Automated assessment of Ki-67 in breast cancer: the utility of digital image analysis using virtual triple staining and whole slide imaging

Ki-67 evaluation using deep-learning model-assisted digital image analysis in breast cancer

AIMS

To test the efficacy of artificial intelligence (AI)-assisted Ki-67 digital image analysis in invasive breast carcinoma (IBC) with quantitative assessment of AI model performance.

METHODS AND RESULTS

This study used 494 cases of Ki-67 slide images of IBC core needle biopsies. The methods were divided into two steps: (i) construction of a deep-learning model (DL); and (ii) DL implementation for Ki-67 analysis. First, a DL tissue classifier model (DL-TC) and a DL nuclear detection model (DL-ND) were constructed using HALO AI DenseNet V2 algorithm with 31,924 annotations in 300 Ki-67 digital slide images. Whether the class predicted by DL-TC in the test set was correct compared with the annotation of ground truth at the pixel level was evaluated. Second, DL-TC- and DL-ND-assisted digital image analysis (DL-DIA) was performed in the other 194 luminal-type cases and correlations with manual counting and clinical outcome were investigated to confirm the accuracy and prognostic potential of DL-DIA. The performance of DL-TC was excellent and invasive carcinoma nests were well segmented from other elements (average precision: 0.851; recall: 0.878; F1-score: 0.858). Ki-67 index data and the number of nuclei from DL-DIA were positively correlated with data from manual counting (ρ = 0.961, and 0.928, respectively). High Ki-67 index (cutoff 20%) cases showed significantly worse recurrence-free survival and breast cancer-specific survival (P = 0.024, and 0.032, respectively).

CONCLUSION

The performances of DL-TC and DL-ND were excellent. DL-DIA demonstrated a high degree of concordance with manual counting of Ki-67 and the results of this approach have prognostic potential.

Advancing Ki67 hotspot detection in breast cancer: a comparative analysis of automated digital image analysis algorithms

AIM

Manual detection and scoring of Ki67 hotspots is difficult and prone to variability, limiting its clinical utility. Automated hotspot detection and scoring by digital image analysis (DIA) could improve the assessment of the Ki67 hotspot proliferation index (PI). This study compared the clinical performance of Ki67 hotspot detection and scoring DIA algorithms based on virtual dual staining (VDS) and deep learning (DL) with manual Ki67 hotspot PI assessment.

METHODS

Tissue sections of 135 consecutive invasive breast carcinomas were immunohistochemically stained for Ki67. Two DIA algorithms, based on VDS and DL, automatically determined the Ki67 hotspot PI. For manual assessment; two independent observers detected hotspots and calculated scores using a validated scoring protocol.

RESULTS

Automated hotspot detection and assessment by VDS and DL could be performed in 73% and 100% of the cases, respectively. Automated hotspot detection by VDS and DL led to higher Ki67 hotspot PIs (mean 39.6% and 38.3%, respectively) compared to manual consensus Ki67 PIs (mean 28.8%). Comparing manual consensus Ki67 PIs with VDS Ki67 PIs revealed substantial correlation (r = 0.90), while manual consensus versus DL Ki67 PIs demonstrated high correlation (r = 0.95).

CONCLUSION

Automated Ki67 hotspot detection and analysis correlated strongly with manual Ki67 assessment and provided higher PIs compared to manual assessment. The DL-based algorithm outperformed the VDS-based algorithm in clinical applicability, because it did not depend on virtual alignment of slides and correlated stronger with manual scores. Use of a DL-based algorithm may allow clearer Ki67 PI cutoff values, thereby improving the clinical usability of Ki67.

