Changes in HER2low status in 47 advanced breast carcinoma biopsies, matching surgical specimens, and distant metastases assessed by conventional light microscopy, digital pathology, and artificial intelligence

Changes in HER2low status in 47 advanced breast carcinoma biopsies, matching surgical specimens, and distant metastases assessed by conventional light microscopy, digital pathology, and artificial intelligence | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Changes in HER2low status in 47 advanced breast carcinoma biopsies, matching surgical specimens, and distant metastases assessed by conventional light microscopy, digital pathology, and artificial intelligence Anikó Kovács, Leif Klint, Linderholm Barbro, Toshima Parris This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6557172/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 22 Jul, 2025 Read the published version in Breast Cancer Research and Treatment → Version 1 posted 9 You are reading this latest preprint version Abstract Background HER2-targeted therapies have improved survival in HER2-positive breast cancer, and recent data suggest potential benefits for patients with HER2-low tumors (defined as immunohistochemistry (IHC) 1 + or 2 + and, in situ hybridization (ISH)-negative). HER2-low tumors are heterogenous, spanning the hormone receptor-positive and triple-negative subtypes. Assessing HER2-low status remains challenging, especially across specimen types. Aims This study aims to (1) compare HER2 assessment using conventional microscopy, digital pathology, and an artificial intelligence (AI) model, and (2) investigate changes in HER2-low status between core biopsies, surgical specimens, and metastases. Materials and Methods IHC slides from 47 HER2-low breast carcinomas were analyzed using conventional microscopy, digital pathology, and an AI model developed on Aiforia® Create. HER2 statuses were categorized as low, ultralow, and null. Changes in HER2 expression across specimen types were evaluated using agreement measures and visualization tools. Results The AI model identified more HER2-low and HER2-ultralow cases compared to conventional methods, improving detection accuracy. HER2 expression differed between specimen types, with metastases exhibiting increased HER2 expression compared to surgical specimens and core biopsies. Digital pathology also showed stronger membranous staining and identified more HER2-positive tumor cells than microscopy. Conclusions AI evaluation is a more sensitive method for HER2-low assessment and reveals expression changes across disease progression. These findings emphasize the need for standardized HER2 assessment to ensure accurate therapy eligibility, particularly for novel treatments like Trastuzumab-Deruxtecan. HER2-low breast cancer HER2-ultralow breast cancer HER2-null breast cancer Artificial Intelligence Changes in HER2-low status Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Introduction The availability of HER2-targeted therapies over the past two decades has significantly improved clinical outcomes for patients with HER2-positive breast cancer (BC) 1 . In this context, HER2-positive BC refers to tumors with high levels of HER2 protein expression, as assessed by immunohistochemistry (IHC), and/or HER2 amplification, as assessed by in situ hybridization (ISH) 2 . HER2-low is a new term to describe patient tumors with low HER2 expression and is defined as IHC 1 + or IHC 2 + and ISH-negative. BC patients classified as HER2-low represent a heterogeneous population, including luminal hormone receptor-positive BC and triple-negative BC 3 . According to current HER2 testing guidelines, BC patients with HER2-low disease are categorized as having HER2-negative BC. As traditional HER2-targeted therapies have not shown efficacy in this subpopulation, these therapies are not recommended in clinical guidelines 4 . Trastuzumab deruxtecan (T-Dxd) is a novel HER2-targeted antibody-drug conjugate (ADC) developed to deliver a potent topoisomerase I inhibitor drug to HER2-expressing cancer cells with potentially reduced systemic toxicity. The drug is conjugated to a humanized anti-HER2 antibody via a cleavable, plasma-stable, peptide-based linker. Upon cleavage of the linker by lysosomal cathepsins, which are upregulated in cancer cells, the released drug becomes cell membrane-permeable. In preclinical studies, T-Dxd demonstrated antitumor activity in various tumor types, including those with low HER2 expression 5 . The antitumor effect of T-Dxd in heterogeneous or HER2-low tumors may be related to the bystander effect, the released drug can act on all nearby tumor cells 6 . This effect, combined with the high drug-to-antibody ratio (payload) and the high potency of the drug ensures high cytotoxicity at the tumor site. Two randomized phase 3 studies (DESTINY-Breast04 & DESTINY-Breast06) involving patients with metastatic BC defined as HER2-low showed that T-Dxd was more effective than chemotherapy in prolonging progression free survival 7 , 8 . Additionally, in the DESTINY-Breast06 study, around 17% of patients were classified as having so-called HER2ultra-low expression, representing a HER2-low disease with faint, incomplete membrane staining of 1 + in ≤ 10% of tumor cells. It is important to evaluate how HER2 expression changes in the metastatic setting to assess the need for additional biopsies and to guide treatment decisions with new therapies like T-Dxd. However, assessing HER2-low status by IHC is challenging and further confirmation with molecular assays is currently not available 9 , 10 . In most cases, IHC 1 + and 2 + regions are heterogeneously distributed within invasive breast carcinomas and distant metastases. Assessment is particularly difficult when the numbers of tumor cells is near the 10% cut-off point (for HER2-low versus HER2-ultralow status) or the 1% cut-off point (for HER2-ultralow versus HER2-null status) 11 . This sparked our interest in investigating whether modern techniques, such as artificial intelligence (AI), could be used as a tool to better define HER2 expression according to these subgroups compared to conventional methods, such as conventional microscopy, as well as digital pathology. This study aimed to (1) compare the assessment and outcome of HER2status using conventional microscopy, digital pathology, and AI; and (2) assess changes in HER2-low status in core biopsies, matching surgical specimens, and distant metastases. Materials and methods Patient samples For invasive breast carcinomas with HER2-low status (n = 47), IHC HercepTest slides for core needle biopsies as well as, matching surgical specimens, and distant metastases were reevaluated using three modalities: (1) Conventional microscopy (eyeballing in a light microscope), (2) Visual estimation of the scanned digital image on the screen (eyeballing on the screen), and (3) evaluation with an AI model (Aiforia®) providing an exact percentage for the HER2 score. AI analysis was preceded by deep learning using the pathologist’s (A.K.) annotations of the IHC images, with assistance from Aiforia® staff. The original IHC HercepTest slides (stained routinely between 2013 and 2023 at the Department of Clinical Pathology, Sahlgrenska University Hospital, Gothenburg, Sweden) were retrieved from the archives. The HercepTest slides were not restained. In brief, 4 µm sections were prepared from formalin-fixed paraffin-embedded blocks. After pretreatment using the Dako PTLink system (Dako, Carpinteria, CA, USA), the sections were processed further on an automated DAKO Autostainer platform with HercepTest (Dako, Cat. SK001). For samples with HercepTest scores of 2+, an additional Ventana dual silver in situ hybridization (SISH) test was performed. The original HercepTest scoring was perfomed by six board-certified pathologists. Int the present study, reevaluation of HercepTest scoring was done by one of these pathologists (A.K.). The revised scores were assigned based on the scanned full-faced slides or the digital images. According to the HER2-low concept, the HercepTest scores were 0, 1 + and 2 + without consideration of the staining intensity. No HER2-positive or HER2-amplified cases were included in the study. A three-tiered classification system was then used to determine the HER2 status: (1) a score of 2 + for HER2-low, (2) a score of 1 + for HER2-ultralow, and (3) a score of 0 for HER2null. The original HercepTest slides were first scanned using a NanoZoomer S210/Hamamatsu scanner (Oncotopix® Scan by Visiopharm) at 40x magnification, and subsequently uploaded to the Aiforia® website. In total, 400 estimations and assessments were done in this study ( Fig. 1 ) : 40 core biopsies with 3 modalities = 120 examinations (7 cases were diagnosed by cytology) 47 surgical specimens with 3 modalities = 141 examinations 47 distant metastases with 3 modalities = 139 examinations (one metastasis had only a scanned image, here only eye-balling on the screen was possible). AI model description and development. Our HER2-low model was developed utilizing whole-slide images (WSI) on a cloud-based, supervised deep-learning AI-platform - Aiforia® Create 6.0 (Aiforia Technologies Plc, Helsinki, Finland). The AI model was trained as a set of individual convolutional neural networks (CNNs), called “layers”. Layers were trained independently of each other but built in a “layer-tree” - in a “parent-child” fashion, where each “child” layer was exclusively trained on the areas identified by the “parent” layer. Each individual layer was designed to identify a single or multiple classes. Annotations were provided as training data for each class within a given layer. Our HER2-low model consisted of three layers, with each layer having a specific set of classes (total of six) (Table 1 ). Table 1 Her2 Low model layers and classes. RL – region layer; OL – object layer; FoV – field of view; VC – very complex. AI model Layer AI model class Morphological feature detected Output FoV (µm) Complexity RL 1 Tissue Total breast tissue Tissue area (mm 2 ) 100 VC RL 2 Invasive Epithelium Malignant epithelial cells infiltrating the surrounding breast tissue Invasive epithelium area (mm 2 ) 50 VC OL 1 Her 2_score 0 Malignant epithelial cells without Her2 staining Count - VC Her 2_score 1+ Malignant epithelial cells with faint incomplete Her2 staining Count - VC Her 2_score 2+ Malignant epithelial cells with weak to moderate complete Her2 staining Count - VC Her 2_score 3+ Malignant epithelial cells with strong complete Her2 staining Count - VC This HER2-low model was developed by using 90% of transfer learning from an earlier developed HER2 model, with an additional 78 annotated regions for calibration purposes. For advanced training parameters, a defined field of view and complexity level (determined by the complexity of the feature of interest) was defined for the different layers (Table 1 ). A total of 7000 iterations were executed to train the AI model on three WSI, with an overall training loss of 0.0015. The HER2-low model was