Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2025 Aug 11;15:29289. doi: 10.1038/s41598-025-13820-8

AI microscope facilitates accurate interpretation of HER2 immunohistochemical scores 0 and 1+ in invasive breast cancer

Zong La 1,2,3, Jie Chen 2, Xunxi Lu 1,2, Chuanfen Lei 1,2, Fengling Li 1,2, Lin Zhao 2, Yuhao Yi 1,2,✉
PMCID: PMC12339971  PMID: 40789886

Abstract

Accurate interpretation of human epidermal growth factor receptor 2 (HER2) immunohistochemistry (IHC) scores 0 and 1+ is crucial for treating HER2-low breast cancer patients with antibody–drug conjugates. To improve diagnostic precision, we developed models using 698 retrospectively collected HER2 IHC slides of breast cancer and tested them on an additional 501 slides reviewed by one junior and three senior pathologists. The artificial intelligence (AI)-based models included an invasive breast cancer (IBC) region segmentation model (Model I) and a nuclei detection model (Model II). Model I achieved mean intersection over union (MIoU) scores of 0.879 and 0.880 at 20× and 40× magnifications, and Model II’s F1-scores were 0.866 and 0.878. The proposed AI microscope based on Models I and II achieved F1 scores of 0.878 and 0.906 and accuracies of 0.856 and 0.890 for interpreting IHC scores of 0 and 1+ at 20× and 40×, respectively, which was superior to that of a junior pathologist with an F1 score of 0.871 and an accuracy of 0.848. Additionally, the AI microscope showed high consistency with the interpretation results from the senior pathologists, reaching kappa values of 0.703 at 20× and 0.774 at 40×. This AI microscope has the potential to enhance the interpretation accuracy of HER2 IHC score in clinical settings.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-13820-8.

Keywords: Breast cancer, Human epidermal growth factor receptor 2, Immunohistochemistry, Artificial intelligence

Subject terms: Breast cancer, Image processing

Introduction

In breast cancer, human epidermal growth factor receptor 2 (HER2) levels can identify patients who may benefit from anti-HER2 therapy, so accurate evaluation of HER2 expression levels is required1. In the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) HER2 testing guidelines, HER2 immunohistochemistry (IHC) scores of 0 and 1+ are both regarded as HER2-negative in breast cancer2,3. Recently, the concept of HER2-low status has gained increasing emphasis in breast cancer, and HER2 expression is classified into three levels: HER2-positive (IHC 3+ and 2+ /amplified fluorescence in situ hybridization [FISH +]), HER2-low (IHC 1+ and 2+ /non-amplified FISH [FISH−]), and HER2-negative (IHC 0)4–6. Antibody drug conjugates (ADCs) showed encouraging responses in breast cancer patients with HER2-low expression, especially the third-generation drug DS8201 (trastuzumab deruxtecan)7–11. The latest study found that HER2-low breast cancer has unique clinicopathological and prognostic characteristics. Therefore, accurate interpretation of HER2-IHC scores is particularly important, especially in differentiating between scores 0 and 1+ 12,13.

The HER2 testing guidelines include detailed instructions for distinguishing IHC scores 0 and 1+. However, since scores 0 and 1+ are both classified as HER2-negative in the current testing guidelines, pathologists may not strictly distinguish between scores 0 and 1+ in daily practice, resulting in inconsistencies or inaccurate interpretations. One study found that the false-negative results due to inaccurate detection of HER2-IHC scores 0 and 1+ caused approximately 2.27% of breast cancer patients to miss their opportunity for targeted therapy every year, indicating the need to re-interpret existing IHC scores 0 and 1+ results or to add FISH testing14. One study analyzed the differences in HER2 evaluation between local and central laboratories. The results showed that the disagreement rate between IHC scores 0 and 1+ was the highest, approaching 50%, although the overall consistency was kappa = 0.7915. A phase 1b study on HER2-low treatment reported that the consistency rates between local and central laboratories in assigning IHC scores of 1+ and 2+ were 70% and 40%, respectively16. Another study reported that 85% (87/102) of IHC 0 slides evaluated by local laboratories were re-evaluated as IHC 1+ or 2+ (false negative) by the central laboratory17. IHC scores 0 and 1+ correspond to two subtypes with different DNA, RNA, and protein levels15,18, as well as prognoses19–21. A recent authoritative multicenter analysis showed that 19% of cases read by laboratories generated results with less than or equal to 70% concordance for IHC 0 vs 1+ . When 18 pathologists interpreted the scanned slides, only 26% concordance was noted between scores 0 and 1+ compared with 58% concordance between scores 2+ and 3+. This inaccuracy in the real world may lead to misallocation of therapy with trastuzumab deruxtecan (T-DXd) for many patients22.