Current status and prospects of artificial intelligence in breast cancer pathology: convolutional neural networks to prospective Vision Transformers

Breast cancer is the most prevalent cancer among women, and its diagnosis requires the accurate identification and classification of histological features for effective patient management. Artificial intelligence, particularly through deep learning, represents the next frontier in cancer diagnosis and management. Notably, the use of convolutional neural networks and emerging Vision Transformers (ViT) has been reported to automate pathologists' tasks, including tumor detection and classification, in addition to improving the efficiency of pathology services. Deep learning applications have also been extended to the prediction of protein expression, molecular subtype, mutation status, therapeutic efficacy, and outcome prediction directly from hematoxylin and eosin-stained slides, bypassing the need for immunohistochemistry or genetic testing. This review explores the current status and prospects of deep learning in breast cancer diagnosis with a focus on whole-slide image analysis. Artificial intelligence applications are increasingly applied to many tasks in breast pathology ranging from disease diagnosis to outcome prediction, thus serving as valuable tools for assisting pathologists and supporting breast cancer management.

Precision HER2: a comprehensive AI system for accurate and consistent evaluation of HER2 expression in invasive breast Cancer

BACKGROUND

With the development of novel anti-HER2 targeted drugs, such as ADCs, it has become increasingly important to accurately interpret HER2 expression in breast cancer. Previous studies have demonstrated high intra-observer and inter-observer variabilities in evaluating HER2 staining by human eyes. There exists a strong requirement to develop artificial intelligence (AI) systems to achieve high-precision HER2 expression scoring for better clinical therapy.

METHODS

In the present study, we collected breast cancer tissue samples and stained consecutive sections with anti-Calponin and anti-HER2 antibodies. High-quality digital images were selected from immunohistochemical slides and interpreted as HER2 3+, 2+, 1+, and 0. AI models were trained and assessed using annotated training and testing sets. The AI model was trained to automatically identify ductal carcinoma in situ (DCIS) by Calponin staining and myoepithelial annotation and filter out DCIS components in HER2-stained slides using image-overlapping techniques. Furthermore, we organized two-phase validation studies. In phase one, pathologists interpreted 112 HER2 whole-slide images (WSIs) without AI assistance, whereas in phase two, pathologists read the same slides using the AI system after a washing period of 2 weeks.

RESULTS

Our AI model greatly improved the accuracy of reading (0.902 vs. 0.710). The number of HER2 1 + patients misdiagnosed as HER2 0 was significantly reduced (32/279 vs. 65/279), and they benefitted from ADC drugs. In addition, the AI algorithm improved the intra-group consistency of HER2 readings by pathologists with different years of experience (intra-class correlation coefficient [ICC]: 0.872-0.926 vs. 0.818-0.908), with the improvement most pronounced among junior pathologists (0.885 vs. 0.818).

CONCLUSIONS

We proposed a high-precision AI system to identify and filter out DCIS components and automatically evaluate HER2 expression in invasive breast cancer.

Synergies and Challenges in the Preclinical and Clinical Implementation of Pathology Artificial Intelligence Applications

Recent introduction of digitalization in pathology has disrupted the field greatly with the potential to change the area immensely. Digital pathology has created the potential of applying advanced quantitative analysis and artificial intelligence (AI) to the domain. In this study, we present an overview of what pathology AI applications have the greatest potential of widespread adoption in the preclinical domain and subsequently, in the clinical setting. We also discuss the major challenges in AI adoption being faced by the field of digital and computational pathology. We review the research literature in the domain and present a detailed analysis of the most promising areas of digital and computational pathology AI research and identify applications that are likely to see the first adoptions of AI technology. Our analysis shows that certain areas and fields of application have received more attention and can potentially affect the field of digital and computational pathology more favorably, leading to the advancement of the field. We also present the main challenges that are faced by the field and provide a comparative analysis of various aspects that are likely to influence the field for the long term in the future.

