Reproducibility of Her2/neu scoring in gastric cancer and assessment of the 10% cut-off rule

HER2-low across solid tumours: different incidences and definitions

Antibody-drug conjugates, particularly trastuzumab deruxtecan (T-DXd), have emerged as effective therapies for various solid tumours. Clinical trials show that T-DXd improves survival in both HER2-positive and HER2-low breast cancer patients. Additionally, it improves survival in HER2-positive gastro-oesophageal cancer and elicits objective responses in HER2-low tumours. Responses have also been noted in lung and gynaecological cancers with HER2 expression, although subgroup analyses for HER2-low cases are lacking. This review assesses HER2 protein expression levels and gene amplification across solid tumours where T-DXd shows potential benefits. We focus on the accuracy and limitations of HER2 testing methods, particularly for identifying HER2-low cancer. A semi-systematic approach was employed, searching EMBASE, Medline, Cochrane, and PubMed databases. We calculated median incidences of HER2-positive, HER2-low, and HER2-0 by immunohistochemistry (IHC), and HER2 amplification by in situ hybridisation (ISH). A total of 144 studies were included, covering breast (n=57), gastro-oesophageal (n=33), lung (n=17), gynaecological (n=24), and various other carcinomas (n=13). The median incidences of HER2-low were 52%, 16%, 58%, and 17% in breast, gastro-oesophageal, endometrial, and ovarian cancers, respectively, with unknown incidences in lung and cervical cancers. Factors influencing HER2-low detection include tumour heterogeneity, antibody clones, observer variability, and lack of validated scoring criteria. Given the significant proportion of HER2-low cases, many patients could benefit from T-DXd, but limitations in detection accuracy necessitate further research and standardisation in diagnostic methods and criteria to advance the clinical utility of T-DXd for HER2-low tumours.

Intratumoral heterogeneity affects tumor regression and Ki67 proliferation index in perioperatively treated gastric carcinoma

BACKGROUND

Intratumoral heterogeneity (ITH) is a major problem in gastric cancer (GC). We tested Ki67 and tumor regression for ITH after neoadjuvant/perioperative chemotherapy.

METHODS

429 paraffin blocks were obtained from 106 neoadjuvantly/perioperatively treated GCs (one to five blocks per case). Serial sections were stained with Masson's trichrome, antibodies directed against cytokeratin and Ki67, and finally digitalized. Tumor regression and three different Ki67 proliferation indices (PI), i.e., maximum PI (KiH), minimum PI (KiL), and the difference between KiH/KiL (KiD) were obtained per block. Statistics were performed in a block-wise (all blocks irrespective of their case-origin) and case-wise manner.

RESULTS

Ki67 and tumor regression showed extensive ITH in our series (maximum ITH within a case: 31% to 85% for KiH; 4.5% to 95.6% for tumor regression). In addition, Ki67 was significantly associated with tumor regression (p < 0.001). Responders (<10% residual tumor, p = 0.016) exhibited prolonged survival. However, there was no significant survival benefit after cut-off values were increased ≥20% residual tumor mass. Ki67 remained without prognostic value.

CONCLUSIONS

Digital image analysis in tumor regression evaluation might help overcome inter- and intraobserver variability and validate classification systems. Ki67 may serve as a sensitivity predictor for chemotherapy and an indicator of ITH.

Artificial intelligence in gastric cancer: applications and challenges

Gastric cancer (GC) is one of the most common malignant tumors with high mortality. Accurate diagnosis and treatment decisions for GC rely heavily on human experts' careful judgments on medical images. However, the improvement of the accuracy is hindered by imaging conditions, limited experience, objective criteria, and inter-observer discrepancies. Recently, the developments of machine learning, especially deep-learning algorithms, have been facilitating computers to extract more information from data automatically. Researchers are exploring the far-reaching applications of artificial intelligence (AI) in various clinical practices, including GC. Herein, we aim to provide a broad framework to summarize current research on AI in GC. In the screening of GC, AI can identify precancerous diseases and assist in early cancer detection with endoscopic examination and pathological confirmation. In the diagnosis of GC, AI can support tumor-node-metastasis (TNM) staging and subtype classification. For treatment decisions, AI can help with surgical margin determination and prognosis prediction. Meanwhile, current approaches are challenged by data scarcity and poor interpretability. To tackle these problems, more regulated data, unified processing procedures, and advanced algorithms are urgently needed to build more accurate and robust AI models for GC.