IHC scores of 0 and 1+ correspond to two subtypes with low reproducibility of interpretations, differences in molecular characteristics, and prognosis. Scores 0 and 1+ must be accurately distinguished to ensure that patients do not miss the opportunity for combined targeted therapy. This study aims to construct an artificial intelligence (AI) microscope-assisted model for accurate interpretation and differentiation of HER2 IHC scores 0 and 1+, thus contributing to appropriate diagnosis and treatment options for breast cancer with HER2-low expression.

Materials and methods

Data collection

This study was approved by the Ethics Committee on Biomedical Research, West China Hospital of Sichuan University (No. 20220764). The HER2 IHC detection was performed using the VENTANA anti-HER2/neu (4B5) rabbit monoclonal primary antibody (Roche Diagnostics GmbH). Data included in this study were obtained from the cases of invasive breast cancer (IBC) at West China Hospital of Sichuan University, whose pathological diagnosis reports and HER2 IHC slides were collected retrospectively.

A total of 698 HER2 IHC slides with different expression levels of HER2 from January 2017 to December 2017 were collected to develop the AI microscope for IBC region segmentation and nuclei detection. Moreover, 544 HER2 IHC 0 and 1+ slides from January 2019 to December 2019 were collected to test the interpretation performance of the AI microscope. For the test set, the inclusion criteria were breast cancer with HER2 IHC 0 and 1+, and the exclusion criteria were HER2 IHC 2+ and 3+ , carcinoma in situ, and slides with poor quality. A junior pathologist screened the slides first, and then three senior pathologists reinterpreted them according to the 2023 HER2 testing guidelines2,3. The gold standard is established as follows: each slide is initially evaluated by two senior pathologists. If their interpretations match, this consensus becomes the gold standard. If differ, a third senior pathologist makes the final determination. Finally, the test set included 501 IBC slides of HER2 IHC 0 and 1+, and the screening process was shown in Supplementary Fig. 1. We divided 501 cases into 5 subsets for cross-validation. The first four subsets each contained 100 randomly selected cases, and the last subset contained the remaining 101 cases. In the i-th subset of cross-validation, we used the data not in the i-th subset for threshold selection and reported the results on the i-th subset.

Data preprocessing

The 698 HER2 IHC slides used for training the IBC region segmentation model (Model I) and nucleus detection model (Model II) were labeled by a pathologist. The IBC regions were labeled using the open-source graphical image annotation tool Labelme (version number v3.19.0). A closed polygon was drawn to outline the IBC region based on the pre-segmentation results, and the outline was as close as possible to the outer edge of the IBC region. We used the Mark Point tool (version number v1.0.0.3) and developed a method to label the nuclei dots. Dot labeling was used to modify each nucleus in the field of view based on the pre-detection results, making the labeled point as close as possible to the nucleus centroid.

IBC region segmentation model (model I)

A bilateral segmentation network (BiSeNet v2)23 was used to train an IBC region segmentation model (Model I). This segmentation model has the advantage of processing low-level network details and high-level semantic classification separately, thereby achieving high-precision, high-efficiency, and real-time semantic segmentation. We iterated each epoch on the training set and evaluated the model’s mean intersection over union (MIoU) on the validation set until the model’s MIoU no longer increased beyond 10 epochs. The model’s best weight on the validation set was defined, and the true performance was evaluated on the test set.

Nuclei detection model (model II)

Nucleus detection is a small and dense target detection task. The target detection algorithm requires annotation of the outer rectangular box of cell nuclei, which causes an immense workload. Traditional saliency detection algorithms rely on a priori knowledge of the data, and cell nuclei tend to overlap and are easily missed. In this study, we used a standard fully convolutional network (FCN) to develop the nucleus segmentation and detection model24-Model II. Data division, iterations at different magnifications, and the prediction evaluation were similar to those of Model I.

Identification of the optimal thresholds 1 and 2

HER2 0 and 1+ interpretation involves the assessment of the membrane staining cell percentage and staining intensity, so determining the appropriate thresholds for the membrane staining percentage (threshold 1, th1) and staining intensity (threshold 2, th2) is critical. The optimal th1 and th2 were determined using the 501 cases with the gold standard from three senior pathologists. First, the tumor cells of these cases were interpreted based on Model I and Model II. Then, the optimal th1 and th2 were identified by five-subsets. The searching range of the mean membrane staining intensity was [0–255] with a step size of 0.1, and that of the proportion of weakly-stained cells was [0–100%] with a step size of 1%. Based on all possible combinations of th1 and th2, the true positive ratio (TPR), false positive ratio (FPR), receiver operating characteristic (ROC) curve, and the area under the ROC curve (AUC) value were investigated to examine the interpretation performances.