Use of a Novel Deep Learning Open-Source Model for Quantification of Ki-67 in Breast Cancer Patients in Pakistan: A Comparative Study between the Manual and Automated Methods

Introduction: Breast cancer is the most common cancer in women; its early detection plays a crucial role in improving patient outcomes. Ki-67 is a biomarker commonly used for evaluating the proliferation of cancer cells in breast cancer patients. The quantification of Ki-67 has traditionally been performed by pathologists through a manual examination of tissue samples, which can be time-consuming and subject to inter- and intra-observer variability. In this study, we used a novel deep learning model to quantify Ki-67 in breast cancer in digital images prepared by a microscope-attached camera. Objective: To compare the automated detection of Ki-67 with the manual eyeball/hotspot method. Place and duration of study: This descriptive, cross-sectional study was conducted at the Jinnah Sindh Medical University. Glass slides of diagnosed cases of breast cancer were obtained from the Aga Khan University Hospital after receiving ethical approval. The duration of the study was one month. Methodology: We prepared 140 digital images stained with the Ki-67 antibody using a microscope-attached camera at 10×. An expert pathologist (P1) evaluated the Ki-67 index of the hotspot fields using the eyeball method. The images were uploaded to the DeepLiif software to detect the exact percentage of Ki-67 positive cells. SPSS version 24 was used for data analysis. Diagnostic accuracy was also calculated by other pathologists (P2, P3) and by AI using a Ki-67 cut-off score of 20 and taking P1 as the gold standard. Results: The manual and automated scoring methods showed a strong positive correlation as the kappa coefficient was significant. The p value was <0.001. The highest diagnostic accuracy, i.e., 95%, taking P1 as gold standard, was found for AI, compared to pathologists P2 and P3. Conclusions: Use of quantification-based deep learning models can make the work of pathologists easier and more reproducible. Our study is one of the earliest studies in this field. More studies with larger sample sizes are needed in future to develop a cohort.

The Evolution of Ki-67 and Breast Carcinoma: Past Observations, Present Directions, and Future Considerations

The 1983 discovery of a mouse monoclonal antibody-the Ki-67 antibody-that recognized a nuclear antigen present only in proliferating cells represented a seminal discovery for the pathologic assessment of cellular proliferation in breast cancer and other solid tumors. Cellular proliferation is a central determinant of prognosis and response to cytotoxic chemotherapy in patients with breast cancer, and since the discovery of the Ki-67 antibody, Ki-67 has evolved as an important biomarker with both prognostic and predictive potential in breast cancer. Although there is universal recognition among the international guideline recommendations of the value of Ki-67 in breast cancer, recommendations for the actual use of Ki-67 assays in the prognostic and predictive evaluation of breast cancer remain mixed, primarily due to the lack of assay standardization and inconsistent inter-observer and inter-laboratory reproducibility. The treatment of high-risk ER-positive/human epidermal growth factor receptor-2 (HER2) negative breast cancer with the recently FDA-approved drug abemaciclib relies on a quantitative assessment of Ki-67 expression in the treatment decision algorithm. This further reinforces the urgent need for standardization of Ki-67 antibody selection and staining interpretation, which will hopefully lead to multidisciplinary consensus on the use of Ki-67 as a prognostic and predictive marker in breast cancer. The goals of this review are to highlight the historical evolution of Ki-67 in breast cancer, summarize the present literature on Ki-67 in breast cancer, and discuss the evolving literature on the use of Ki-67 as a companion diagnostic biomarker in breast cancer, with consideration for the necessary changes required across pathology practices to help increase the reliability and widespread adoption of Ki-67 as a prognostic and predictive marker for breast cancer in clinical practice.

Computational pathology in 2030: a Delphi study forecasting the role of AI in pathology within the next decade

BACKGROUND

Artificial intelligence (AI) is rapidly fuelling a fundamental transformation in the practice of pathology. However, clinical integration remains challenging, with no AI algorithms to date in routine adoption within typical anatomic pathology (AP) laboratories. This survey gathered current expert perspectives and expectations regarding the role of AI in AP from those with first-hand computational pathology and AI experience.

METHODS

Perspectives were solicited using the Delphi method from 24 subject matter experts between December 2020 and February 2021 regarding the anticipated role of AI in pathology by the year 2030. The study consisted of three consecutive rounds: 1) an open-ended, free response questionnaire generating a list of survey items; 2) a Likert-scale survey scored by experts and analysed for consensus; and 3) a repeat survey of items not reaching consensus to obtain further expert consensus.

