Loading
Life sciences · Journal article
Scientific Reports · October 12, 2026
No summary has been generated for this record yet. What follows is drawn from its source metadata only.
Journal article.
No findings were extractable from the material analysed.
Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.
The source did not state who this applies to in practice.
Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.
This record has not been graded across any dimension yet. Treat the label above as provisional and read the source.
What is missing. This record has no bottom line, key findings, reported figures, evidence dimensions. That is a gap in the analysis, not a judgement about the study.
Accurate assessment of estrogen receptor (ER) status is critical in breast cancer diagnosis and treatment planning, as it directly determines eligibility for endocrine therapy. ER expression is routinely evaluated on immunohistochemistry (IHC) slides using the semi-quantitative Allred scoring system, which combines the proportion of ER-positive tumor nuclei with staining intensity into a composite score of 0–8. However, manual Allred scoring is labor-intensive, subjective, and prone to inter- and intra-observer variability, particularly in cases that are borderline. In this work, we present a whole-slide image (WSI)-level Allred score estimation pipeline built around MOSAIC, our previously developed multi-scale, orientation-aware nuclei analysis network. The pipeline incorporates a patch-based WSI inference strategy that requires only nucleus-level annotations for training; a centroid-guided instance separation and majority-voting scheme that converts pixel-level segmentation outputs into clinically interpretable proportion scores and intensity scores; and a systematic aggregation protocol that derives the final Allred score without requiring slide-level supervision. MOSAIC integrates multi-scale context extraction to handle nuclear size variability, rotation-enhanced fusion to improve robustness to arbitrary nuclear orientation, and a transformer-inspired multi-scale integration hub to model long-range contextual dependencies across heterogeneous tissue regions. Experimental evaluation on an institutional ER-IHC cohort comprising 37 WSIs from different patients demonstrated strong pixel-level segmentation accuracy, robust instance-level nuclei detection, and improved stability in borderline ER cases. At the WSI level, the pipeline achieved the highest concordance with pathologist-assigned hormonal therapy decisions (89.2%) among all compared approaches, requiring an average of 3.5 minutes to analyze the pathologist-defined tumor regions of each WSI, compared with 11 minutes to 2 hours for the baseline methods evaluated under the present experimental protocol.