Mindpeak Breast HER2

When a compatible HER2 image is opened, analysis runs automatically across the entire slide—no region drawing, no manual counting. The algorithm detects invasive tumor, excludes ductal carcinoma in situ, and classifies every tumor cell on every tissue fragment, returning a color-coded overlay within seconds. Because coverage is complete rather than a hand-picked field, the result does not depend on where a researcher happens to look-one of the best-documented sources of variability in HER2 scoring. At low magnification the overlay shows immediately how HER 2 expression is distributed across the slide and how much scoreable tumor it contains. Because the score is derived from the same cells displayed in the overlay, the overlay also provides a clear audit trail.

Region-level interrogation of the whole slide analysis covers the whole slide, but the 0-click analysis covers the whole slide, but the output is not locked to it. Using the standard annotation tools, subregions can be drawn on top of the completed analysis and scored individually, each returning its own HER2 score and per-class cell counts. This supports two outputs that whole slide scoring alone cannot provide. Tissue can be excluded from consideration, such as a second core, a poorly fixed fragment, crush artifact, or an area the researcher judges unrepresentative. HER2 expression can also be compared between regions, making intratumoral heterogeneity measurable rather than impressionistic: per core, per tumor focus, or across any regions of interest the study design calls for.

At the cellular level, HER2 scoring depends on researcher judgments that are genuinely difficult: whether membrane staining is complete or incomplete, faint or weak, and whether a barely perceptible rim counts at all. These are the interpretations that separate HER2 0 from HER2-low and ultralow.

The same field, with every membrane assessed. Mindpeak Breast HER2 delineates each tumor cell and classifies its HER2 membrane staining, distinguishing negative cells, faint incomplete staining, weak-to-moderate complete staining, and intense staining—applying the same criterion to every cell on the slide rather than to a sampled subset.

Results are returned as a HER2 score, together with the per-class cell counts and percentages that produced it, making the basis for the score visible rather than inferred. The panel also records the provenance of the analysis, including the AI scenario used, the processor version, and the laboratory identifier. This information is important because the staining-intensity threshold is calibrated to each laboratory's own staining during onboarding. Individual markup classes can be shown or hidden while reviewing.