batch-analyzed on WSI or regions of interest for the whole set of images used in this study. Statistical analysis A descriptive analysis of HER2 status was conducted across patient samples over time (core biopsy, matching surgical specimen, and matching distant metastasis) and across the three different analysis modalities (conventional microscopy, visual estimation of the scanned image on the screen, and AI), with Sankey diagrams employed for a visual presentation. Agreement in HER2 status between the modalities was evaluated using Gwett's AC1 coefficient and weighted kappa coefficient, and illustrated graphically with bubble plots and Bland-Altman plots with 95% limits of agreement across each sample type (core biopsy, surgical specimen, and distant metastasis). Correlations between methods were evaluated using Spearman’s non-parametric rank correlation coefficient. Statistical analyses were conducted using SAS/STAT® Software, Version 9.4 (SAS Institute Inc., Cary, NC, USA). Results Patients We identified 47 patients diagnosed with invasive BC through core biopsy or cytology, who subsequently underwent surgery with curative intent. Of the 47 patients, 37 (79%) were classified as having hormone receptor-positive BC (luminal), and 10 (21%) as having triple-negative BC (TNBC). Ten patients diagnosed with BC received neoadjuvant chemotherapy before surgery (6 luminal BC and 4 TNBC). All 47 patients experienced a recurrence of their disease in the form of distant metastases. The most common site of metastasis was the liver, followed by the skeleton. In all cases, a biopsy of the metastatic site was performed, and in 2 cases, biopsies were taken from 2 metastatic sites, resulting in a total of 49 biopsies from the metastases. (Fig. 1 ) HER2 status assessment by digital image analysis using AI Digital image analysis (DIA) using AI confirmed the HER2-low status in the majority of cases. Moreover, AI evaluated HER2-ultralow status with confidence (tumor cells with weak membranous staining: > score 0 < score 1+) giving an exact percentage of tumor cells showing score 1+. This was also the case when identifying HER2-null status (score 0 or 1 + in < 1% of tumor cells) or by evaluating the presence of some tumor cells with score 3 + in < 10% within an obvious HER2low tumor. Seven cases needed to be re-evaluated by AI because AI incorrectly included the DCIS component of the carcinoma (1 core biopsy and 6 surgical specimens). Comparison of HER2 status assessment by conventional light microscopy, digital pathology and artificial intelligence Comparison of the three modalities in core biopsies : Digital visual estimation and AI scored the same number of cases as HER2-low status, which was 10% higher compared to conventional microscopy. There was no discrepancy regarding HER2null status among the three modalities. In core biopsies, there was discordance for 5 cases, which were classified as HER2-ultralow by conventional microscopy, but HER2-low by digital visual estimation (13% discordance: 5 of 40 cases). (Table 2 ) The observed agreement was 90% (80%, 100%) and Weighted Kappa 0.9 (0.8, 1.0) for both AI and digital estimation vs conventional microscopy. The observed agreement for visual digital estimation vs AI was 100% (90%, 100%). (Figs. 2 – 3 ) Table 2 Discordant cases among core biopsies Patient No. Eye balling microscope Digital image estimation Eye balling on the screen AI Case 11 HER2ultralow HER2low HER2low Case 19 HER2ultralow HER2low HER2low Case 31 HER2ultralow HER2low HER2low Case 32 HER2ultralow HER2low HER2low Case 37 HER2ultralow HER2low HER2low Comparison of the three modalities in surgical specimens: Digital visual evaluation and AI classified marginally more patient samples as HER2low (51% vs 53% vs 47%) compared to microscopy. AI scored more patients as HER2-ultralow compared to microscopy and digital visual evaluation (28% vs 23% vs 19%). Consequently, more patients with HER2null status were identified using the microscopic evaluation, further highlighting that the digital image shows stronger membranous staining and more cells could be identified with these staining patterns. In surgical specimens, there was discordance in: 10 cases (21% discordance: 10 of 47 cases) (Table 3). The observed agreement was 80% (70%, 90%) and Weighted Kappa was 0.8 (0.7, 0.9) for both AI and digital estimation vs the conventional microscope, the observed agreement digital estimation vs AI was 100% (90%, 100%) and Weighted Kappa 0.9 (0.9, 1.0). (Figs. 4 – 5 ) Table 3 Discordant cases among surgical specimens Comparison of the three modalities in the distant metastases: AI classified more patient samples as HER2low than digital visual evaluation (13% vs 11%). In metastatic specimens, there was discordance in: 7 cases (15% discordance: 7 of 47 cases) (Table 4 ) The observed agreement was 90% (80%, 100%) and Weighted Kappa 0.8 (0.6, 1.0) for digital estimation vs the conventional microscope, whereas the observed agreement was 80% (70%, 90%) and Weighted Kappa 0.6 (0.3, 0.9) for AI vs conventional microscope. Moreover, the observed agreement was 90% (80%, 100%) and Weighted Kappa 0.9 (0.9, 1.0) for visual digital estimation vs AI. (Figs. 6 – 7 ) Table 4 Discordance among distant metastases Patient No. in Excel Table Eye balling microscope Digital image estimation Eye balling on the screen AI Case 3. HER2ultralow HER2low HER2low Case 10. null HER2ultralow HER2ultralow Case 25. HER2ultralow HER2low HER2low Case 28. HER2ultralow HER2low HER2low Case 29. HER2ultralow HER2ultralow HER2low Case 44. HER2ultralow HER2low HER2low Case 47. HER2ultralow HER2low HER2low Comparison of HER2 status in the three sample types (core biopsy, matching surgical specimen, and matching distant metastasis) Comparison of the three sample types using conventional microscopy: Using conventional microscopy, no difference was found in the number of core biopsies and matching distant metastases samples classified as HER2-low or HER2-ultralow in the three patient sample categories. None of the core biopsies were classified as HER2-null. (Fig. 8 ) Comparison of the three sample types using digital visual evaluation: As with conventional microscopy, the number of samples classified as HER2-low using digital visual evaluation was similar for the core biopsies and distant metastasis samples. However, more surgical specimens and distant metastasis samples were classified as HER2-ultralow compared to core biopsies. None of the core biopsy samples were HER2-null, thereby matching the results of conventional microscopy. More HER2-low cases were identified by digital visual evaluation compared to conventional microscope. (Fig. 9 ) Comparison of the three sample types using AI: Consistent with the other modalities, the number of samples classified as HER2-low using AI was similar in the core biopsies and distant metastasis samples. Similar to the other modalities, the number of HER2-ultralow cases was highest in the surgical specimens, followed by the distant metastasis samples, and lowest in the core biopsies. The number of distant metastasis samples with HER2-low status was 25% higher than in surgical specimens. Fewer distant metastases were classified as HER2null compared to core biopsies and surgical specimens. (Fig. 10 ) Association between treatment and changes in HER2 status Ten patients received neoadjuvant treatment, and for most of these patients, the time between the core biopsy and the surgical specimen was relatively short, without any antitumoral treatment administered in between. All patients received adjuvant treatment after surgery, which could have influenced HER2 status at the time of recurrence with metastatic disease. Therefore, we were interested in examining how HER2 status differed between the surgical specimen and the distant metastasis biopsy. We found that the most common change was an increase in HER2 expression (43%), while a smaller proportion of samples had a decrease in HER2 expression (17%). (Table 5 ) Table 5 Changes in HER2-status between the specimens Variable Breast core biopsies (n = 47) Matching operation specimens (n = 47) Their distant metastases (n = 47) Difference Matching operation specimens - Breast core biopsies Difference Their distant metastases - Matching operation specimens AI n/a HER2null HER2ultralow HER2low 7 (14.9%) 0 (0.0%) 1 (2.1%) 39 (83.0%) 0 (0.0%) 10 (21.3%) 13 (27.7%) 24 (51.1%) 1 (2.1%) 4 (8.5%) 6 (12.8%) 36 (76.6%) Decrease17 (36.2%) Equal23 (48.9%) Increase7 (14.9%) Decrease8 (17.0%) Equal19 (40.4%) Increase20 (42.6%) For categorical variables n (%) is presented. Effect size is estimated using the Cohen method (difference in mean/SD at first timepoint). Discussion In this study, we compared three modalities – conventional microscopy, digital pathology, and an AI model – for the assessment of HER2 status and changes in HER2 expression in core biopsies, surgical specimens, and metastasis biopsies from 47 patients with advanced BC. During digital visual evaluation, the scanned images often showed a stronger membranous staining intensity than in imaging by conventional microscopy, which was consistent with previous reports. This led to higher HER2 expression scores on digital visual evaluation compared to conventional microscopy, e.g., HER2-null being scored as HER2-ultra-low and HER2-ultra-low as HER2-low. Furthermore, more tumor cells could be identified with digital visual evaluation, which resulted in a higher percentage of positively stained tumor cells. The AI model classified more patient samples as HER2-ultra-low compared to conventional microscopy and digital visual evaluation, which can be explained by the number of tumor cells with membranous staining near the cutoff value of 10% (for HER2-low versus HER2-ultra-low status) or the cutoff value of 1% (for HER2-ultra-low versus HER2-null status). Therefore, more HER2-low cases were identified digitally and found by the AI than by conventional microscopy. Clinically, when a patient shows radiological signs of recurrence, it is important to know how often HER2 expression changes to a lower level in the metastatic setting, particularly whether a previously HER2-low/ultra-low tumor has converted to HER2-null or 0 9 . This information is crucial in determining whether a biopsy of the metastasis is needed to confirm a prior HER2-low/ultra-low diagnosis and if the location of the metastasis is difficult to biopsy, such as in the case of brain metastases. Moreover, a patient who has undergone surgery for a HER2-low tumor is not offered adjuvant HER2-targeted therapy. However, in cases where the metastases retain HER2-low or even HER2-ultra-low status, targeted treatment with T-Dxd could offer benefits. Using all three methods, we demonstrated, that it was overwhelmingly more common for HER2 expression to be higher in the metastasis compared to the surgical specimens. In only a few cases did a HER2-low/ultralow tumor in the surgical specimen convert to HER2-null or 0, particularly in patients with a prior HER2-low tumor. Discordance between surgical