HER2 0 and 1+ interpretation by the AI model

To interpret the HER2 staining intensity, we first captured a field of vision with a size of 3008 × 3008 pixels (View I) using the image acquisition device. Then, Model I was used to identify the IBC region S in View I. Model II was used to locate the nucleus centroid in the IBC region S, which produced the coordinates of the tumor cell set D. Further, the Image I’ was obtained by extracting the diaminobenzidine (DAB) channel from View I in RGB color space and normalizing the channel to [0–255]. In Image I’, the regions of interest (ROIs) used to determine the HER2 IHC score consisted of surrounding pixels centered at the cell nuclei of the tumor cell set D (20× : d = 30, 0.5 μm/pixel; 40×: d = 60, 0.25 μm/pixel). The mean value was calculated based on each ROI to determine the cell membrane staining intensity.

The staining intensity of all tumor cells was recorded, and then the tumor cells were classified as unstained or weakly-stained according to the classification th2. HER2 membrane staining intensity of each case was interpreted based on the results of different fields of vision of the entire slide. The proportion of weakly-stained tumor cells was calculated and compared with the th1. If it was greater than the th1, the case was interpreted as HER2 IHC 1+; otherwise, the case was interpreted as HER2 IHC 0.

HER2 IHC evaluation by pathologists

According to the knowledge gained from published studies25–27 and clinical workflows, a pathologist first determined the location of the lesion at 4 × , then selected multiple fields of vision with the IBC at 10 × , and finally interpreted five fields at 20 × and 10 fields at 40×. For some slides, the IBC regions were not sufficiently large to generate enough fields, so the IBC regions were completely collected. The microscope was Nikon (ECLIPSE Ci-L), and the microscope camera system was constructed using TOUPCAM (E3ISPM09000KPB).

Statistical analysis

The performance of Model I for IBC region segmentation was evaluated using MIoU, which is a standard metric for semantic segmentation, representing the ratio of intersection and union. The MIoU is the average of the intersection to union ratio of each class in the dataset. The performances of the nuclei detection and AI microscope classification were evaluated using the F1-score. Consistency analysis was assessed using the Cohen’s kappa coefficient test28. All analyses were performed using SPSS Statistics (version 26, IBM Corporation) and GraphPad Prism 8.0 (GraphPad Software, Inc.). AI algorithms were conducted using Python 3.8 and Scikit-learn (version 0.23.2).

Results

Performance of model I and model II

The process for staining classification by the AI models was shown in Fig. 1. To train the AI models, we divided the data of 698 HER2 IHC slides into the training set, validation set, and test set at a ratio of 3:1:1. As shown in Supplementary Table 1, 28,199 patches (1024 × 1024 pixels) at 20 × magnification and 38,086 patches (1024 × 1024 pixels) at 40× magnification were identified. The labeling interface is shown in Supplementary y Fig. 2. The representative images of the Model I and II was shown in Supplementary Fig. 3. Model I achieved MIoU values of 0.879 and 0.880 in the test set, at 20 × and 40×, respectively; Model II showed F1-scores of 0.866 and 0.878 in the test set at 20× and 40×, respectively. The detailed results are shown in Table 1. The performance of the two models in the test set showed their capability to accurately identify IBC regions and detect the nuclei of the cells.

Fig. 1.

Fig. 1

Open in a new tab

HER2 IHC scores 0 and 1+ classification system.

Table 1.

Metrics results of IBC region segmentation and nuclei detection.

Models Metrics Magnification 20× Magnification 40×
IBC segmentation mIOU 0.879 0.880
Accuracy 0.937 0.937
Kappa 0.872 0.872
Nuclear detection F1-score 0.866 0.878
Recall 0.863 0.852
Precision 0.870 0.905
Open in a new tab

