Driving · 2026-11-02

A small HER2 reader for routine H&E slides: real-world deployment

Thang Tran (CloudKites AI Lab, New South Wales, Australia), Lan Dang (Monash Business School, Monash University, Victoria, Australia)

arXiv:2610.40662PDF

AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.

TL;DR

The authors build an end-to-end HER2 reader for routine H&E slides: a 22M-parameter ViT-S/16 tile encoder trained on public TCGA slides via self-supervision, distillation from H-optimus-0, and two JEPA refinement rounds, with five cross-validated attention-based slide classifiers. They report strong external AUCs on Yale-HER2 and other hospitals and provide a deployed CPU research-use pre-read with public weights and per-slide predictions.

Why it matters

HER2 therapy decisions depend on IHC and, for equivocal IHC 2+ cases, ISH. The authors propose an H&E-based probability and call to prioritize ISH for equivocal cases. They evaluate on multiple external hospital cohorts (including HEROHE’s IHC 2+ cases) and emphasize deployment without a GPU, plus publicly released model artifacts for reproducibility and research use.

Method

  • Train a ViT-S/16 (22M) tile encoder on public TCGA H&E by self-supervised pretraining, distillation from H-optimus-0, and two JEPA refinement rounds; no Yale-HER2 slides enter training and no stage uses HER2 labels.
  • Use CLAM-SB attention-based slide classifiers in five cross-validation folds; average probabilities (not logits), and fix an operating point on training data to produce a proposed call with attention heatmaps.
  • Deploy a single verified model pack for CPU pre-read in a pathology platform; weights, per-slide predictions, and the data behind every figure are public.
Abstract (from arXiv)

HER2 status decides who receives HER2-targeted therapy. Pathologists score HER2 protein by immunohistochemistry (IHC); equivocal IHC 2+ tumours, their hardest calls, wait for in situ hybridisation (ISH). We built a HER2 reader for routine H&E slides, owned end to end, as a baseline for pathologists. A 22M-parameter ViT-S/16 tile encoder was trained on public TCGA slides by self-supervision, distillation from H-optimus-0 and two rounds of JEPA refinement; five cross-validated attention-based slide classifiers average their probabilities; and an operating point is fixed on training data. Compared under one paired ten-seed protocol, readers on our encoder rank first of 7 on external validation from TCGA-BRCA to Yale-HER2 (AUC 0.812), level with readers on encoders 4 and 50 times larger. We also rebuilt PathBench's encoder benchmark on public cohorts; our replica reproduces its published order of reference encoders, and under it our encoder ranks first of 7 on TCGA-BRCA (0.689) and sits level with H-optimus-0 and Phikon on Yale-HER2 (0.897). On three external hospitals our deployed reader reaches AUC 0.82 on Yale-HER2 (about 0.81 with training slides balanced for tissue size), 0.75 on HEROHE and 0.63 on BCNB biopsies, and among HEROHE's 85 IHC 2+ cases it ranks ISH-amplified above non-amplified tumours at 0.76, best among our TCGA-trained readers. Per-laboratory score normalisation and a laboratory's own slides improve it further on a new site. Without a GPU our reader needs 5-6 s per slide. It serves as a research-use pre-read in a pathology platform, where it proposes a probability and a call beside IHC and helps prioritise ISH for equivocal cases. Weights, per-slide predictions and the data behind every figure are public at https://huggingface.co/cloudkites/her2-reader.

Related papers

Publication record

Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.

First public release: Date not supplied · Indexed by Pub-AI: 2026-11-02 UTC

preprint

A small HER2 reader for routine H&E slides: real-world deployment

Thang Tran, Lan Dang

Source date: Date not supplied · Original record ↗

arxiv
2610.40662

    Imported from the existing registry. No new publication milestone asserted.

    Abstract · arxiv

    HER2 status decides who receives HER2-targeted therapy. Pathologists score HER2 protein by immunohistochemistry (IHC); equivocal IHC 2+ tumours, their hardest calls, wait for in situ hybridisation (ISH). We built a HER2 reader for routine H&E slides, owned end to end, as a baseline for pathologists. A 22M-parameter ViT-S/16 tile encoder was trained on public TCGA slides by self-supervision, distillation from H-optimus-0 and two rounds of JEPA refinement; five cross-validated attention-based slide classifiers average their probabilities; and an operating point is fixed on training data. Compared under one paired ten-seed protocol, readers on our encoder rank first of 7 on external validation from TCGA-BRCA to Yale-HER2 (AUC 0.812), level with readers on encoders 4 and 50 times larger. We also rebuilt PathBench's encoder benchmark on public cohorts; our replica reproduces its published order of reference encoders, and under it our encoder ranks first of 7 on TCGA-BRCA (0.689) and sits level with H-optimus-0 and Phikon on Yale-HER2 (0.897). On three external hospitals our deployed reader reaches AUC 0.82 on Yale-HER2 (about 0.81 with training slides balanced for tissue size), 0.75 on HEROHE and 0.63 on BCNB biopsies, and among HEROHE's 85 IHC 2+ cases it ranks ISH-amplified above non-amplified tumours at 0.76, best among our TCGA-trained readers. Per-laboratory score normalisation and a laboratory's own slides improve it further on a new site. Without a GPU our reader needs 5-6 s per slide. It serves as a research-use pre-read in a pathology platform, where it proposes a probability and a call beside IHC and helps prioritise ISH for equivocal cases. Weights, per-slide predictions and the data behind every figure are public at https://huggingface.co/cloudkites/her2-reader.