GasHisSDB: A new gastric histopathology image dataset for computer aided diagnosis of gastric cancer

BACKGROUND AND OBJECTIVE

Gastric cancer is the fifth most common cancer globally, and early detection of gastric cancer is essential to save lives. Histopathological examination of gastric cancer is the gold standard for the diagnosis of gastric cancer. However, computer-aided diagnostic techniques are challenging to evaluate due to the scarcity of publicly available gastric histopathology image datasets.

METHODS

In this paper, a noble publicly available Gastric Histopathology Sub-size Image Database (GasHisSDB) is published to identify classifiers' performance. Specifically, two types of data are included: normal and abnormal, with a total of 245,196 tissue case images. In order to prove that the methods of different periods in the field of image classification have discrepancies on GasHisSDB, we select a variety of classifiers for evaluation. Seven classical machine learning classifiers, three Convolutional Neural Network classifiers, and a novel transformer-based classifier are selected for testing on image classification tasks.

RESULTS

This study performed extensive experiments using traditional machine learning and deep learning methods to prove that the methods of different periods have discrepancies on GasHisSDB. Traditional machine learning achieved the best accuracy rate of 86.08% and a minimum of just 41.12%. The best accuracy of deep learning reached 96.47% and the lowest was 86.21%. Accuracy rates vary significantly across classifiers.

CONCLUSIONS

To the best of our knowledge, it is the first publicly available gastric cancer histopathology dataset containing a large number of images for weakly supervised learning. We believe that GasHisSDB can attract researchers to explore new algorithms for the automated diagnosis of gastric cancer, which can help physicians and patients in the clinical setting.

[Application and Progress of Convolutional Neural Network-based Pathological Diagnosis of Gastric Cancer]

The incidence of gastric cancer is the highest among all kinds of malignant tumors in China. Because gastric cancer is very hard to identify in its early stage, the early diagnosis rate of gastric cancer in China is relatively low. At present, the pathological diagnosis of gastric cancer mainly depends on the diagnosis of pathologists. However, the gradual improvement of people's living standards and the growing demand for medical and health care have exacerbated the shortage of medical resources, which has become a even more serious problem. Therefore, there is an urgent need for new technologies to help deal with this challenge. In recent years, with the rapid development of artificial intelligence (AI) and digital pathology, AI-aided pathological diagnosis based on convolutional neural network (CNN) as the core technology is showing promises for improving the diagnostic efficiency of gastric cancer. It is also of great significance for the early diagnosis and treatment of the disease and the reduction of its high incidence and mortality. We herein summarize the application and progress of deep-learning CNN in pathological diagnosis of gastric cancer, as well as the existing problems and prospects of future development.

First-in-human phase 0 study of 111In-CHX-A"-DTPA trastuzumab for HER2 tumor imaging

Introduction:

Tumors over-expressing the human epithelial receptor 2 (HER2) or exhibiting amplification or mutation of its proto-oncogene have a poorer prognosis. Using trastuzumab and/or other HER2 targeted therapies can increase overall survival in patients with HER2(+) tumors making it critical to accurately identify patients who may benefit. We report on a Phase 0 study of the imaging agent, ¹¹¹In-CHX-A”-DTPA trastuzumab, in patients with known HER2 status to evaluate its safety and biodistribution and to obtain preliminary data regarding its ability to provide an accurate, whole-body, non-invasive means to determine HER2 status.

Methods:

¹¹¹In-CHX-A”-DTPA trastuzumab was radiolabeled on-site and slowly infused into 11 patients who underwent single (n=5) or multiple (n=6) ɣ-camera (n=6) and/or SPECT (n=8) imaging sessions.