AI microscope interpretation of HER2 IHC 0 and 1+

The representative images of AI microscope in IBC region segmentation and cell nuclei detection are shown in Fig. 2. We evaluated the classification performance of AI microscope model on five subsets. The average F1-scores for the five subsets at magnifications of 20× and 40× were 0.879 (95%CI: 0.864–0.894 ) and 0.906 (95%CI: 0.886–0.926), respectively. The F1-scores and other performance metrics for each subset were provided in Supplementary Table 2. Since subset 3 achieved the highest F1-score (0.890) at the 20× magnification, it was selected as the dataset for subsequent analysis. The strategy for threshold determination was to calculate the effect of changing th2 under different th1 levels on the classification results. As shown in Fig. 3, within subset 3, we fixed th1, traversed th2, and evaluated the TPR and FPR under each threshold combination. Then we plotted ROC curves and calculated AUC values. The ROC curves for the remaining four subsets were shown in Supplementary Fig. 4A–H. At a fixed th1 of 5%, the average AUC across the five subsets at 20× magnification was 0.930 (95%CI: 0.926–0.934); at a fixed th1 of 2%, the average AUC across the five subsets at 40× magnification was 0.953 (95%CI: 0.951–0.955). Detailed AUC results under different thresholds were provided in Supplementary Table 3.

Fig. 2.

Fig. 2

Open in a new tab

An example of the segmentation of IBC regions and nucleus detection on HER2 IHC scores 0 and 1+ images using the AI microscope. A1, B1, C1, and D1 are the images of HER2 IHC slides that were interpreted as 0 or 1+ by the pathologist at 20× and 40×; A2, B2, C2, and D2 are the corresponding predicted probability maps of the IBC area. The more the color shifts toward red, the higher the probability that this region belonging to IBC, while blue means that the probability of this region belonging to IBC is zero; A3, B3, C3, and D3 denote the segmentation masks of the corresponding IBC region prediction probability map after binarization, where black is the background indicating non-IBC regions and white is the IBC region; A4, B4, C4, and D4 are the corresponding nucleus detection probability maps, where white dots represent the probability that a given pixel can be characterized as the nucleus centroid and black regions represent that pixels are not the nucleus centroid; A5, B5, C5, and D5 are the renderings of the predicted results of the integrated IBC region segmentation and nucleus detection, where red is the nucleus centroid of the tumor cell marked in the original image.

Fig. 3.

Fig. 3

Open in a new tab

The performances of the thresholds in subset 3. Magnifications of 20× (A) and 40× (B). The red line showed the best AUC, and the green line showed the lowest AUC.

Interpretation analysis of HER2 IHC 0 and 1+ in the test set

The AI microscope interpretation performances at 20× and 40× were higher than that of the junior pathologist (F1-scores of 20×: 0.878, 40×: 0.906, junior pathologist: 0.871). The detailed results are shown in Table 2. As shown in Table 3, among the 501 IBC cases, 211 cases were diagnosed as IHC 0, and 290 cases were diagnosed as IHC 1+ by the junior pathologist. Compared with the gold standard, the junior pathologist interpreted 42 cases of IHC 1+ as IHC 0 and 34 cases of IHC 0 as IHC 1+. AI microscope achieved better interpretation at 20× and 40× (kappa = 0.703 and 0.774, P < 0.001) than the junior pathologist (kappa = 0.687, P < 0.001). In both tables, 95% CIs are provided for evaluating AI microscope predictions, obtained through bootstrapping with 10,000 replications across all cases.

Table 2.

The classification performances of different methods for interpreting HER2 0 and 1+

Metrics Junior pathologist AI 20× AI 40×
TN 169 169 180
FP 34 34 23
FN 42 38 32
TP 256 260 266
Recall 0.859

0.872

(0.833, 0.909)

0.893

(0.856, 0.927)
Precision 0.883

0.884

(0.847, 0.920)

0.920

(0.888, 0.950)
F1-score 0.871

0.878

(0.849, 0.905)

0.906

(0.880, 0.930)
TPR 0.859

0.872

(0.833, 0.909)

0.893

(0.856, 0.927)
FPR 0.168

0.167

(0.117, 0.222)

0.113

(0.072, 0.159)
Accuracy 0.848

0.856

(0.824, 0.886)

0.890

(0.862, 0.916)
Open in a new tab

95% CIs are provided for evaluation metrics of AI predictions.

Table 3.

The consistency of different methods with the gold standard.

Gold standard Kappa P
HER2-IHC 1 HER2-IHC 0 Total
298 203 501
Junior HER2-IHC 1 256 34 290 0.687  < 0.001
HER2-IHC 0 42 169 211
AI 20× HER2-IHC 1 260 34 294

0.703

(0.637, 0.765)

  < 0.001
HER2-IHC 0 38 169 207
AI 40× HER2-IHC 1 266 23 289

0.774

(0.715, 0.827)

  < 0.001
HER2-IHC 0 32 180 212
Open in a new tab

Discussion

Using AI technology to help interpret HER2 IHC staining has been studied in recent years. Many studies have focused on whole slide image (WSI)-based AI interpretation29–36. WSI-based AI analysis relies on the premise that digital scanning of slides has been completed. However, since slide digitization is not part of the ordinary clinical workflow, it remains a significant obstacle for widespread application in clinical practice.