specimens and metastatic biopsies has been described in other studies, which have also reported that HER2 expression most commonly increases. However, these studies compared HER2-low with HER2 0 status and did not consider HER2-ultralow as a variable 9 , 12 , 13 . On the other hand, studies examining HER2-positive and HER2-negative status found that it was more common for the disease to transition from being HER2-positive in the primary tumor to HER2-negative in the metastasis biopsy. This change in status has been described in up to 20% of cases, likely due to adaptation following adjuvant HER2 therapy 14 . This indicates that while the need for a new biopsy of the metastasis exists, in cases where the metastasis is in a location difficult to biopsy, it remains fairly unlikely that the disease has converted to HER2-null or 0. One can contemplate whether the change in HER2 status concerning distant metastases may be related to the smaller biopsy size. When obtaining surgical specimens, there are larger tumor pieces available for HER2 analysis, but for distant metastases, mainly core needle biopsies are taken which represent only a random sample of the entire metastatic mass. However, we did not observe “edge artifact” in the needle biopsies by HER immunostaining. Decalcification of bone metastases did not modify the scores either or did not make the evaluation more difficult. Similarities in HER2low status between core biopsies and distant metastases can be explained that both represent only a small part of the tumor (a random biopsy), not the entire tumor. Our findings raise the question of how digital pathology and AI assessment could change the assessment of HER2 status and eligibility determination for T-Dxd compared to current standard methods. Which modality will be the gold standard? Which one should be taken into account as authoritative? The usage of digital pathology varies greatly across European countries and worldwide. Pathology departments using conventional microscopy may identify fewer advanced BCs with HER2-low status. The relatively low interobservers´ concordance in identifying HER2-ultralow status warrants the need for increased precision in HER2 assessment by digital tools using AI 15 . Recent studies have shown that AI was an accurate method for reducing the number of equivocal cases and did not affect the sensitivity of the assessment 16 – 20 . In our study, AI often included DCIS in the assessment, which demanded revision by excluding the areas of DCIS before the re-run. In subsequent studies, annotations should be performed in order to teach AI to exclude DCIS 21 . A discrepancy in HER2-low status using different antibodies had been reported (Ventana versus Dako´s two antibodies). Ventana´s antibody and Dako´s monoclonal antibody yield higher HER2 staining scores 22 , 23 . Conclusion The HER2low concept causes diagnostic challenges for pathologists. We found that digital pathology with AI evaluation is a more sensitive method for HER2-low assessment of metastatic BC compared to conventional techniques (conventional microscopy and digital visual assessment) and can show HER2 expression changes across disease progression. AI evaluation provided an exact percentage of each HER2 score within seconds with very high accuracy. The Aiforia® solution highlighted specific findings and showed visual feedback, allowing users to check the results down to the pixel level. The findings emphasize the need for standardized HER2 assessment to accurately determine treatment eligibility, particularly for novel therapies like T-Dxd. Further clinical studies are needed to verify the predictive value of the scoring modalities. Declarations Ethics approval and consent to participate This study was approved by the Regional Ethical Committee (registration no. 287-15, and its amendment, 2023-03030-02 for AI usage) in Gothenburg, Sweden. The local ethical committee waived the requirement for informed consent since the patients’ core biopsies, operation specimens, and distant metastases were decoded. Consent for publication Not applicable Availability of supporting data The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. Competing interests The authors declare that they have no competing interests Funding A.K. received research grants from Sahlgrenska Comprehensive Cancer Centre in 2023 to perform the study. Authors´contribution Leif Klint and Anikó Kovács wrote the main manuscript and prepared all the figures and tables. All authors reviewed the manuscript. Acknowledgements We thank representatives of Aiforia® Ms. Rute Pedrosa and Mr. Darshan Kumar for their teaching and assistance with the assessment by AI. We thank Mr. Seyed Abbas Rezazadeh Fakhar and Mr. Bora Todorovic for retrieving the histological slides from the slide archive. References Jørgensen JT. Twenty-five years with HER2 targeted therapy. Ann Transl Med 2024;12(3):53. (In eng). DOI: 10.21037/atm-23-153. Amin MB, Greene FL, Edge SB, et al. The Eighth Edition AJCC Cancer Staging Manual: Continuing to build a bridge from a population-based to a more "personalized" approach to cancer staging. CA: a cancer journal for clinicians 2017;67(2):93-99. (In eng). DOI: 10.3322/caac.21388. Giuliani S, Ciniselli CM, Leonardi E, et al. In a cohort of breast cancer screened patients the proportion of HER2 positive cases is lower than that earlier reported and pathological characteristics differ between HER2 3+ and HER2 2+/Her2 amplified cases. Virchows Arch 2016;469(1):45-50. (In eng). DOI: 10.1007/s00428-016-1940-y. Cardoso F, Senkus E, Costa A, et al. 4th ESO-ESMO International Consensus Guidelines for Advanced Breast Cancer (ABC 4)†. Annals of oncology : official journal of the European Society for Medical Oncology 2018;29(8):1634-1657. (In eng). DOI: 10.1093/annonc/mdy192. Ogitani Y, Aida T, Hagihara K, et al. DS-8201a, A Novel HER2-Targeting ADC with a Novel DNA Topoisomerase I Inhibitor, Demonstrates a Promising Antitumor Efficacy with Differentiation from T-DM1. Clinical cancer research : an official journal of the American Association for Cancer Research 2016;22(20):5097-5108. (In eng). DOI: 10.1158/1078-0432.Ccr-15-2822. Nakada T, Masuda T, Naito H, et al. Novel antibody drug conjugates containing exatecan derivative-based cytotoxic payloads. Bioorg Med Chem Lett 2016;26(6):1542-1545. (In eng). DOI: 10.1016/j.bmcl.2016.02.020. Bardia A, Hu X, Dent R, et al. Trastuzumab Deruxtecan after Endocrine Therapy in Metastatic Breast Cancer. The New England journal of medicine 2024 (In eng). DOI: 10.1056/NEJMoa2407086. Modi S, Jacot W, Yamashita T, et al. Trastuzumab Deruxtecan in Previously Treated HER2-Low Advanced Breast Cancer. The New England journal of medicine 2022;387(1):9-20. (In eng). DOI: 10.1056/NEJMoa2203690. Tarantino P, Gandini S, Nicolò E, et al. Evolution of low HER2 expression between early and advanced-stage breast cancer. European journal of cancer (Oxford, England : 1990) 2022;163:35-43. (In eng). DOI: 10.1016/j.ejca.2021.12.022. Gaudio M, Jacobs F, Benvenuti C, et al. Unveiling the HER2-low phenomenon: exploring immunohistochemistry and gene expression to characterise HR-positive HER2-negative early breast cancer. Breast cancer research and treatment 2024;203(3):487-495. (In eng). DOI: 10.1007/s10549-023-07151-3. Marchiò C, Annaratone L, Marques A, Casorzo L, Berrino E, Sapino A. Evolving concepts in HER2 evaluation in breast cancer: Heterogeneity, HER2-low carcinomas and beyond. Semin Cancer Biol 2021;72:123-135. (In eng). DOI: 10.1016/j.semcancer.2020.02.016. Almstedt K, Krauthauser L, Kappenberg F, et al. Discordance of HER2-Low between Primary Tumors and Matched Distant Metastases in Breast Cancer. Cancers (Basel) 2023;15(5) (In eng). DOI: 10.3390/cancers15051413. Miglietta F, Griguolo G, Bottosso M, et al. Evolution of HER2-low expression from primary to recurrent breast cancer. NPJ Breast Cancer 2021;7(1):137. (In eng). DOI: 10.1038/s41523-021-00343-4. Schrijver W, Suijkerbuijk KPM, van Gils CH, van der Wall E, Moelans CB, van Diest PJ. Receptor Conversion in Distant Breast Cancer Metastases: A Systematic Review and Meta-analysis. J Natl Cancer Inst 2018;110(6):568-580. (In eng). DOI: 10.1093/jnci/djx273. Mehta S, Iyengar A, Barman H, et al. Prevalence of “HER2 ultra-low” among patients with advanced breast cancer with historical IHC0 status. Journal of Clinical Oncology 2024;42(16_suppl):e13156-e13156. DOI: 10.1200/JCO.2024.42.16_suppl.e13156. Jakobsen MR, Teerapakpinyo C, Shuangshoti S, Keelawat S. Comparison between digital image analysis and visual assessment of immunohistochemical HER2 expression in breast cancer. Pathol Res Pract 2018;214(12):2087-2092. (In eng). DOI: 10.1016/j.prp.2018.10.015. Palm C, Connolly CE, Masser R, et al. Determining HER2 Status by Artificial Intelligence: An Investigation of Primary, Metastatic, and HER2 Low Breast Tumors. Diagnostics (Basel) 2023;13(1) (In eng). DOI: 10.3390/diagnostics13010168. Wu S, Yue M, Zhang J, et al. The Role of Artificial Intelligence in Accurate Interpretation of HER2 Immunohistochemical Scores 0 and 1+ in Breast Cancer. Mod Pathol 2023;36(3):100054. (In eng). DOI: 10.1016/j.modpat.2022.100054. Krishnamurthy S, Schnitt SJ, Vincent-Salomon A, et al. Fully Automated Artificial Intelligence Solution for Human Epidermal Growth Factor Receptor 2 Immunohistochemistry Scoring in Breast Cancer: A Multireader Study. JCO Precis Oncol 2024;8:e2400353. (In eng). DOI: 10.1200/po.24.00353. Gustavson M, Haneder S, Spitzmueller A, et al. Abstract PD6-01: Novel approach to HER2 quantification: Digital pathology coupled with AI-based image and data analysis delivers objective and quantitative HER2 expression analysis for enrichment of responders to trastuzumab deruxtecan (T-DXd; DS-8201), specifically in HER2-low patients. Cancer Research 2021;81(4_Supplement):PD6-01-PD6-01. DOI: 10.1158/1538-7445.Sabcs20-pd6-01. Xiong Z, Liu K, Liu S, et al. Precision HER2: a comprehensive AI system for accurate and consistent evaluation of HER2 expression in invasive breast Cancer. BMC cancer 2024;24(1):1204. (In eng). DOI: 10.1186/s12885-024-12980-6. Rüschoff J, Friedrich M, Nagelmeier I, et al. Comparison of HercepTest™ mAb pharmDx (Dako Omnis, GE001) with Ventana PATHWAY anti-HER-2/neu (4B5) in breast cancer: correlation with HER2 amplification and HER2 low status. Virchows Arch 2022;481(5):685-694. (In eng). DOI: 10.1007/s00428-022-03378-5. Zhang H, Katerji H, Turner BM, Hicks DG. HER2-Low Breast Cancers. Am J Clin Pathol 2022;157(3):328-336. DOI: 10.1093/ajcp/aqab117. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 22 Jul, 2025 Read the published version in Breast Cancer Research and Treatment → Version 1 posted Editorial decision: Revision requested 23 May, 2025 Reviews received at journal 22 May, 2025 Reviews received at journal 10 May, 2025 Reviewers agreed at journal 10 May, 2025 Reviewers agreed at journal 06 May, 2025 Reviewers invited by journal 06 May, 2025 Editor assigned by journal 02 May, 2025 Submission checks completed at journal 02 May, 2025 First submitted to journal 29 Apr, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6557172","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":454791966,"identity":"2c15985e-7b91-482e-b77b-50d77c8ae666","order_by":0,"name":"Anikó Kovács","email":"","orcid":"","institution":"University of Gothenburg","correspondingAuthor":false,"prefix":"","firstName":"Anikó","middleName":"","lastName":"Kovács","suffix":""},{"id":454791967,"identity":"a07d2624-a018-4177-a053-77a4bae9a326","order_by":1,"name":"Leif Klint","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBElEQVRIie3SMUvDQBTA8XcELsuLricW8hWuOFihJF/lQqCTH8DB4YKQLtW530Po5PBKIC7B2bGh4JxFiYt6SQtOvWYUvP+ScNyP9wIBcLn+YCfAdAM0QQCP+hN+jHBg2RJIIHCuhhJgHTFvKIctxv27TEMlRpfh4uPs/WkaPfj3xNobC8G1Ia8Cr/JgdT56m6U5vigPKwsRyXwDjUBZBo8XgoqUi2vpsdxCwtpM6QluDfnuCfv8sk1hu8UMYXVDFHUEAm37liTTqurIbLwFShXHShZYHian/nOtm3Iay6LYrFuK4nC+GNft7WHSp/ZPz/wDSbcSHQG/sRYgHnzb5XK5/k0/UoVO3ZEgctgAAAAASUVORK5CYII=","orcid":"","institution":"University of Gothenburg","correspondingAuthor":true,"prefix":"","firstName":"Leif","middleName":"","lastName":"Klint","suffix":""},{"id":454791968,"identity":"ad6b05b5-b844-4561-991e-63b158329dd9","order_by":2,"name":"Linderholm Barbro","email":"","orcid":"","institution":"University of Gothenburg","correspondingAuthor":false,"prefix":"","firstName":"Linderholm","middleName":"","lastName":"Barbro","suffix":""},{"id":454791969,"identity":"84328d6b-2a13-498c-8338-81905eac2720","order_by":3,"name":"Toshima Parris","email":"","orcid":"","institution":"University of Gothenburg","correspondingAuthor":false,"prefix":"","firstName":"Toshima","middleName":"","lastName":"Parris","suffix":""}],"badges":[],"createdAt":"2025-04-29 13:53:29","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6557172/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6557172/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s10549-025-07776-6","type":"published","date":"2025-07-22T15:57:37+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":82584852,"identity":"fea99516-8d74-45e7-8d5e-bcc43c474c9f","added_by":"auto","created_at":"2025-05-13 06:57:50","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":202045,"visible":true,"origin":"","legend":"\u003cp\u003ePatient characteristics\u003c/p\u003e\n\u003cp\u003eThe figure shows the number of samples and their results from the three different sites.\u003c/p\u003e","description":"","filename":"fig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/9bd5c630b507ab91347a355a.jpg"},{"id":82583336,"identity":"a4727610-e777-4c2e-bfa8-1a497631b766","added_by":"auto","created_at":"2025-05-13 06:49:50","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":117980,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of results by the three modalities in the core biopsies\u003c/p\u003e\n\u003cp\u003eThe figure illustrates the concordance between the three different assessment methods in the core biopsies.\u003c/p\u003e","description":"","filename":"fig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/8e4ca7fe17a333813cda3686.jpg"},{"id":82584853,"identity":"d6dbf1df-a322-4191-8e98-b9ca95e78141","added_by":"auto","created_at":"2025-05-13 06:57:50","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":141912,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of results by the three modalities in the core biopsies\u003c/p\u003e\n\u003cp\u003eThe figure illustrates the difference in results between the three different assessment methods in the core biopsies\u003c/p\u003e","description":"","filename":"fig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/4e06e306a0971fde58a8654d.jpg"},{"id":82583337,"identity":"ffd39293-7bf5-440b-b2c7-923f55f5c608","added_by":"auto","created_at":"2025-05-13 06:49:50","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":114125,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of results by the three modalities in the surgical specimens\u003c/p\u003e\n\u003cp\u003eThe figure illustrates the concordance between the three different assessment methods in the surgical specimens\u003c/p\u003e","description":"","filename":"fig4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/67a0ebc53497b996c7bd2962.jpg"},{"id":82583339,"identity":"cec644ff-461f-41f5-a42d-ac2e2e2fb107","added_by":"auto","created_at":"2025-05-13 06:49:50","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":138052,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of results by the three modalities in the surgical specimens\u003c/p\u003e\n\u003cp\u003eThe figure illustrates the difference in results between the three different assessment methods in the surgical specimens\u003c/p\u003e","description":"","filename":"fig5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/95ce91b699af2f7eefbae9cb.jpg"},{"id":82584850,"identity":"bcb4c3a3-2589-4d0e-bfff-c2ab4dd152cb","added_by":"auto","created_at":"2025-05-13 06:57:50","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":111621,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of results by the three modalities in the surgical specimens\u003c/p\u003e\n\u003cp\u003eThe figure illustrates the concordance between the three different assessment methods in the matching distant metastases\u003c/p\u003e","description":"","filename":"fig6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/f343052b203b7177a09f7d74.jpg"},{"id":82583343,"identity":"65786cfa-0014-48d7-82b7-f814b754b909","added_by":"auto","created_at":"2025-05-13 06:49:50","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":141777,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of results by the three modalities in the surgical specimens\u003c/p\u003e\n\u003cp\u003eThe figure illustrates the difference in results between the three different assessment methods in the matching distant metastases\u003c/p\u003e","description":"","filename":"fig7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/c8ce4078b2d8cdd5f951fcff.jpg"},{"id":82583345,"identity":"43a84cb4-a971-48e4-a0f5-3131445ce4be","added_by":"auto","created_at":"2025-05-13 06:49:50","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":163553,"visible":true,"origin":"","legend":"\u003cp\u003eComparing HER2 status in the three patient samples by the microscope\u003c/p\u003e\n\u003cp\u003eThe figure illustrates changes in HER2 status in individual patients between different biopsy sites examined under the microscope.\u003c/p\u003e","description":"","filename":"fig8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/82c66087fd56b4b2da18a1dc.jpg"},{"id":82583349,"identity":"baad1dc1-c97a-4a74-a906-d48f504b5c11","added_by":"auto","created_at":"2025-05-13 06:49:50","extension":"jpg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":161604,"visible":true,"origin":"","legend":"\u003cp\u003eComparing HER2 status in the three patient samples by eye balling\u003c/p\u003e\n\u003cp\u003eThe figure illustrates changes in HER2 status in individual patients between different biopsy sites examined by eye balling.\u003c/p\u003e","description":"","filename":"fig9.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/93dd145272cc7d82ea086033.jpg"},{"id":82583344,"identity":"d957104a-ccf6-4e25-a295-b503f140b1fb","added_by":"auto","created_at":"2025-05-13 06:49:50","extension":"jpg","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":157646,"visible":true,"origin":"","legend":"\u003cp\u003eComparing HER2 status in the three patient samples by AI\u003c/p\u003e\n\u003cp\u003eThe figure illustrates changes in HER2 status in individual patients between different biopsy sites examined by AI.\u003c/p\u003e","description":"","filename":"fig10.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/0366d5e5ad61f749b91337a5.jpg"},{"id":87756696,"identity":"813e79da-bb62-4546-aad8-c7200f481722","added_by":"auto","created_at":"2025-07-28 16:07:49","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2661316,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6557172/v1/149bf743-a3b2-492a-b356-ce11b2075637.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eChanges in HER2low status in 47 advanced breast carcinoma biopsies, matching surgical specimens, and distant metastases assessed by conventional light microscopy, digital pathology, and artificial intelligence\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe availability of HER2-targeted therapies over the past two decades has significantly improved clinical outcomes for patients with HER2-positive breast cancer (BC)\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. In this context, HER2-positive BC refers to tumors with high levels of HER2 protein expression, as assessed by immunohistochemistry (IHC), and/or HER2 amplification, as assessed by in situ hybridization (ISH)\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. HER2-low is a new term to describe patient tumors with low HER2 expression and is defined as IHC 1\u0026thinsp;+\u0026thinsp;or IHC 2\u0026thinsp;+\u0026thinsp;and ISH-negative. BC patients classified as HER2-low represent a heterogeneous population, including luminal hormone receptor-positive BC and triple-negative BC\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. According to current HER2 testing guidelines, BC patients with HER2-low disease are categorized as having HER2-negative BC. As traditional HER2-targeted therapies have not shown efficacy in this subpopulation, these therapies are not recommended in clinical guidelines\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eTrastuzumab deruxtecan (T-Dxd) is a novel HER2-targeted antibody-drug conjugate (ADC) developed to deliver a potent topoisomerase I inhibitor drug to HER2-expressing cancer cells with potentially reduced systemic toxicity. The drug is conjugated to a humanized anti-HER2 antibody via a cleavable, plasma-stable, peptide-based linker. Upon cleavage of the linker by lysosomal cathepsins, which are upregulated in cancer cells, the released drug becomes cell membrane-permeable. In preclinical studies, T-Dxd demonstrated antitumor activity in various tumor types, including those with low HER2 expression\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. The antitumor effect of T-Dxd in heterogeneous or HER2-low tumors may be related to the bystander effect, the released drug can act on all nearby tumor cells\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. This effect, combined with the high drug-to-antibody ratio (payload) and the high potency of the drug ensures high cytotoxicity at the tumor site. Two randomized phase 3 studies (DESTINY-Breast04 \u0026amp; DESTINY-Breast06) involving patients with metastatic BC defined as HER2-low showed that T-Dxd was more effective than chemotherapy in prolonging progression free survival\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Additionally, in the DESTINY-Breast06 study, around 17% of patients were classified as having so-called HER2ultra-low expression, representing a HER2-low disease with faint, incomplete membrane staining of 1\u0026thinsp;+\u0026thinsp;in \u0026le;\u0026thinsp;10% of tumor cells.