FINDINGS

Consensus opinions were reached on 141 of 180 survey items (78.3%). Experts agreed that AI would be routinely and impactfully used within AP laboratory and pathologist clinical workflows by 2030. High consensus was reached on 100 items across nine categories encompassing the impact of AI on (1) pathology key performance indicators (KPIs) and (2) the pathology workforce and specific tasks performed by (3) pathologists and (4) AP lab technicians, as well as (5) specific AI applications and their likelihood of routine use by 2030, (6) AI's role in integrated diagnostics, (7) pathology tasks likely to be fully automated using AI, and (8) regulatory/legal and (9) ethical aspects of AI integration in pathology.

INTERPRETATION

This systematic consensus study details the expected short-to-mid-term impact of AI on pathology practice. These findings provide timely and relevant information regarding future care delivery in pathology and raise key practical, ethical, and legal challenges that must be addressed prior to AI's successful clinical implementation.

FUNDING

No specific funding was provided for this study.

Artificial intelligence in breast cancer histopathology

This is a review on the use of artificial intelligence for digital breast pathology. A systematic search on PubMed was conducted, identifying 17,324 research papers related to breast cancer pathology. Following a semimanual screening, 664 papers were retrieved and pursued. The papers are grouped into six major tasks performed by pathologists-namely, molecular and hormonal analysis, grading, mitotic figure counting, ki-67 indexing, tumour-infiltrating lymphocyte assessment, and lymph node metastases identification. Under each task, open-source datasets for research to build artificial intelligence (AI) tools are also listed. Many AI tools showed promise and demonstrated feasibility in the automation of routine pathology investigations. We expect continued growth of AI in this field as new algorithms mature.

Comparison Between Manual and Automated Assessment of Ki-67 in Breast Carcinoma: Test of a Simple Method in Daily Practice

BACKGROUND

In the era of "precision medicine," the availability of high-quality tumor biomarker tests is critical and tumor proliferation evaluated by Ki-67 antibody is one of the most important prognostic factors in breast cancer. But the evaluation of Ki-67 index has been shown to suffer from some interobserver variability. The goal of the study is to develop an easy, automated, and reliable Ki-67 assessment approach for invasive breast carcinoma in routine practice.

PATIENTS AND METHODS

A total of 151 biopsies of invasive breast carcinoma were analyzed. The Ki-67 index was evaluated by 2 pathologists with MIB-1 antibody as a global tumor index and also in a hotspot. These 2 areas were also analyzed by digital image analysis (DIA).

RESULTS

For Ki-67 index assessment, in the global and hotspot tumor area, the concordances were very good between DIA and pathologists when DIA focused on the annotations made by pathologist (0.73 and 0.83, respectively). However, this was definitely not the case when DIA was not constrained within the pathologist's annotations and automatically established its global or hotspot area in the whole tissue sample (concordance correlation coefficients between 0.28 and 0.58).

CONCLUSIONS

The DIA technique demonstrated a meaningful concordance with the indices evaluated by pathologists when the tumor area is previously identified by a pathologist. In contrast, basing Ki-67 assessment on automatic tissue detection was not satisfactory and provided bad concordance results. A representative tumoral zone must therefore be manually selected prior to the measurement made by the DIA.

Implementation and Clinical Adoption of Precision Oncology Workflows Across a Healthcare Network

BACKGROUND

Precision oncology relies on molecular diagnostics, and the value-proposition of modern healthcare networks promises a higher standard of care across partner sites. We present the results of a clinical pilot to standardize precision oncology workflows.

METHODS

Workflows are defined as the development, roll-out, and updating of disease-specific molecular order sets. We tracked the timeline, composition, and effort of consensus meetings to define the combination of molecular tests. To assess clinical impact, we examined order set adoption over a two-year period (before and after roll-out) across all gastrointestinal and hepatopancreatobiliary (GI) malignancies, and by provider location within the network.