In recent years, AI microscope for helping predict molecular marker expression and diagnose lesions has emerged in a human–machine collaborative manner37–40. AI models can help us quickly complete IBC region segmentation, nucleus detection, and cell membrane staining intensity grading by eliminating interfering factors (edge staining and carcinoma in situ), thereby achieving accurate and rapid interpretation of HER2 scores. Zhang et al. used AI microscope to perform classification-assisted interpretation of 285 breast cancer HER2 IHC slides, including negative, equivocal, and positive cases (0/1+, 2+, and 3+, respectively), and the results showed that the accuracy rate reached 95%25. Yue et al. compared different interpretation methods on 50 IBC slides, and the results showed that the consistency of AI microscope-assisted interpretation (kappa = 0.735) was higher than that of manual interpretation (kappa = 0.715). A survey on the acceptance of AI microscope–assisted interpretation among pathologists showed that the acceptance rate of AI microscope–assisted interpretation by pathologists was 90%. However, the study did not include IHC 0 cases27. In summary, currently published AI microscope-assisted interpretation studies have not specifically addressed the problem of accurate classification of HER2-low and HER2 0.

We developed an AI microscope-assisted interpretation model and performed cross-validation on an independent test set of HER2-IHC 0 and 1+ cases. The model showed excellent classification performance, and the best F1-score was 0.906, the highest interpretation value was 0.890, and the consistency with the gold standard was also high (kappa = 0.774). The AI microscope-assisted interpretation performance suggested that 40× magnification was better than 20× magnification (F1-score 0.906 vs. 0.878), and the accuracy of interpretation was also better at 40× magnification than at 20× magnification (accuracy 0.890 vs. 0.856). The AI microscope-assisted interpretation model proposed here achieved more accurate, objective, and repeatable interpretation. The results show that the interpretation results at 40× magnification are better than those at 20× magnification, which is mainly because the AI microscope can detect the cell nuclei more accurately and capture the subtle features of tumor cell membrane staining more clearly at high magnification. The results suggested that in clinical practice, pathologists should also observe cell membrane morphology and staining characteristics under high magnification for interpretation. We provided clinical pathologists with a way to help them interpret slides under different magnifications.

The integrated IBC region segmentation and nuclei detection model proposed in this study can accurately capture all tumor cells in the selected field of vision and is not affected by interstitial cells, such as lymphocytes. It avoids the problem of missing some tumor cell nuclei when only tumor cell nuclei are labeled but not non-tumor cell nuclei, and lays a good foundation for subsequent grading of cell membrane staining. According to the latest breast cancer HER2 testing guidelines, the distinction between IHC scores 0 and 1+ focuses on the staining intensity and the number of stained cell membranes rather than the integrity of the membrane. In the final staining intensity grading algorithm, we did not consider the staining integrity of the cell membrane. Instead, the normalized mean value of the DAB channels was calculated by directly cutting the ROIs at the nucleus centroid to characterize the staining intensity of the cell membrane. The interpretation of HER2-IHC scores 0 and 1+ involves assessing the percentage of stained membrane cells (th1) and the intensity of staining (th2). When determining the best threshold, interestingly, the AUC did not reach its highest value when the th1 of the percentage of tumor cells with weak membrane staining was set at 10%. Under different magnifications, the highest AUC in each subset was found when the above percentage was between 1 and 5%, and the AUC values tended to monotonically decrease with increasing thresholds above 5%, which may suggest that we can use this method of traversing the threshold combinations to determine the optimal interpretation threshold.

AI microscope-assisted interpretation does not require full digital scanning of the slides in advance. Pathologists can select the field of view during the interpretation process to call the AI model for auxiliary interpretation and obtain AI interpretation results within 1 s, and the results are presented in both quantitative and qualitative forms. Our interpretation method does not break up a doctor’s existing clinical workflow while significantly reducing subjective factors in the doctor’s image reading. It saves time and is highly practical in clinical applications. Next, we will carry out more extensive real-world clinical research and offer more reliable and conclusive practical evidence to demonstrate the advantages of this model so as to alleviate the overloaded work schedule of pathologists, promoting wide acceptance of using AI microscope-assisted interpretation among pathologists. The human–machine collaboration model not only improves cost-effectiveness but also offers significant application value for junior pathologists who lack experience. We focused on the current clinical challenges and the newly proposed concept of HER2-low. Combining clinical workflow with AI microscope, we achieve accurate interpretation of IHC scores 0 and 1+, which leads to improvement in the classification and interpretation system of HER2 expression to achieve more accurate stratified diagnosis and treatment.