\u003c/p\u003e \u003cp\u003eIt is important to evaluate how HER2 expression changes in the metastatic setting to assess the need for additional biopsies and to guide treatment decisions with new therapies like T-Dxd. However, assessing HER2-low status by IHC is challenging and further confirmation with molecular assays is currently not available\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. In most cases, IHC 1\u0026thinsp;+\u0026thinsp;and 2\u0026thinsp;+\u0026thinsp;regions are heterogeneously distributed within invasive breast carcinomas and distant metastases. Assessment is particularly difficult when the numbers of tumor cells is near the 10% cut-off point (for HER2-low versus HER2-ultralow status) or the 1% cut-off point (for HER2-ultralow versus HER2-null status)\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. This sparked our interest in investigating whether modern techniques, such as artificial intelligence (AI), could be used as a tool to better define HER2 expression according to these subgroups compared to conventional methods, such as conventional microscopy, as well as digital pathology.\u003c/p\u003e \u003cp\u003eThis study aimed to (1) compare the assessment and outcome of HER2status using conventional microscopy, digital pathology, and AI; and (2) assess changes in HER2-low status in core biopsies, matching surgical specimens, and distant metastases.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatient samples\u003c/h2\u003e \u003cp\u003eFor invasive breast carcinomas with HER2-low status (n\u0026thinsp;=\u0026thinsp;47), IHC HercepTest slides for core needle biopsies as well as, matching surgical specimens, and distant metastases were reevaluated using three modalities: (1) Conventional microscopy (eyeballing in a light microscope), (2) Visual estimation of the scanned digital image on the screen (eyeballing on the screen), and (3) evaluation with an AI model (Aiforia\u0026reg;) providing an exact percentage for the HER2 score. AI analysis was preceded by deep learning using the pathologist\u0026rsquo;s (A.K.) annotations of the IHC images, with assistance from Aiforia\u0026reg; staff.\u003c/p\u003e \u003cp\u003eThe original IHC HercepTest slides (stained routinely between 2013 and 2023 at the Department of Clinical Pathology, Sahlgrenska University Hospital, Gothenburg, Sweden) were retrieved from the archives. The HercepTest slides were not restained. In brief, 4 \u0026micro;m sections were prepared from formalin-fixed paraffin-embedded blocks. After pretreatment using the Dako PTLink system (Dako, Carpinteria, CA, USA), the sections were processed further on an automated DAKO Autostainer platform with HercepTest (Dako, Cat. SK001). For samples with HercepTest scores of 2+, an additional Ventana dual silver in situ hybridization (SISH) test was performed.\u003c/p\u003e \u003cp\u003e The original HercepTest scoring was perfomed by six board-certified pathologists. Int the present study, reevaluation of HercepTest scoring was done by one of these pathologists (A.K.). The revised scores were assigned based on the scanned full-faced slides or the digital images. According to the HER2-low concept, the HercepTest scores were 0, 1\u0026thinsp;+\u0026thinsp;and 2\u0026thinsp;+\u0026thinsp;without consideration of the staining intensity. No HER2-positive or HER2-amplified cases were included in the study. A three-tiered classification system was then used to determine the HER2 status: (1) a score of 2\u0026thinsp;+\u0026thinsp;for HER2-low, (2) a score of 1\u0026thinsp;+\u0026thinsp;for HER2-ultralow, and (3) a score of 0 for HER2null. The original HercepTest slides were first scanned using a NanoZoomer S210/Hamamatsu scanner (Oncotopix\u0026reg; Scan by Visiopharm) at 40x magnification, and subsequently uploaded to the Aiforia\u0026reg; website.\u003c/p\u003e \u003cp\u003eIn total, 400 estimations and assessments were done in this study \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e:\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e40 core biopsies with 3 modalities\u0026thinsp;=\u0026thinsp;\u003cb\u003e120\u003c/b\u003e examinations (7 cases were diagnosed by cytology)\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e47 surgical specimens with 3 modalities\u0026thinsp;=\u0026thinsp;\u003cb\u003e141\u003c/b\u003e examinations\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e47 distant metastases with 3 modalities\u0026thinsp;=\u0026thinsp;\u003cb\u003e139\u003c/b\u003e examinations (one metastasis had only a scanned image, here only eye-balling on the screen was possible).\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eAI model description and development.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eOur HER2-low model was developed utilizing whole-slide images (WSI) on a cloud-based, supervised deep-learning AI-platform - Aiforia\u0026reg; Create 6.0 (Aiforia Technologies Plc, Helsinki, Finland). The AI model was trained as a set of individual convolutional neural networks (CNNs), called \u0026ldquo;layers\u0026rdquo;. Layers were trained independently of each other but built in a \u0026ldquo;layer-tree\u0026rdquo; - in a \u0026ldquo;parent-child\u0026rdquo; fashion, where each \u0026ldquo;child\u0026rdquo; layer was exclusively trained on the areas identified by the \u0026ldquo;parent\u0026rdquo; layer. Each individual layer was designed to identify a single or multiple classes. Annotations were provided as training data for each class within a given layer. Our HER2-low model consisted of three layers, with each layer having a specific set of classes (total of six) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eHer2 Low model layers and classes. RL \u0026ndash; region layer; OL \u0026ndash; object layer; FoV \u0026ndash; field of view; VC \u0026ndash; very complex.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAI model Layer\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAI model class\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMorphological feature detected\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOutput\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eFoV (\u0026micro;m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eComplexity\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRL 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTissue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTotal breast tissue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTissue area (mm\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eVC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRL 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInvasive Epithelium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMalignant epithelial cells infiltrating the surrounding breast tissue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eInvasive epithelium area (mm\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eVC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eOL 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHer 2_score 0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMalignant epithelial cells without Her2 staining\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCount\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eVC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHer 2_score 1+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMalignant epithelial cells with faint incomplete Her2 staining\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCount\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eVC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHer 2_score 2+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMalignant epithelial cells with weak to moderate complete Her2 staining\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCount\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eVC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHer 2_score 3+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMalignant epithelial cells with strong complete Her2 staining\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCount\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eVC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThis HER2-low model was developed by using 90% of transfer learning from an earlier developed HER2 model, with an additional 78 annotated regions for calibration purposes. For advanced training parameters, a defined field of view and complexity level (determined by the complexity of the feature of interest) was defined for the different layers (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). A total of 7000 iterations were executed to train the AI model on three WSI, with an overall training loss of 0.0015. The HER2-low model was batch-analyzed on WSI or regions of interest for the whole set of images used in this study.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eA descriptive analysis of HER2 status was conducted across patient samples over time (core biopsy, matching surgical specimen, and matching distant metastasis) and across the three different analysis modalities (conventional microscopy, visual estimation of the scanned image on the screen, and AI), with Sankey diagrams employed for a visual presentation.\u003c/p\u003e \u003cp\u003eAgreement in HER2 status between the modalities was evaluated using Gwett's AC1 coefficient and weighted kappa coefficient, and illustrated graphically with bubble plots and Bland-Altman plots with 95% limits of agreement across each sample type (core biopsy, surgical specimen, and distant metastasis). Correlations between methods were evaluated using Spearman\u0026rsquo;s non-parametric rank correlation coefficient.\u003c/p\u003e \u003cp\u003eStatistical analyses were conducted using SAS/STAT\u0026reg; Software, Version 9.4 (SAS Institute Inc., Cary, NC, USA).\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003ePatients\u003c/h2\u003e \u003cp\u003eWe identified 47 patients diagnosed with invasive BC through core biopsy or cytology, who subsequently underwent surgery with curative intent. Of the 47 patients, 37 (79%) were classified as having hormone receptor-positive BC (luminal), and 10 (21%) as having triple-negative BC (TNBC). Ten patients diagnosed with BC received neoadjuvant chemotherapy before surgery (6 luminal BC and 4 TNBC). All 47 patients experienced a recurrence of their disease in the form of distant metastases. The most common site of metastasis was the liver, followed by the skeleton. In all cases, a biopsy of the metastatic site was performed, and in 2 cases, biopsies were taken from 2 metastatic sites, resulting in a total of 49 biopsies from the metastases. (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e)\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eHER2 status assessment by digital image analysis using AI\u003c/h3\u003e\n\u003cp\u003eDigital image analysis (DIA) using AI confirmed the HER2-low status in the majority of cases. Moreover, AI evaluated HER2-ultralow status with confidence (tumor cells with weak membranous staining: \u0026gt; score 0\u0026thinsp;\u0026lt;\u0026thinsp;score 1+) giving an exact percentage of tumor cells showing score 1+. This was also the case when identifying HER2-null status (score 0 or 1\u0026thinsp;+\u0026thinsp;in \u0026lt;\u0026thinsp;1% of tumor cells) or by evaluating the presence of some tumor cells with score 3\u0026thinsp;+\u0026thinsp;in \u0026lt;\u0026thinsp;10% within an obvious HER2low tumor.