Results:

No safety issues were identified. Visual and semi-quantitative imaging data were concordant with tissue HER2 expression profiling in all but 1 patient. The biodistribution showed intense peak liver activity at the initial imaging timepoint (3.3h) and a single-phase clearance fit of the average time-activity curve (TAC) estimated t_1/2=46.9h (R²=0.97; 95%CI 41.8 to 53h). This was followed by high gastrointestinal (GI) tract activity peaking by 52h. Linear regression predicted GI clearance by 201.2h (R² =0.96; 95%CI 188.5 to 216.9h). Blood pool had lower activity with its maximum on the initial images. Non-linear regression fit projected a t_1/2=34.2h (R² =0.96; 95%CI 25.3 to 46.3h). Assuming linear whole-body clearance, linear regression projected complete elimination (x-intercept) at 256.5hr (R²=0.96; 95%CI 186.1 to 489.2h).

Conclusion:

¹¹¹In-CHX-A”-DTPA trastuzumab can be safely imaged in humans. The biodistribution allowed for visual and semiquantitative analysis with results concordant with tissue expression profiling in 10 of 11 patients. Advances in Knowledge and Implications for Patient Care Using readily available components and on-site radiolabeling ¹¹¹In-CHX-A”-DTPA trastuzumab SPECT imaging may provide an economical, non-invasive means to detect HER2 over-expression.

Physical basis of the 'magnification rule' for standardized Immunohistochemical scoring of HER2 in breast and gastric cancer

Background

Detection of HER2/neu receptor overexpression and/or amplification is a prerequisite for efficient anti-HER2 treatment of breast and gastric carcinomas. Immunohistochemistry (IHC) of the HER2 protein is the most common screening test, thus precise and reproducible IHC-scoring is of utmost importance. Interobserver variance still is a problem; in particular in gastric carcinomas the reliable differentiation of IHC scores 2+ and 1+ is challenging.

Herein we describe the physical basis of what we called the ‘magnification rule’: Different microscope objectives are employed to reproducibly subdivide the continuous spectrum of IHC staining intensities into distinct categories (1+, 2+, 3+).

Methods

HER2-IHC was performed on 120 breast cancer biopsy specimens (n = 40 per category). Width and color-intensity of membranous DAB chromogen precipitates were measured by whole-slide scanning and digital morphometry. Image-analysis data were related to semi-quantitative manual scoring according to the magnification rule and to the optical properties of the employed microscope objectives.

Results

The semi-quantitative manual HER2-IHC scores are correlated to color-intensity measured by image-analysis and to the width of DAB-precipitates. The mean widths ±standard deviations of precipitates were: IHC-score 1+, 0.64 ± 0.1 μm; score 2+, 1.0 ± 0.23 μm; score 3+, 2.14 ± 0.4 μm. The width of precipitates per category matched the optical resolution of the employed microscope objective lenses: Approximately 0.4 μm (40×), 1.0 μm (10×) and 2.0 μm (5×).

Conclusions

Perceived intensity, width of the DAB chromogen precipitate, and absolute color-intensity determined by image-analysis are linked. These interrelations form the physical basis of the ‘magnification rule’: 2+ precipitates are too narrow to be observed with 5× microscope objectives, 1+ precipitates are too narrow for 10× objectives. Thus, the rule uses the optical resolution windows of standard diagnostic microscope objectives to derive the width of the DAB-precipitates. The width is in turn correlated with color-intensity. Hereby, the more or less subjective estimation of IHC scores based only on the staining-intensity is replaced by a quasi-morphometric measurement. The principle seems universally applicable to immunohistochemical stainings of membrane-bound biomarkers that require an intensity-dependent scoring.

Electronic supplementary material

The online version of this article (10.1186/s13000-018-0696-x) contains supplementary material, which is available to authorized users.

Poor efficacy response to trastuzumab therapy in advanced gastric cancer with homogeneous HER2 positive and non-intestinal type

Introduction

Factors affecting trastuzumab efficacy in advanced gastric cancer (GC) are largely unknown. Heterogeneity is a notable feature of HER2 in GC. Whether the heterogeneity influences trastuzumab efficacy is still unknown.

Results

The HER2^homogeneous group and HER2^{heterogeneous} group showed no statistical difference in RR (46.4% vs 55.0%, P = 0.558), PFS (5.80 vs 6.30 months, P = 0.804) and OS (16.00 vs 16.00 months, P = 0.787). The Lauren^intestinal group and Lauren^{non-intestinal} group demonstrated no discrepancy in PFS (6.00 vs 6.00 months, P = 0.912) and OS (16.50 vs 14.00 months, P = 0.227). However, by combining HER2 heterogeneity and Lauren classification, PFS and OS of HER2^homogeneous/Lauren^{non-intestinal} subgroup was the shortest among the 4 subgroups (P = 0.012 and P = 0.037), which was much shorter than the other patients (PFS:3.00 vs 6.30 months, P = 0.003; OS: 4.50 vs 16.50 months, P = 0.004). Univariate and multivariate analysis showed that HER2 heterogeneity combined with Lauren classification was an independent prognostic factor in both PFS (P = 0.031 and P = 0.002) and OS (P = 0.039 and P = 0.013).

Materials and Methods

48 patients with HER2 positive advanced GCs accepting trastuzumab treatment were retrospectively analyzed. Based on HER2 heterogeneity, the patients were divided into a HER2^homogeneous group and a HER2^{heterogeneous} group. Response rate (RR), progression free survival (PFS), and overall survival (OS) were compared. Main clinicopathological factors including Lauren classification were subjected to subgroup analysis.

Conclusions

HER2 heterogeneity alone may not correlate with trastuzumab efficacy in HER2 positive advanced GCs. HER2 heterogeneity combined with Lauren classification may help to identify a subgroup with poor response to trastuzumab which is homogeneous HER2 positive and non-intestinal type.

Clinicopathologic Characteristics of Microsatellite Instable Gastric Carcinomas Revisited: Urgent Need for Standardization

Microsatellite instable gastric cancer (MSI-GC) is a specific molecular subtype of GC. We studied the phenotypes, genotypes, and clinicopathologic characteristics of MSI-GC in a white GC cohort and compared our findings with an extended literature review. The study cohort consisted of 482 patients. Specimens were available from 452 cases and were used for immunostaining (MLH1, PMS2, MSH2, MSH6) and molecular biological analyses (BAT-25, BAT-26, NR-21, NR-24, NR-27; Epstein-Barr virus in situ hybridization). Thirty-four (7.5%) GCs were MSI. Loss of MLH1 and/or PMS2 was found in 30 (88%) MSI-GC, 3 (9%) showed loss of MSH2 and/or MSH6. One (3%) MSI-GC was identified only by molecular biological testing. A single case was heterogeneous and contained microsatellite-stable and instable tumor areas. Twenty-one (62%) MSI-GCs showed unusual histologic features. MSI-GC was not found in diffuse-type or Epstein-Barr virus-positive GC. MSI-GC was significantly more prevalent in elderly patients, distal stomach, and was associated with a significantly lower number of lymph node metastases and a significantly better overall and tumor-specific survival. MSI-GC constitutes a small but relevant subgroup of GC with distinct clinicopathologic characteristics. Our literature review illustrates the shortcomings of missing standardized testing algorithms with prevalences of MSI-GC ranging from 0% to 44.5%. Future studies should test the hypothesis that patients with MSI-GCs may not need adjuvant/perioperative chemotherapy. However, this will require a standardized, quality-controlled diagnostic algorithm of MSI for GC.

Deep convolutional neural networks for automatic classification of gastric carcinoma using whole slide images in digital histopathology

Deep learning using convolutional neural networks is an actively emerging field in histological image analysis. This study explores deep learning methods for computer-aided classification in H&E stained histopathological whole slide images of gastric carcinoma. An introductory convolutional neural network architecture is proposed for two computerized applications, namely, cancer classification based on immunohistochemical response and necrosis detection based on the existence of tumor necrosis in the tissue. Classification performance of the developed deep learning approach is quantitatively compared with traditional image analysis methods in digital histopathology requiring prior computation of handcrafted features, such as statistical measures using gray level co-occurrence matrix, Gabor filter-bank responses, LBP histograms, gray histograms, HSV histograms and RGB histograms, followed by random forest machine learning. Additionally, the widely known AlexNet deep convolutional framework is comparatively analyzed for the corresponding classification problems. The proposed convolutional neural network architecture reports favorable results, with an overall classification accuracy of 0.6990 for cancer classification and 0.8144 for necrosis detection.