Our study also has shortcomings in that we did not extend the method from single-center to multicenter studies. In addition, the field of vision of microscopic interpretation is not as large as that of WSI, and the selection of fields is somewhat subjective. We are currently mining larger datasets and more clinical-pathological information to further study AI-assisted prediction of HER2 scores.

Conclusion

In this study, we constructed AI microscope-assisted interpretation models, including a cancer region segmentation model, a tumor cell nuclei detection model, and a membrane staining intensity grading model. The results suggested that the AI microscope interpretation performance was robust, and the accuracy and consistency of HER2-IHC score 0 and 1+ interpretation at 20× and 40× magnifications were higher than those of the original pathological diagnosis. The segmentation detection and interpretation performance of AI microscope at 40× were also found to be better than those at 20×. To conclude, AI microscope interpretation can facilitate accurate classification when screening patients with low HER2 expression for targeted therapy, thus contributing to clinical treatment decisions.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (2.8MB, docx)

Author contributions

Y.Y. is the corresponding authors. Y.Y. supervised the study designation, data acquisition, implementation, analysis, and manuscript edits. Z.L. made contributions to the study designation, data acquisition, sample screening, pathological factors evaluation, statistic analysis, and manuscript drafting. J.C. made contributions to study designation, image processing, AI-based approach designation and implementation, figure creation, statistic analysis, and manuscript edits. X.X. and C.F. made contributions to data acquisition, pathological factors evaluations. F.L. contributed to the partial the discussion of study designation and manuscript edits. L.Z. made contributions to the data acquisition, sample screening. All authors read and approved the final manuscript.

Funding

This work was supported by the National Key Research and Development Program (2017YFC0113908); the Technological Innovation Project of Chengdu New Industrial Technology Research Institute (2017-CY02–00026-GX); the 1·3·5 project for disciplines of excellence (ZYGD18012).

Data availability

The data that support the findings of this study are available on request from the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

Our study was approved by the ethical committee of West China Hospital, Sichuan University (No.20220764), and abided by the Declaration of Helsinki before using tissue samples for scientific research purposes only.

Consent to participate

The written informed consent was waived by the ethical committee for this retrospective study.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Zong La and Jie Chen: Contributed equally to this work.