\u003c/p\u003e \u003cp\u003eSeven cases needed to be re-evaluated by AI because AI incorrectly included the DCIS component of the carcinoma (1 core biopsy and 6 surgical specimens).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eComparison of HER2 status assessment by conventional light microscopy, digital pathology and artificial intelligence\u003c/h2\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e\u003cem\u003eComparison of the three modalities in core biopsies\u003c/em\u003e:\u003c/h2\u003e \u003cp\u003eDigital visual estimation and AI scored the same number of cases as HER2-low status, which was 10% higher compared to conventional microscopy.\u003c/p\u003e \u003cp\u003eThere was no discrepancy regarding HER2null status among the three modalities.\u003c/p\u003e \u003cp\u003eIn core biopsies, there was discordance for 5 cases, which were classified as HER2-ultralow by conventional microscopy, but HER2-low by digital visual estimation (13% discordance: 5 of 40 cases). (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) The observed agreement was 90% (80%, 100%) and Weighted Kappa 0.9 (0.8, 1.0) for both AI and digital estimation vs conventional microscopy. The observed agreement for visual digital estimation vs AI was 100% (90%, 100%). (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDiscordant cases among core biopsies\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePatient No.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEye balling microscope\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDigital image estimation\u003c/p\u003e \u003cp\u003eEye balling on the screen\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAI\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCase 11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHER2ultralow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCase 19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHER2ultralow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCase 31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHER2ultralow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCase 32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHER2ultralow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCase 37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHER2ultralow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHER2low\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e\n\u003ch3\u003eComparison of the three modalities in surgical specimens:\u003c/h3\u003e\n\u003cp\u003eDigital visual evaluation and AI classified marginally more patient samples as HER2low (51% vs 53% vs 47%) compared to microscopy. AI scored more patients as HER2-ultralow compared to microscopy and digital visual evaluation (28% vs 23% vs 19%). Consequently, more patients with HER2null status were identified using the microscopic evaluation, further highlighting that the digital image shows stronger membranous staining and more cells could be identified with these staining patterns.\u003c/p\u003e \u003cp\u003eIn surgical specimens, there was discordance in: 10 cases (21% discordance: 10 of 47 cases) (Table\u0026nbsp;3). The observed agreement was 80% (70%, 90%) and Weighted Kappa was 0.8 (0.7, 0.9) for both AI and digital estimation vs the conventional microscope, the observed agreement digital estimation vs AI was 100% (90%, 100%) and Weighted Kappa 0.9 (0.9, 1.0). (Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e\u003cstrong\u003eTable 3\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiscordant cases among surgical specimens\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cimg width=\"605\" height=\"429\" src=\"data:image/emf;base64,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\" alt=\"image\"\u003e\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eComparison of the three modalities in the distant metastases:\u003c/h2\u003e \u003cp\u003eAI classified more patient samples as HER2low than digital visual evaluation (13% vs 11%).\u003c/p\u003e \u003cp\u003eIn metastatic specimens, there was discordance in: 7 cases (15% discordance: 7 of 47 cases) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e4\u003c/span\u003e) The observed agreement was 90% (80%, 100%) and Weighted Kappa 0.8 (0.6, 1.0) for digital estimation vs the conventional microscope, whereas the observed agreement was 80% (70%, 90%) and Weighted Kappa 0.6 (0.3, 0.9) for AI vs conventional microscope. Moreover, the observed agreement was 90% (80%, 100%) and Weighted Kappa 0.9 (0.9, 1.0) for visual digital estimation vs AI. (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e)\u003c/p\u003e \n\u003cp\u003e\u003cstrong\u003eTable 4\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiscordance among distant metastases\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"629\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.2222%;\"\u003e\n \u003cp\u003ePatient No. in Excel Table\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.6825%;\"\u003e\n \u003cp\u003eEye balling microscope\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 30.3175%;\"\u003e\n \u003cp\u003eDigital image estimation\u003c/p\u003e\n \u003cp\u003eEye balling on the screen\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.7778%;\"\u003e\n \u003cp\u003eAI\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.2222%;\"\u003e\n \u003cp\u003eCase 3.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.6825%;\"\u003e\n \u003cp\u003eHER2ultralow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 30.3175%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.7778%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.2222%;\"\u003e\n \u003cp\u003eCase 10.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.6825%;\"\u003e\n \u003cp\u003enull\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 30.3175%;\"\u003e\n \u003cp\u003eHER2ultralow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.7778%;\"\u003e\n \u003cp\u003eHER2ultralow\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.2222%;\"\u003e\n \u003cp\u003eCase 25.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.6825%;\"\u003e\n \u003cp\u003eHER2ultralow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 30.3175%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.7778%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.2222%;\"\u003e\n \u003cp\u003eCase 28.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.6825%;\"\u003e\n \u003cp\u003eHER2ultralow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 30.3175%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.7778%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.2222%;\"\u003e\n \u003cp\u003eCase 29.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.6825%;\"\u003e\n \u003cp\u003eHER2ultralow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 30.3175%;\"\u003e\n \u003cp\u003eHER2ultralow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.7778%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.2222%;\"\u003e\n \u003cp\u003eCase 44.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.6825%;\"\u003e\n \u003cp\u003eHER2ultralow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 30.3175%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.7778%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.2222%;\"\u003e\n \u003cp\u003eCase 47.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.6825%;\"\u003e\n \u003cp\u003eHER2ultralow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 30.3175%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17.7778%;\"\u003e\n \u003cp\u003eHER2low\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\u003cp\u003e \u003cb\u003eComparison of HER2 status in the three sample types\u003c/b\u003e (core biopsy, matching surgical specimen, and matching distant metastasis)\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eComparison of the three sample types using conventional microscopy:\u003c/h2\u003e \u003cp\u003eUsing conventional microscopy, no difference was found in the number of core biopsies and matching distant metastases samples classified as HER2-low or HER2-ultralow in the three patient sample categories.\u003c/p\u003e \u003cp\u003eNone of the core biopsies were classified as HER2-null. (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eComparison of the three sample types using digital visual evaluation:\u003c/h2\u003e \u003cp\u003eAs with conventional microscopy, the number of samples classified as HER2-low using digital visual evaluation was similar for the core biopsies and distant metastasis samples.\u003c/p\u003e \u003cp\u003eHowever, more surgical specimens and distant metastasis samples were classified as HER2-ultralow compared to core biopsies.\u003c/p\u003e \u003cp\u003eNone of the core biopsy samples were HER2-null, thereby matching the results of conventional microscopy.\u003c/p\u003e \u003cp\u003eMore HER2-low cases were identified by digital visual evaluation compared to conventional microscope. (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eComparison of the three sample types using AI:\u003c/h2\u003e \u003cp\u003eConsistent with the other modalities, the number of samples classified as HER2-low using AI was similar in the core biopsies and distant metastasis samples.\u003c/p\u003e \u003cp\u003eSimilar to the other modalities, the number of HER2-ultralow cases was highest in the surgical specimens, followed by the distant metastasis samples, and lowest in the core biopsies.\u003c/p\u003e \u003cp\u003eThe number of distant metastasis samples with HER2-low status was 25% higher than in surgical specimens.