HER2 Testing and Clinical Decision Making in Gastroesophageal Adenocarcinoma: Guideline From the College of American Pathologists, American Society for Clinical Pathology, and the American Society of Clinical Oncology

Context ERBB2 (erb-b2 receptor tyrosine kinase 2 or HER2) is currently the only biomarker established for selection of a specific therapy for patients with advanced gastroesophageal adenocarcinoma (GEA). However, there are no comprehensive guidelines for the assessment of HER2 in patients with GEA. Objectives To establish an evidence-based guideline for HER2 testing in patients with GEA, formalize the algorithms for methods to improve the accuracy of HER2 testing while addressing which patients and tumor specimens are appropriate, and to provide guidance on clinical decision making. Design The College of American Pathologists (CAP), American Society for Clinical Pathology (ASCP), and the American Society of Clinical Oncology (ASCO) convened an Expert Panel to conduct a systematic review of the literature to develop an evidence-based guideline with recommendations for optimal HER2 testing in patients with GEA. Results The Panel is proposing 11 recommendations with strong agreement from the open comment participants. Recommendations The Panel recommends that tumor specimen(s) from all patients with advanced GEA, who are candidates for HER2-targeted therapy, should be assessed for HER2 status before the initiation of HER2-targeted therapy. Clinicians should offer combination chemotherapy and an HER2-targeted agent as initial therapy for all patients with HER2-positive advanced GEA. For pathologists, guidance is provided for morphologic selection of neoplastic tissue, testing algorithms, scoring methods, interpretation and reporting of results, and laboratory quality assurance. Conclusion This guideline provides specific recommendations for assessment of HER2 in patients with advanced GEA while addressing pertinent technical issues and clinical implications of the results.

HER2 Testing and Clinical Decision Making in Gastroesophageal Adenocarcinoma: Guideline From the College of American Pathologists, American Society for Clinical Pathology, and American Society of Clinical Oncology

CONTEXT

- ERBB2 (erb-b2 receptor tyrosine kinase 2 or HER2) is currently the only biomarker established for selection of a specific therapy for patients with advanced gastroesophageal adenocarcinoma (GEA). However, there are no comprehensive guidelines for the assessment of HER2 in patients with GEA.

OBJECTIVES

- To establish an evidence-based guideline for HER2 testing in patients with GEA, to formalize the algorithms for methods to improve the accuracy of HER2 testing while addressing which patients and tumor specimens are appropriate, and to provide guidance on clinical decision making.

DESIGN

- The College of American Pathologists, American Society for Clinical Pathology, and American Society of Clinical Oncology convened an expert panel to conduct a systematic review of the literature to develop an evidence-based guideline with recommendations for optimal HER2 testing in patients with GEA.

RESULTS

- The panel is proposing 11 recommendations with strong agreement from the open-comment participants.

RECOMMENDATIONS

- The panel recommends that tumor specimen(s) from all patients with advanced GEA, who are candidates for HER2-targeted therapy, should be assessed for HER2 status before the initiation of HER2-targeted therapy. Clinicians should offer combination chemotherapy and a HER2-targeted agent as initial therapy for all patients with HER2-positive advanced GEA. For pathologists, guidance is provided for morphologic selection of neoplastic tissue, testing algorithms, scoring methods, interpretation and reporting of results, and laboratory quality assurance.

CONCLUSIONS

- This guideline provides specific recommendations for assessment of HER2 in patients with advanced GEA while addressing pertinent technical issues and clinical implications of the results.

HER2 Testing and Clinical Decision Making in Gastroesophageal Adenocarcinoma: Guideline From the College of American Pathologists, American Society for Clinical Pathology, and American Society of Clinical Oncology

CONTEXT

ERBB2 (erb-b2 receptor tyrosine kinase 2 or HER2) is currently the only biomarker established for selection of a specific therapy for patients with advanced gastroesophageal adenocarcinoma (GEA). However, there are no comprehensive guidelines for the assessment of HER2 in patients with GEA.