References

  • 1.Ahn, S. K. & Jung, S. Y. Current biomarkers for precision medicine in breast cancer. Adv. Exp. Med. Biol.1187, 363–379. 10.1007/978-981-32-9620-6_18 (2021). [DOI] [PubMed] [Google Scholar]
  • 2.Wolff, A. C. et al. human epidermal growth factor receptor 2 testing in breast cancer: American society of clinical oncology/college of American pathologists clinical practice guideline focused update. J. Clin. Oncol.: Off. J. Am. Soc. Clin. Oncol.36, 2105–2122. 10.1200/jco.2018.77.8738 (2018). [DOI] [PubMed] [Google Scholar]
  • 3.Wolff, A. C. et al. Human Epidermal growth factor receptor 2 testing in breast cancer: ASCO-college of american pathologists guideline update. J. Clin. Oncol.41, 3867–3872. 10.1200/jco.22.02864 (2023). [DOI] [PubMed] [Google Scholar]
  • 4.Tarantino, P. et al. HER2-low breast cancer: Pathological and clinical landscape. J. Clin. Oncol.: Off. J. Am. Soc. Clin. Oncol.38, 1951–1962. 10.1200/jco.19.02488 (2020). [DOI] [PubMed] [Google Scholar]
  • 5.Marchiò, C. et al. Evolving concepts in HER2 evaluation in breast cancer: Heterogeneity, HER2-low carcinomas and beyond. Semin. Cancer Biol.72, 123–135. 10.1016/j.semcancer.2020.02.016 (2021). [DOI] [PubMed] [Google Scholar]
  • 6.Tarantino, P. et al. ESMO expert consensus statements (ECS) on the definition, diagnosis, and management of HER2-low breast cancer. Ann Oncol34, 645–659. 10.1016/j.annonc.2023.05.008 (2023). [DOI] [PubMed] [Google Scholar]
  • 7.Clifton, G. T. et al. Results of a Randomized Phase IIb Trial of Nelipepimut-S + Trastuzumab versus Trastuzumab to prevent recurrences in patients with high-risk HER2 low-expressing breast cancer. Clin. Cancer Res.: Off. J. Am. Assoc. Cancer Res.26, 2515–2523. 10.1158/1078-0432.Ccr-19-2741 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Dekker, T. J. A. HER2-targeted therapies in HER2-low-expressing breast cancer. J. Clin. Oncol.: Off. J. Am. Soc. Clin. Oncol.38, 3350–3351. 10.1200/jco.20.00657 (2020). [DOI] [PubMed] [Google Scholar]
  • 9.Kostova, V., Désos, P., Starck, J. B. & Kotschy, A. The Chemistry Behind ADCs. Pharmaceuticals10.3390/ph14050442 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Andrikopoulou, A. et al. Trastuzumab Deruxtecan (DS-8201a): The latest research and advances in breast cancer. Clin. Breast Cancer21, e212–e219. 10.1016/j.clbc.2020.08.006 (2021). [DOI] [PubMed] [Google Scholar]
  • 11.Nicolò, E., Zagami, P. & Curigliano, G. Antibody-drug conjugates in breast cancer: The chemotherapy of the future?. Curr. Opin. Oncol.32, 494–502. 10.1097/cco.0000000000000656 (2020). [DOI] [PubMed] [Google Scholar]
  • 12.Denkert, C. et al. Clinical and molecular characteristics of HER2-low-positive breast cancer: Pooled analysis of individual patient data from four prospective, neoadjuvant clinical trials. Lancet Oncol.22, 1151–1161. 10.1016/s1470-2045(21)00301-6 (2021). [DOI] [PubMed] [Google Scholar]
  • 13.Modi, S. et al. Trastuzumab Deruxtecan in PREVIOUSLY TREATED HER2-low advanced breast cancer. New Engl. J. Med.10.1056/NEJMoa2203690 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Garrison, L. P. Jr. et al. Assessing the potential cost-effectiveness of retesting IHC0, IHC1+, or FISH-negative early stage breast cancer patients for HER2 status. Cancer119, 3113–3122. 10.1002/cncr.28196 (2013). [DOI] [PubMed] [Google Scholar]
  • 15.Schettini, F. et al. Clinical, pathological, and PAM50 gene expression features of HER2-low breast cancer. NPJ Breast Cancer7, 1. 10.1038/s41523-020-00208-2 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Modi, S. et al. Antitumor activity and safety of Trastuzumab Deruxtecan in patients with HER2-low-expressing advanced breast cancer: Results from a phase Ib study. J. Clin. Oncol.:Off. J. Am. Soc. Clin. Oncol.38, 1887–1896. 10.1200/jco.19.02318 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Lambein, K. et al. Distinguishing score 0 from score 1+ in HER2 immunohistochemistry-negative breast cancer: Clinical and pathobiological relevance. Am. J. Clin. Pathol.140, 561–566. 10.1309/ajcp4a7ktayhzsoe (2013). [DOI] [PubMed] [Google Scholar]
  • 18.Nakagawa, M. et al. Among triple-negative breast cancers, HER2(0) breast cancer shows a strong tendency to be basal-like compared with HER2(1+) breast cancer: Preliminary results. Breast Cancer (Tokyo, Japan)19, 54–59. 10.1007/s12282-011-0265-6 (2012). [DOI] [PubMed] [Google Scholar]