\u003c/p\u003e \u003cp\u003eFewer distant metastases were classified as HER2null compared to core biopsies and surgical specimens. (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eAssociation between treatment and changes in HER2 status\u003c/h2\u003e \u003cp\u003eTen patients received neoadjuvant treatment, and for most of these patients, the time between the core biopsy and the surgical specimen was relatively short, without any antitumoral treatment administered in between. All patients received adjuvant treatment after surgery, which could have influenced HER2 status at the time of recurrence with metastatic disease. Therefore, we were interested in examining how HER2 status differed between the surgical specimen and the distant metastasis biopsy. We found that the most common change was an increase in HER2 expression (43%), while a smaller proportion of samples had a decrease in HER2 expression (17%). (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e5\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eChanges in HER2-status between the specimens\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBreast core biopsies\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;47)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMatching operation specimens\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;47)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTheir distant metastases\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;47)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDifference\u003c/p\u003e \u003cp\u003eMatching operation specimens - Breast core biopsies\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eDifference\u003c/p\u003e \u003cp\u003eTheir distant metastases - Matching operation specimens\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAI\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e n/a \u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e HER2null \u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e HER2ultralow\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e HER2low\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7 (14.9%)\u003c/p\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003cp\u003e1 (2.1%)\u003c/p\u003e \u003cp\u003e39 (83.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003cp\u003e10 (21.3%)\u003c/p\u003e \u003cp\u003e13 (27.7%)\u003c/p\u003e \u003cp\u003e24 (51.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (2.1%)\u003c/p\u003e \u003cp\u003e4 (8.5%)\u003c/p\u003e \u003cp\u003e6 (12.8%)\u003c/p\u003e \u003cp\u003e36 (76.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDecrease17 (36.2%)\u003c/p\u003e \u003cp\u003eEqual23 (48.9%)\u003c/p\u003e \u003cp\u003eIncrease7 (14.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eDecrease8 (17.0%)\u003c/p\u003e \u003cp\u003eEqual19 (40.4%)\u003c/p\u003e \u003cp\u003eIncrease20 (42.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e \u003cp\u003eFor\u0026nbsp;categorical\u0026nbsp;variables\u0026nbsp;n\u0026nbsp;(%)\u0026nbsp;is\u0026nbsp;presented.\u003c/p\u003e \u003cp\u003eEffect\u0026nbsp;size\u0026nbsp;is\u0026nbsp;estimated\u0026nbsp;using\u0026nbsp;the\u0026nbsp;Cohen\u0026nbsp;method\u0026nbsp;(difference\u0026nbsp;in\u0026nbsp;mean/SD\u0026nbsp;at\u0026nbsp;first\u0026nbsp;timepoint).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we compared three modalities \u0026ndash; conventional microscopy, digital pathology, and an AI model \u0026ndash; for the assessment of HER2 status and changes in HER2 expression in core biopsies, surgical specimens, and metastasis biopsies from 47 patients with advanced BC. During digital visual evaluation, the scanned images often showed a stronger membranous staining intensity than in imaging by conventional microscopy, which was consistent with previous reports. This led to higher HER2 expression scores on digital visual evaluation compared to conventional microscopy, e.g., HER2-null being scored as HER2-ultra-low and HER2-ultra-low as HER2-low. Furthermore, more tumor cells could be identified with digital visual evaluation, which resulted in a higher percentage of positively stained tumor cells. The AI model classified more patient samples as HER2-ultra-low compared to conventional microscopy and digital visual evaluation, which can be explained by the number of tumor cells with membranous staining near the cutoff value of 10% (for HER2-low versus HER2-ultra-low status) or the cutoff value of 1% (for HER2-ultra-low versus HER2-null status).\u003c/p\u003e \u003cp\u003eTherefore, more HER2-low cases were identified digitally and found by the AI than by conventional microscopy.\u003c/p\u003e \u003cp\u003eClinically, when a patient shows radiological signs of recurrence, it is important to know how often HER2 expression changes to a lower level in the metastatic setting, particularly whether a previously HER2-low/ultra-low tumor has converted to HER2-null or 0\u003csup\u003e9\u003c/sup\u003e. This information is crucial in determining whether a biopsy of the metastasis is needed to confirm a prior HER2-low/ultra-low diagnosis and if the location of the metastasis is difficult to biopsy, such as in the case of brain metastases. Moreover, a patient who has undergone surgery for a HER2-low tumor is not offered adjuvant HER2-targeted therapy. However, in cases where the metastases retain HER2-low or even HER2-ultra-low status, targeted treatment with T-Dxd could offer benefits.\u003c/p\u003e \u003cp\u003eUsing all three methods, we demonstrated, that it was overwhelmingly more common for HER2 expression to be higher in the metastasis compared to the surgical specimens. In only a few cases did a HER2-low/ultralow tumor in the surgical specimen convert to HER2-null or 0, particularly in patients with a prior HER2-low tumor. Discordance between surgical specimens and metastatic biopsies has been described in other studies, which have also reported that HER2 expression most commonly increases. However, these studies compared HER2-low with HER2 0 status and did not consider HER2-ultralow as a variable\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. On the other hand, studies examining HER2-positive and HER2-negative status found that it was more common for the disease to transition from being HER2-positive in the primary tumor to HER2-negative in the metastasis biopsy. This change in status has been described in up to 20% of cases, likely due to adaptation following adjuvant HER2 therapy\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. This indicates that while the need for a new biopsy of the metastasis exists, in cases where the metastasis is in a location difficult to biopsy, it remains fairly unlikely that the disease has converted to HER2-null or 0. One can contemplate whether the change in HER2 status concerning distant metastases may be related to the smaller biopsy size. When obtaining surgical specimens, there are larger tumor pieces available for HER2 analysis, but for distant metastases, mainly core needle biopsies are taken which represent only a random sample of the entire metastatic mass. However, we did not observe \u0026ldquo;edge artifact\u0026rdquo; in the needle biopsies by HER immunostaining. Decalcification of bone metastases did not modify the scores either or did not make the evaluation more difficult. Similarities in HER2low status between core biopsies and distant metastases can be explained that both represent only a small part of the tumor (a random biopsy), not the entire tumor.\u003c/p\u003e \u003cp\u003eOur findings raise the question of how digital pathology and AI assessment could change the assessment of HER2 status and eligibility determination for T-Dxd compared to current standard methods. Which modality will be the gold standard? Which one should be taken into account as authoritative? The usage of digital pathology varies greatly across European countries and worldwide. Pathology departments using conventional microscopy may identify fewer advanced BCs with HER2-low status. The relatively low interobservers\u0026acute; concordance in identifying HER2-ultralow status warrants the need for increased precision in HER2 assessment by digital tools using AI\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. Recent studies have shown that AI was an accurate method for reducing the number of equivocal cases and did not affect the sensitivity of the assessment\u003csup\u003e\u003cspan additionalcitationids=\"CR17 CR18 CR19\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. In our study, AI often included DCIS in the assessment, which demanded revision by excluding the areas of DCIS before the re-run. In subsequent studies, annotations should be performed in order to teach AI to exclude DCIS\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. A discrepancy in HER2-low status using different antibodies had been reported (Ventana versus Dako\u0026acute;s two antibodies). Ventana\u0026acute;s antibody and Dako\u0026acute;s monoclonal antibody yield higher HER2 staining scores\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThe HER2low concept causes diagnostic challenges for pathologists. We found that digital pathology with AI evaluation is a more sensitive method for HER2-low assessment of metastatic BC compared to conventional techniques (conventional microscopy and digital visual assessment) and can show HER2 expression changes across disease progression. AI evaluation provided an exact percentage of each HER2 score within seconds with very high accuracy. The Aiforia\u0026reg; solution highlighted specific findings and showed visual feedback, allowing users to check the results down to the pixel level. The findings emphasize the need for standardized HER2 assessment to accurately determine treatment eligibility, particularly for novel therapies like T-Dxd. Further clinical studies are needed to verify the predictive value of the scoring modalities.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cu\u003eEthics approval and consent to participate \u003c/u\u003eThis study was approved by the Regional Ethical Committee (registration no. 287-15, and its amendment, 2023-03030-02 for AI usage) in Gothenburg, Sweden. The local ethical committee waived the requirement for informed consent since the patients\u0026rsquo; core biopsies, operation specimens, and distant metastases were decoded. \u003c/p\u003e\n\u003cp\u003e\u003cu\u003eConsent for publication\u003c/u\u003e Not applicable\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eAvailability of supporting data\u003c/u\u003e The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eCompeting interests\u003c/u\u003e The authors declare that they have no competing interests\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eFunding\u003c/u\u003e A.K. received research grants from Sahlgrenska Comprehensive Cancer Centre in 2023 to perform the study.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eAuthors\u0026acute;contribution\u003c/u\u003e Leif Klint and Anik\u0026oacute; Kov\u0026aacute;cs wrote the main manuscript and prepared all the figures and tables. All authors reviewed the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eAcknowledgements\u003c/u\u003e We thank representatives of Aiforia\u0026reg; Ms. Rute Pedrosa and Mr. Darshan Kumar for their teaching and assistance with the assessment by AI. We thank Mr. Seyed Abbas Rezazadeh Fakhar and Mr. Bora Todorovic for retrieving the histological slides from the slide archive. \u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eJ\u0026oslash;rgensen JT. Twenty-five years with HER2 targeted therapy. Ann Transl Med 2024;12(3):53. (In eng). DOI: 10.21037/atm-23-153.\u003c/li\u003e\n\u003cli\u003eAmin MB, Greene FL, Edge SB, et al. The Eighth Edition AJCC Cancer Staging Manual: Continuing to build a bridge from a population-based to a more \u0026quot;personalized\u0026quot; approach to cancer staging. CA: a cancer journal for clinicians 2017;67(2):93-99. (In eng). DOI: 10.3322/caac.21388.\u003c/li\u003e\n\u003cli\u003eGiuliani S, Ciniselli CM, Leonardi E, et al. In a cohort of breast cancer screened patients the proportion of HER2 positive cases is lower than that earlier reported and pathological characteristics differ between HER2 3+ and HER2 2+/Her2 amplified cases. Virchows Arch 2016;469(1):45-50. (In eng). DOI: 10.1007/s00428-016-1940-y.\u003c/li\u003e\n\u003cli\u003eCardoso F, Senkus E, Costa A, et al. 4th ESO-ESMO International Consensus Guidelines for Advanced Breast Cancer (ABC 4)\u0026dagger;. Annals of oncology : official journal of the European Society for Medical Oncology 2018;29(8):1634-1657. (In eng). DOI: 10.1093/annonc/mdy192.\u003c/li\u003e\n\u003cli\u003eOgitani Y, Aida T, Hagihara K, et al. DS-8201a, A Novel HER2-Targeting ADC with a Novel DNA Topoisomerase I Inhibitor, Demonstrates a Promising Antitumor Efficacy with Differentiation from T-DM1. Clinical cancer research : an official journal of the American Association for Cancer Research 2016;22(20):5097-5108. (In eng). DOI: 10.1158/1078-0432.Ccr-15-2822.\u003c/li\u003e\n\u003cli\u003eNakada T, Masuda T, Naito H, et al. Novel antibody drug conjugates containing exatecan derivative-based cytotoxic payloads. Bioorg Med Chem Lett 2016;26(6):1542-1545. (In eng). DOI: 10.1016/j.bmcl.2016.02.020.\u003c/li\u003e\n\u003cli\u003eBardia A, Hu X, Dent R, et al. Trastuzumab Deruxtecan after Endocrine Therapy in Metastatic Breast Cancer. The New England journal of medicine 2024 (In eng). DOI: 10.1056/NEJMoa2407086.\u003c/li\u003e\n\u003cli\u003eModi S, Jacot W, Yamashita T, et al. Trastuzumab Deruxtecan in Previously Treated HER2-Low Advanced Breast Cancer. The New England journal of medicine 2022;387(1):9-20. (In eng). DOI: 10.1056/NEJMoa2203690.\u003c/li\u003e\n\u003cli\u003eTarantino P, Gandini S, Nicol\u0026ograve; E, et al. Evolution of low HER2 expression between early and advanced-stage breast cancer. European journal of cancer (Oxford, England : 1990) 2022;163:35-43. (In eng). DOI: 10.1016/j.ejca.2021.12.022.\u003c/li\u003e\n\u003cli\u003eGaudio M, Jacobs F, Benvenuti C, et al. Unveiling the HER2-low phenomenon: exploring immunohistochemistry and gene expression to characterise HR-positive HER2-negative early breast cancer. Breast cancer research and treatment 2024;203(3):487-495. (In eng). DOI: 10.1007/s10549-023-07151-3.\u003c/li\u003e\n\u003cli\u003eMarchi\u0026ograve; C, Annaratone L, Marques A, Casorzo L, Berrino E, Sapino A. Evolving concepts in HER2 evaluation in breast cancer: Heterogeneity, HER2-low carcinomas and beyond. Semin Cancer Biol 2021;72:123-135. (In eng). DOI: 10.1016/j.semcancer.2020.02.016.\u003c/li\u003e\n\u003cli\u003eAlmstedt K, Krauthauser L, Kappenberg F, et al. Discordance of HER2-Low between Primary Tumors and Matched Distant Metastases in Breast Cancer. Cancers (Basel) 2023;15(5) (In eng). DOI: 10.3390/cancers15051413.\u003c/li\u003e\n\u003cli\u003eMiglietta F, Griguolo G, Bottosso M, et al. Evolution of HER2-low expression from primary to recurrent breast cancer. NPJ Breast Cancer 2021;7(1):137. (In eng). DOI: 10.1038/s41523-021-00343-4.\u003c/li\u003e\n\u003cli\u003eSchrijver W, Suijkerbuijk KPM, van Gils CH, van der Wall E, Moelans CB, van Diest PJ. Receptor Conversion in Distant Breast Cancer Metastases: A Systematic Review and Meta-analysis. J Natl Cancer Inst 2018;110(6):568-580. (In eng). DOI: 10.1093/jnci/djx273.\u003c/li\u003e\n\u003cli\u003eMehta S, Iyengar A, Barman H, et al. Prevalence of \u0026ldquo;HER2 ultra-low\u0026rdquo; among patients with advanced breast cancer with historical IHC0 status. Journal of Clinical Oncology 2024;42(16_suppl):e13156-e13156. DOI: 10.1200/JCO.2024.42.16_suppl.e13156.\u003c/li\u003e\n\u003cli\u003eJakobsen MR, Teerapakpinyo C, Shuangshoti S, Keelawat S. Comparison between digital image analysis and visual assessment of immunohistochemical HER2 expression in breast cancer. Pathol Res Pract 2018;214(12):2087-2092. (In eng). DOI: 10.1016/j.prp.2018.10.015.\u003c/li\u003e\n\u003cli\u003ePalm C, Connolly CE, Masser R, et al. Determining HER2 Status by Artificial Intelligence: An Investigation of Primary, Metastatic, and HER2 Low Breast Tumors. Diagnostics (Basel) 2023;13(1) (In eng). DOI: 10.3390/diagnostics13010168.\u003c/li\u003e\n\u003cli\u003eWu S, Yue M, Zhang J, et al. The Role of Artificial Intelligence in Accurate Interpretation of HER2 Immunohistochemical Scores 0 and 1+ in Breast Cancer. Mod Pathol 2023;36(3):100054. (In eng). DOI: 10.1016/j.modpat.2022.100054.\u003c/li\u003e\n\u003cli\u003eKrishnamurthy S, Schnitt SJ, Vincent-Salomon A, et al. Fully Automated Artificial Intelligence Solution for Human Epidermal Growth Factor Receptor 2 Immunohistochemistry Scoring in Breast Cancer: A Multireader Study. JCO Precis Oncol 2024;8:e2400353. (In eng). DOI: 10.1200/po.24.00353.\u003c/li\u003e\n\u003cli\u003eGustavson M, Haneder S, Spitzmueller A, et al. Abstract PD6-01: Novel approach to HER2 quantification: Digital pathology coupled with AI-based image and data analysis delivers objective and quantitative HER2 expression analysis for enrichment of responders to trastuzumab deruxtecan (T-DXd; DS-8201), specifically in HER2-low patients. Cancer Research 2021;81(4_Supplement):PD6-01-PD6-01. DOI: 10.1158/1538-7445.Sabcs20-pd6-01.\u003c/li\u003e\n\u003cli\u003eXiong Z, Liu K, Liu S, et al. Precision HER2: a comprehensive AI system for accurate and consistent evaluation of HER2 expression in invasive breast Cancer. BMC cancer 2024;24(1):1204. (In eng). DOI: 10.1186/s12885-024-12980-6.\u003c/li\u003e\n\u003cli\u003eR\u0026uuml;schoff J, Friedrich M, Nagelmeier I, et al. Comparison of HercepTest\u0026trade; mAb pharmDx (Dako Omnis, GE001) with Ventana PATHWAY anti-HER-2/neu (4B5) in breast cancer: correlation with HER2 amplification and HER2 low status. Virchows Arch 2022;481(5):685-694. (In eng). DOI: 10.1007/s00428-022-03378-5.\u003c/li\u003e\n\u003cli\u003eZhang H, Katerji H, Turner BM, Hicks DG. HER2-Low Breast Cancers. Am J Clin Pathol 2022;157(3):328-336. DOI: 10.1093/ajcp/aqab117.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"breast-cancer-research-and-treatment","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"brea","sideBox":"Learn more about [Breast Cancer Research and Treatment](https://www.springer.com/journal/10549)","snPcode":"10549","submissionUrl":"https://submission.nature.com/new-submission/10549/3","title":"Breast Cancer Research and Treatment","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"HER2-low breast cancer; HER2-ultralow breast cancer, HER2-null breast cancer, Artificial Intelligence, Changes in HER2-low status","lastPublishedDoi":"10.21203/rs.3.rs-6557172/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6557172/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eHER2-targeted therapies have improved survival in HER2-positive breast cancer, and recent data suggest potential benefits for patients with HER2-low tumors (defined as immunohistochemistry (IHC) 1\u0026thinsp;+\u0026thinsp;or 2\u0026thinsp;+\u0026thinsp;and, in situ hybridization (ISH)-negative). HER2-low tumors are heterogenous, spanning the hormone receptor-positive and triple-negative subtypes. Assessing HER2-low status remains challenging, especially across specimen types.\u003c/p\u003e\u003ch2\u003eAims\u003c/h2\u003e \u003cp\u003eThis study aims to (1) compare HER2 assessment using conventional microscopy, digital pathology, and an artificial intelligence (AI) model, and (2) investigate changes in HER2-low status between core biopsies, surgical specimens, and metastases.\u003c/p\u003e\u003ch2\u003eMaterials and Methods\u003c/h2\u003e \u003cp\u003eIHC slides from 47 HER2-low breast carcinomas were analyzed using conventional microscopy, digital pathology, and an AI model developed on Aiforia\u0026reg; Create. HER2 statuses were categorized as low, ultralow, and null. Changes in HER2 expression across specimen types were evaluated using agreement measures and visualization tools.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe AI model identified more HER2-low and HER2-ultralow cases compared to conventional methods, improving detection accuracy. HER2 expression differed between specimen types, with metastases exhibiting increased HER2 expression compared to surgical specimens and core biopsies. Digital pathology also showed stronger membranous staining and identified more HER2-positive tumor cells than microscopy.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eAI evaluation is a more sensitive method for HER2-low assessment and reveals expression changes across disease progression. These findings emphasize the need for standardized HER2 assessment to ensure accurate therapy eligibility, particularly for novel treatments like Trastuzumab-Deruxtecan.\u003c/p\u003e","manuscriptTitle":"Changes in HER2low status in 47 advanced breast carcinoma biopsies, matching surgical specimens, and distant metastases assessed by conventional light microscopy, digital pathology, and artificial intelligence","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-13 06:49:45","doi":"10.21203/rs.3.rs-6557172/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-05-24T02:18:29+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-05-23T03:10:54+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-05-10T19:55:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"91743561878171953405148292183732859137","date":"2025-05-10T19:33:57+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"146970689999894141412535847172090754705","date":"2025-05-06T23:07:17+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-05-06T22:03:05+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-05-02T14:58:31+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-05-02T14:55:29+00:00","index":"","fulltext":""},{"type":"submitted","content":"Breast Cancer Research and Treatment","date":"2025-04-29T13:46:32+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"breast-cancer-research-and-treatment","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"brea","sideBox":"Learn more about [Breast Cancer Research and Treatment](https://www.springer.com/journal/10549)","snPcode":"10549","submissionUrl":"https://submission.nature.com/new-submission/10549/3","title":"Breast Cancer Research and Treatment","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"00ae38b7-5dfc-4efa-89fa-37d7831fd005","owner":[],"postedDate":"May 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-07-28T16:01:30+00:00","versionOfRecord":{"articleIdentity":"rs-6557172","link":"https://doi.org/10.1007/s10549-025-07776-6","journal":{"identity":"breast-cancer-research-and-treatment","isVorOnly":false,"title":"Breast Cancer Research and Treatment"},"publishedOn":"2025-07-22 15:57:37","publishedOnDateReadable":"July 22nd, 2025"},"versionCreatedAt":"2025-05-13 06:49:45","video":"","vorDoi":"10.1007/s10549-025-07776-6","vorDoiUrl":"https://doi.org/10.1007/s10549-025-07776-6","workflowStages":[]},"version":"v1","identity":"rs-6557172","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6557172","identity":"rs-6557172","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}