RESULTS

Development of 12 disease center-specific order sets took ~9 months, and the average number of tests per indication changed from 2.9 to 2.8 (P = .74). After roll-out, we identified significant increases in requests for GI patients (17%; P < .001), compliance with testing recommendations (9%; P < .001), and the fraction of "abnormal" results (6%; P < .001). Of 1088 GI patients, only 3 received targeted agents based on findings derived from non-recommended orders (1 before and 2 after roll-out); indicating that our practice did not negatively affect patient treatments. Preliminary analysis showed 99% compliance by providers in network sites, confirming the adoption of the order sets across the network.

CONCLUSION

Our study details the effort of establishing precision oncology workflows, the adoption pattern, and the absence of harm from the reduction of non-recommended orders. Establishing a modifiable communication tool for molecular testing is an essential component to optimize patient care via precision oncology.

Ki-67 assessment of pancreatic neuroendocrine neoplasms: Systematic review and meta-analysis of manual vs. digital pathology scoring

Ki-67 assessment is a key step in the diagnosis of neuroendocrine neoplasms (NENs) from all anatomic locations. Several challenges exist related to quantifying the Ki-67 proliferation index due to lack of method standardization and inter-reader variability. The application of digital pathology coupled with machine learning has been shown to be highly accurate and reproducible for the evaluation of Ki-67 in NENs. We systematically reviewed all published studies on the subject of Ki-67 assessment in pancreatic NENs (PanNENs) employing digital image analysis (DIA). The most common advantages of DIA were improvement in the standardization and reliability of Ki-67 evaluation, as well as its speed and practicality, compared to the current gold standard approach of manual counts from captured images, which is cumbersome and time consuming. The main limitations were attributed to higher costs, lack of widespread availability (as of yet), operator qualification and training issues (if it is not done by pathologists), and most importantly, the drawback of image algorithms counting contaminating non-neoplastic cells and other signals like hemosiderin. However, solutions are rapidly developing for all of these challenging issues. A comparative meta-analysis for DIA versus manual counting shows very high concordance (global coefficient of concordance: 0.94, 95% CI: 0.83-0.98) between these two modalities. These findings support the widespread adoption of validated DIA methods for Ki-67 assessment in PanNENs, provided that measures are in place to ensure counting of only tumor cells either by software modifications or education of non-pathologist operators, as well as selection of standard regions of interest for analysis. NENs, being cellular and monotonous neoplasms, are naturally more amenable to Ki-67 assessment. However, lessons of this review may be applicable to other neoplasms where proliferation activity has become an integral part of theranostic evaluation including breast, brain, and hematolymphoid neoplasms.

Automated Analysis of Proliferating Cells Spatial Organisation Predicts Prognosis in Lung Neuroendocrine Neoplasms

Simple Summary

Lung neuroendocrine neoplasms (lung NENs) are categorised by morphology, defining a classification sometimes unable to reflect ultimate clinical outcome, particularly for the intermediate domains of adenocarcinomas and large-cell neuroendocrine carcinomas. Moreover, subjectivity and poor reproducibility characterise diagnosis and prognosis assessment of all NENs. The aim of this study was to design and evaluate an objective and reproducible approach to the grading of lung NENs, potentially extendable to other NENs, by exploring a completely new perspective of interpreting the well-recognised proliferation marker Ki-67. We designed an automated pipeline to harvest quantitative information from the spatial distribution of Ki-67-positive cells, analysing its heterogeneity in the entire extent of tumour tissue—which currently represents the main weakness of Ki-67—and employed machine learning techniques to predict prognosis based on this information. Demonstrating the efficacy of the proposed framework would hint at a possible path for the future of grading and classification of NENs.

Abstract

Lung neuroendocrine neoplasms (lung NENs) are categorised by morphology, defining a classification sometimes unable to reflect ultimate clinical outcome. Subjectivity and poor reproducibility characterise diagnosis and prognosis assessment of all NENs. Here, we propose a machine learning framework for tumour prognosis assessment based on a quantitative, automated and repeatable evaluation of the spatial distribution of cells immunohistochemically positive for the proliferation marker Ki-67, performed on the entire extent of high-resolution whole slide images. Combining features from the fields of graph theory, fractality analysis, stochastic geometry and information theory, we describe the topology of replicating cells and predict prognosis in a histology-independent way. We demonstrate how our approach outperforms the well-recognised prognostic role of Ki-67 Labelling Index on a multi-centre dataset comprising the most controversial lung NENs. Moreover, we show that our system identifies arrangement patterns in the cells positive for Ki-67 that appear independently of tumour subtyping. Strikingly, the subset of these features whose presence is also independent of the value of the Labelling Index and the density of Ki-67-positive cells prove to be especially relevant in discerning prognostic classes. These findings disclose a possible path for the future of grading and classification of NENs.

Preoperative ultrasound radiomics analysis for expression of multiple molecular biomarkers in mass type of breast ductal carcinoma in situ

BACKGROUND

The molecular biomarkers of breast ductal carcinoma in situ (DCIS) have important guiding significance for individualized precision treatment. This study was intended to explore the significance of radiomics based on ultrasound images to predict the expression of molecular biomarkers of mass type of DCIS.

METHODS

116 patients with mass type of DCIS were included in this retrospective study. The radiomics features were extracted based on ultrasound images. According to the ratio of 7:3, the data sets of molecular biomarkers were split into training set and test set. The radiomics models were developed to predict the expression of estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), Ki67, p16, and p53 by using combination of multiple feature selection and classifiers. The predictive performance of the models were evaluated using the area under the curve (AUC) of the receiver operating curve.

RESULTS

The investigators extracted 5234 radiomics features from ultrasound images. 12, 23, 41, 51, 31 and 23 features were important for constructing the models. The radiomics scores were significantly (P < 0.05) in each molecular marker expression of mass type of DCIS. The radiomics models showed predictive performance with AUC greater than 0.7 in the training set and test set: ER (0.94 and 0.84), PR (0.90 and 0.78), HER2 (0.94 and 0.74), Ki67 (0.95 and 0.86), p16 (0.96 and 0.78), and p53 (0.95 and 0.74), respectively.

CONCLUSION

Ultrasonic-based radiomics analysis provided a noninvasive preoperative method for predicting the expression of molecular markers of mass type of DCIS with good accuracy.

Improving Ki67 assessment concordance by the use of an artificial intelligence-empowered microscope: a multi-institutional ring study

AIMS

The nuclear proliferation biomarker Ki67 plays potential prognostic and predictive roles in breast cancer treatment. However, the lack of interpathologist consistency in Ki67 assessment limits the clinical use of Ki67. The aim of this article was to report a solution utilising an artificial intelligence (AI)-empowered microscope to improve Ki67 scoring concordance.

METHODS AND RESULTS

We developed an AI-empowered microscope in which the conventional microscope was equipped with AI algorithms, and AI results were provided to pathologists in real time through augmented reality. We recruited 30 pathologists with various experience levels from five institutes to assess the Ki67 labelling index on 100 Ki67-stained slides from invasive breast cancer patients. In the first round, pathologists conducted visual assessment on a conventional microscope; in the second round, they were assisted with reference cards; and in the third round, they were assisted with an AI-empowered microscope. Experienced pathologists had better reproducibility and accuracy [intraclass correlation coefficient (ICC) = 0.864, mean error = 8.25%] than inexperienced pathologists (ICC = 0.807, mean error = 11.0%) in visual assessment. Moreover, with reference cards, inexperienced pathologists (ICC = 0.836, mean error = 10.7%) and experienced pathologists (ICC = 0.875, mean error = 7.56%) improved their reproducibility and accuracy. Finally, both experienced pathologists (ICC = 0.937, mean error = 4.36%) and inexperienced pathologists (ICC = 0.923, mean error = 4.71%) improved the reproducibility and accuracy significantly with the AI-empowered microscope.

CONCLUSION

The AI-empowered microscope allows seamless integration of the AI solution into the clinical workflow, and helps pathologists to obtain higher consistency and accuracy for Ki67 assessment.