  • 19.Dehghani, M. et al. The effects of low HER2/neu expression on the clinicopathological characteristics of triple-negative breast cancer patients. Asian Pac. J. Cancer Prev.: APJCP21, 3027–3032. 10.31557/apjcp.2020.21.10.3027 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Gilcrease, M. Z. et al. Even low-level HER2 expression may be associated with worse outcome in node-positive breast cancer. Am. J. Surg. Pathol.33, 759–767. 10.1097/PAS.0b013e31819437f9 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Rossi, V. et al. Moderate immunohistochemical expression of HER-2 (2+) without HER-2 gene amplification is a negative prognostic factor in early breast cancer. Oncologist17, 1418–1425. 10.1634/theoncologist.2012-0194 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Fernandez, A. I. et al. Examination of low ERBB2 protein expression in breast cancer tissue. JAMA Oncol.10.1001/jamaoncol.2021.7239 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Yu, C. et al. BiSeNet V2: Bilateral network with guided aggregation for real-time semantic segmentation. Int. J. Comput. Vis.129, 3051–3068. 10.1007/s11263-021-01515-2 (2021). [Google Scholar]
  • 24.Shelhamer, E., Long, J. & Darrell, T. Fully convolutional networks for semantic segmentation. IEEE Trans. Pattern Anal. Mach. Intell.39, 640–651. 10.1109/tpami.2016.2572683 (2017). [DOI] [PubMed] [Google Scholar]
  • 25.Zhang, J. et al. Microscope based HER2 scoring system. (2020).
  • 26.Chow, Z. L. et al. Counting mitoses with digital pathology in breast phyllodes tumors. Arch. Pathol. Lab. Med.144, 1397–1400. 10.5858/arpa.2019-0435-OA (2020). [DOI] [PubMed] [Google Scholar]
  • 27.Yue, M. et al. Can AI-assisted microscope facilitate breast HER2 interpretation? A multi-institutional ring study. Virchows Archiv.: Int. J. Pathol.479, 443–449. 10.1007/s00428-021-03154-x (2021). [DOI] [PubMed] [Google Scholar]
  • 28.McHugh, M. L. Interrater reliability: The kappa statistic. Biochemia medica22, 276–282 (2012). [PMC free article] [PubMed] [Google Scholar]
  • 29.Masmoudi, H., Hewitt, S. M., Petrick, N., Myers, K. J. & Gavrielides, M. A. Automated quantitative assessment of HER-2/neu immunohistochemical expression in breast cancer. IEEE Trans. Med. Imag.28, 916–925. 10.1109/tmi.2009.2012901 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Nassar, A. et al. Reading immunohistochemical slides on a computer monitor–a multisite performance study using 180 HER2-stained breast carcinomas. Appl. Immunohistochem. Mol. Morphol.: AIMM19, 212–217. 10.1097/PAI.0b013e3181f5e84d (2011). [DOI] [PubMed] [Google Scholar]
  • 31.Brügmann, A. et al. Digital image analysis of membrane connectivity is a robust measure of HER2 immunostains. Breast Cancer Res. Treat.132, 41–49. 10.1007/s10549-011-1514-2 (2012). [DOI] [PubMed] [Google Scholar]
  • 32.Qaiser, T. et al. HER2 challenge contest: a detailed assessment of automated HER2 scoring algorithms in whole slide images of breast cancer tissues. Histopathology72, 227–238. 10.1111/his.13333 (2018). [DOI] [PubMed] [Google Scholar]
  • 33.Qaiser, T. & Rajpoot, N. M. Learning where to see: A novel attention model for automated immunohistochemical scoring. IEEE trans. Med. Imag.38, 2620–2631. 10.1109/tmi.2019.2907049 (2019). [DOI] [PubMed] [Google Scholar]
  • 34.Tewary, S. & Mukhopadhyay, S. HER2 molecular marker scoring using transfer learning and decision level fusion. J. Digit. Imag.34, 667–677. 10.1007/s10278-021-00442-5 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Bannier, P. A. et al. Development of a deep-learning model tailored for HER2 detection in breast cancer to aid pathologists in interpreting HER2-low cases. Histopathology85, 478–488. 10.1111/his.15274 (2024). [DOI] [PubMed] [Google Scholar]
  • 36.Jung, M. et al. Augmented interpretation of HER2, ER, and PR in breast cancer by artificial intelligence analyzer: enhancing interobserver agreement through a reader study of 201 cases. Breast Cancer Res.26, 31. 10.1186/s13058-024-01784-y (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Tang, H. P. et al. Cervical cytology screening facilitated by an artificial intelligence microscope: A preliminary study. Cancer Cytopathol.129, 693–700. 10.1002/cncy.22425 (2021). [DOI] [PubMed] [Google Scholar]
  • 38.Cai, L. et al. Improving Ki67 assessment concordance by the use of an artificial intelligence-empowered microscope: A multi-institutional ring study. Histopathology79, 544–555. 10.1111/his.14383 (2021). [DOI] [PubMed] [Google Scholar]
  • 39.Wang, X. et al. How can artificial intelligence models assist PD-L1 expression scoring in breast cancer: Results of multi-institutional ring studies. NPJ Breast Cancer7, 61. 10.1038/s41523-021-00268-y (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Chen, P. C. et al. An augmented reality microscope with real-time artificial intelligence integration for cancer diagnosis. Nat. Med.25, 1453–1457. 10.1038/s41591-019-0539-7 (2019). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (2.8MB, docx)

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author.

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES