.
SNF3 showed a different pattern. Among 60 precision-arm patients, 8 were well responders, 37 were partial responders, 7 were poor responders and 8 were extreme poor responders. Favorable response to CDK4/6 and PARP inhibitors was associated with higher baseline HRD scores, a measure of DNA-repair weakness, and higher CDK4/6-sensitivity scores, along with greater activation of BER and cell-cycle programs. The study did not report effect sizes or confidence intervals for these associations.
One exceptional responder, identified as STE068, had an exceptionally high HRD score and mutations in four DNA-repair genes: BRCA1, ATR, NSD2 and XRCC5. Three of those genes were HRR-related. That is a single-patient observation, so it cannot establish predictive value or explain why the treatment worked.
SNF4's pattern was centered on the tumor's surrounding tissue. In SNF4 tumors, pCR was associated with a higher proportion of immune cells and more immune-tumor interactions, while non-pCR tumors showed stromal barriers separating immune cells from tumor cells. In SNF2 tumors, lymphatic-vessel generation and general immune activation were correlated with immunotherapy efficacy. The associations therefore varied by subtype, from nerve-related signals to DNA-repair patterns and tissue organization.
A set of clues, not a finished test
The study's biomarker findings are exploratory, and several signals came from small responder groups or a single exceptional responder. The analyses did not establish that PNI or GABA blockade, stroma-targeting therapy, PARP and CDK4/6 combinations, or lymphatic-vessel features used to select immunotherapy improve human outcomes. They also did not validate any proposed biomarker as a clinical predictive test. The headline comparison was based on pCR at surgery, so it should not be read as proof of long-term benefit.
Paper data and sources
Original title: Multimodal assessment of exceptional responders refines subtyping-directed precision treatment in HR+/HER2– breast cancer
Authors: Xiao Y, Yang F, Qi Y et al.
Journal/Repository: The Journal of clinical investigation
Status: Peer-reviewed
First online: 2026-10-19
DOI: Not available
Original paper · Full text