OBJECTIVES

To establish an evidence-based guideline for HER2 testing in patients with GEA, to formalize the algorithms for methods to improve the accuracy of HER2 testing while addressing which patients and tumor specimens are appropriate, and to provide guidance on clinical decision making.

DESIGN

The College of American Pathologists, American Society for Clinical Pathology, and American Society of Clinical Oncology convened an expert panel to conduct a systematic review of the literature to develop an evidence-based guideline with recommendations for optimal HER2 testing in patients with GEA.

RESULTS

The panel is proposing 11 recommendations with strong agreement from the open-comment participants.

RECOMMENDATIONS

The panel recommends that tumor specimen(s) from all patients with advanced GEA, who are candidates for HER2-targeted therapy, should be assessed for HER2 status before the initiation of HER2-targeted therapy. Clinicians should offer combination chemotherapy and a HER2-targeted agent as initial therapy for all patients with HER2-positive advanced GEA. For pathologists, guidance is provided for morphologic selection of neoplastic tissue, testing algorithms, scoring methods, interpretation and reporting of results, and laboratory quality assurance.

CONCLUSIONS

This guideline provides specific recommendations for assessment of HER2 in patients with advanced GEA while addressing pertinent technical issues and clinical implications of the results.

HER2 heterogeneity in gastric/gastroesophageal cancers: From benchside to practice

HER2 is overexpressed in approximately 10%-20% of gastric and gastroesophageal junction carcinomas. In these types of cancer, accurate assessment of HER2 status is mandatory, for selecting patients who may benefit from targeted therapies with anti-HER2 drugs such as Trastuzumab. This manuscript focuses on HER2 in gastric carcinogenesis, on optimal evaluation of HER2 and on the possible causes which may contribute to inaccurate HER2 evaluation. Similarly to breast cancer HER2 evaluation, standardization of HER2 testing in gastric cancer is necessary in diagnostic practice. The three principle aspects which require consideration are: (1) the choice of sample with regards to cancer morphology - intestinal vs diffuse areas; (2) the choice of scoring criteria - use of HER2 scoring criteria specific for gastric cancer; and (3) the choice of HER2 evaluation methods - use of an algorithm in which both immunohistochemistry and in situ hybridization play a role. Problematic issues include: (1) pre-analytic variables with particular emphasis on fixation; (2) recommended methodology for HER2 assessment (immunohistochemistry vs in situ hybridization); (3) HER2 heterogeneity both within the primary tumor and between primary tumor and metastases; (4) reliability of biopsies in HER 2 evaluation; and (5) quantity of sample (FFPE blocks from surgical specimens or endoscopic biopsies) necessary for an adequate assessment.

MET in gastric cancer--discarding a 10% cutoff rule

Aims

We aimed to develop a putative predictive biomarker score for future hepatocyte growth factor receptor (MET)‐targeted therapy of gastric cancer (GC).

Methods and results

MET expression and MET amplification were analysed by immunohistochemistry (IHC) and chromogenic in‐situ hybridization (CISH) in 470 GC patients. Immunostaining was documented with the HistoScore. The percentage area of MET‐amplified tumour cell clones was assessed by virtual microscopy. The expression of MET was heterogeneous in primary and metastatic GC. Immunostaining intensity (MET‐IHC 2+/3+) correlated with MET amplification and a positive MET status was defined by a combination of MET‐IHC 2+ or 3+ with MET amplification, or MET‐IHC 3+ without MET amplification. The prognostic significance of the MET status was independent from the percentage area of positive tumour cells (e.g. <10 versus ≥10%). MET‐positive GCs were microsatellite stable and of KRAS/PIK3CA wild‐type. MET‐positive GCs had a very poor prognosis, with a median survival of 5.4 months and a hazard ratio of 2.126.

Conclusions

A combination of immunohistochemistry and CISH is suitable to assess MET status. If MET status is used as a predictive biomarker, prospective studies should pay specific attention to adequate tissue sampling, should ignore cutoff values for tumour areas, may consider the KRAS and PIK3CA genotype as negative predictive markers and should carry out the analysis expeditiously.