Image-based multiplex immune profiling of cancer tissues: translational implications. A report of the International Immuno-oncology Biomarker Working Group on Breast Cancer

Recent advances in the field of immuno-oncology have brought transformative changes in the management of cancer patients. The immune profile of tumours has been found to have key value in predicting disease prognosis and treatment response in various cancers. Multiplex immunohistochemistry and immunofluorescence have emerged as potent tools for the simultaneous detection of multiple protein biomarkers in a single tissue section, thereby expanding opportunities for molecular and immune profiling while preserving tissue samples. By establishing the phenotype of individual tumour cells when distributed within a mixed cell population, the identification of clinically relevant biomarkers with high-throughput multiplex immunophenotyping of tumour samples has great potential to guide appropriate treatment choices. Moreover, the emergence of novel multi-marker imaging approaches can now provide unprecedented insights into the tumour microenvironment, including the potential interplay between various cell types. However, there are significant challenges to widespread integration of these technologies in daily research and clinical practice. This review addresses the challenges and potential solutions within a structured framework of action from a regulatory and clinical trial perspective. New developments within the field of immunophenotyping using multiplexed tissue imaging platforms and associated digital pathology are also described, with a specific focus on translational implications across different subtypes of cancer. © 2024 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Pitfalls in machine learning-based assessment of tumor-infiltrating lymphocytes in breast cancer: A report of the International Immuno-Oncology Biomarker Working Group on Breast Cancer

The clinical significance of the tumor-immune interaction in breast cancer is now established, and tumor-infiltrating lymphocytes (TILs) have emerged as predictive and prognostic biomarkers for patients with triple-negative (estrogen receptor, progesterone receptor, and HER2-negative) breast cancer and HER2-positive breast cancer. How computational assessments of TILs might complement manual TIL assessment in trial and daily practices is currently debated. Recent efforts to use machine learning (ML) to automatically evaluate TILs have shown promising results. We review state-of-the-art approaches and identify pitfalls and challenges of automated TIL evaluation by studying the root cause of ML discordances in comparison to manual TIL quantification. We categorize our findings into four main topics: (1) technical slide issues, (2) ML and image analysis aspects, (3) data challenges, and (4) validation issues. The main reason for discordant assessments is the inclusion of false-positive areas or cells identified by performance on certain tissue patterns or design choices in the computational implementation. To aid the adoption of ML for TIL assessment, we provide an in-depth discussion of ML and image analysis, including validation issues that need to be considered before reliable computational reporting of TILs can be incorporated into the trial and routine clinical management of patients with triple-negative breast cancer. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Spatial analyses of immune cell infiltration in cancer: current methods and future directions: A report of the International Immuno-Oncology Biomarker Working Group on Breast Cancer

Modern histologic imaging platforms coupled with machine learning methods have provided new opportunities to map the spatial distribution of immune cells in the tumor microenvironment. However, there exists no standardized method for describing or analyzing spatial immune cell data, and most reported spatial analyses are rudimentary. In this review, we provide an overview of two approaches for reporting and analyzing spatial data (raster versus vector-based). We then provide a compendium of spatial immune cell metrics that have been reported in the literature, summarizing prognostic associations in the context of a variety of cancers. We conclude by discussing two well-described clinical biomarkers, the breast cancer stromal tumor infiltrating lymphocytes score and the colon cancer Immunoscore, and describe investigative opportunities to improve clinical utility of these spatial biomarkers. © 2023 The Pathological Society of Great Britain and Ireland.

Digital image analysis and assisted reading of the HER2 score display reduced concordance: pitfalls in the categorisation of HER2-low breast cancer

AIMS

Digital image analysis (DIA) is used increasingly as an assisting tool to evaluate biomarkers, including human epidermal growth factor receptor 2 (HER2) in invasive breast cancer (BC). DIA can assist pathologists in HER2 evaluation by presenting quantitative information about the HER2 staining in APP assisted reading (AR). Concurrently, the HER2-low category (HER2-1+/2+ without HER2 gene amplification) has gained prominence due to newly developed antibody-drug conjugates. However, major inter- and intraobserver variability have been observed for the entity. The present quality assurance study investigated the concordance between DIA and AR in clinical use, especially concerning the HER2-low category.

METHODS AND RESULTS

HER2 immunohistochemistry (IHC) in 761 tumours from 727 patients was evaluated in tissue microarray (TMA) cores by DIA (Visiopharm HER2-CONNECT) and AR. Overall concordance between HER2-scores were 73% (n = 552, weighted-κ: 0.66), and 88% (n = 669, weighted-κ: 0.70), when combining HER2-0/1+. A total of 205 scores were discordant by one category, while four were discordant by two categories. A heterogeneous HER2 pattern was relatively common in the discordant cases and a pitfall in the categorisation of HER2-low BC. AR more commonly reassigned a lower HER2 score (from HER2-1+ to HER2-0) within the HER2-low subgroup (n = 624) compared with DIA.

CONCLUSION

DIA and AR display moderate agreement with heterogeneous and aberrant staining, representing a source of discordance and a pitfall in the evaluation of HER2.

Abstract 1708: Novel analysis method for in situ spatial phenotyping of cell populations in multimarker imagery

Spatial biology is the study of tissues within their own context. This has become increasingly important in immuno-oncology because of the need to understand complex tumor microenvironment architecture, especially regarding the immune contexture in and around solid tumors. The main method for understanding the spatial biology of tumors has been the use of multi-marker staining and multiplexed imaging methods in FFPE sections. This imaging data is used to segment the tissue into relevant compartments (tumor, stroma, etc.) and perform multiplex phenotyping of the cells. For highplex data (10+ markers), phenotyping has typically used an unsupervised clustering algorithm. This was a good starting point, but has many limitations, including a reliance on having a sufficient population of cells of each type, variability sample to sample, and no means of differentiating between biologically relevant or impossible phenotypes. Improving cellular phenotyping for highplex samples is a requirement for many of the new multiplexed imaging modalities that are being used.

We have developed a novel method of phenotyping cells in situ in multiplexed imaging data that involves the biological concept of dividing markers into two user-defined categories: lineage markers and functional markers. Lineage markers are used to identify cellular phenotypes, while functional markers are measured afterwards for their expression levels in across each phenotype and in each individual cell. Lineage markers (e.g., CD45, α-SMA, CD8, cytokeratin, CD20, etc.) define the phenotype of the cell and are essentially binary; they are either present or not present with the correct biodistribution. Functional markers (PD-1, PD-L1, Ki67, etc.), on the other hand, are quantitatively measured in each cell. This new method first uses a supervised classification methodology using lineage markers to find cells with expected phenotypes. In a second step, the levels of functional markers are measured for each cell. Unexpected phenotypes in unclassified cells can then be explored. These data can then be used for downstream spatial/distance/hotspot analysis. During the setup of the analysis the user can explore and QC the data using an interactive data exploration tool which allows for a constant algorithm supervision and correction.

Here we present data of the novel workflow for 7-plex (Vectra) and 20+-plex (Hyperion/CODEX) data and found that the new method increased the efficiency of setting up complex analysis workflows and improved analysis results. Fewer non-biological phenotypes were found, cell classifications were improved compared to auto-clustering, while still exploring unexpected phenotypes. Interactive exploration of the cell populations enabled quicker validation of results and deeper understanding of the tumor immune contexture. This new method will improve analyses of the next generation of highplex imaging datasets.

Citation Format: Fabian T. Schneider, James R. Mansfield, Alessandro S. Massaro, Simon Haastrup, Rasmus A. Lyngby, Andreas Hussing, Johan Dore Hansen, Jeppe Thagaard. Novel analysis method for in situ spatial phenotyping of cell populations in multimarker imagery [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1708.

Abstract 4144: An AI-based, automated workflow for identification and scoring of invasive tumors in Ki-67 stained breast cancer specimens

Understanding the rate of tumor cell growth in breast cancer specimens may be indicative of disease aggressiveness, a tumor characteristic which can be used to make an informed treatment decision. The nuclear protein Ki-67 is increased in cells as they prepare to divide or proliferate and is therefore widely used as a proliferation marker for tumor progression. This degree of tumor cell proliferation, or the proliferative index, is commonly detailed in pathology reports shared with the patient care team. In this study, we utilized the Ki-67 [K2] immunohistochemistry (IHC) assay to stain 10 breast cancer specimens. Stained slides were imaged using the AT2 scanner (Leica Biosystems, Buffalo Grove, IL) and analyzed using the Visiopharm Image Analysis platform. Previous efforts to assess Ki-67 positivity utilizing image analysis have relied on the use of a secondary stain or manual effort by the pathologist to exclude non-invasive tumor regions. These antiquated methods are costly to the lab as they require additional materials or valuable pathologist time. Our novel image analysis approach utilizes artificial intelligence (AI) to automatically denote non-invasive verses invasive tumor regions, which can then be used to quantify the Ki-67 proliferative index. This valuable tool will allow for greater accuracy, cost-savings, and time efficiency when analyzing breast cancer samples compared to traditional methods.

Citation Format: Bhavika Patel, Stephanie Allen, Sameer Talwalkar, Navi Mehra, Jeppe Thagaard, Thomas W. Ramsing, Agnete Overgaard, Jenifer Caldara. An AI-based, automated workflow for identification and scoring of invasive tumors in Ki-67 stained breast cancer specimens [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 4144.

Abstract 1709: Improved understanding of the biology and pathophysiology of the tumor microenvironment in PDAC samples revealed by InSituPlex, Imaging Mass Cytometry, and advanced image processing

Pancreatic cancer remains a deadly disease due to difficulties hindering its early diagnosis, giving way to metastasis of the tumor and resulting in poor prognosis. While there are many neoplasms of the pancreas, pancreatic invasive ductal adenocarcinoma (PDAC) is the most common, and treatment options are few, with poor overall survival. Recently, several publications have demonstrated improved outcomes with the inclusion of immunotherapy to cytotoxic drug combinations in some patients. However, optimally selecting patients as candidates for immunotherapy-chemotherapy combinations remains a critical challenge.

The complexities of the tumor microenvironment (TME) have been implicated in the failure of chemotherapy, radiation therapy, and immunotherapy. The tumor microenvironment of PDAC is especially rich with multiple interactions between pancreatic epithelial/cancer cells, stromal cells, immune cells, and the extracellular matrix (ECM). PDACs are characterized by a complex ECM of desmoplastic reaction consisting of an extensive and dense fibrotic stroma that surrounds and infiltrates clusters of malignant epithelial cells, together with the loss of basement membrane integrity and an abnormal vasculature. PD-L1 is also expressed in PDACs, and its overexpression has been associated with a poor prognosis.

In the present study we demonstrate a unique tissue phenotyping workflow combining complementary methods that can unravel the complexity of the tumor microenvironment. We highlight how a workflow combining multiple elements provides utility, robustness, and an ability to derive biological insights in PDAC samples. An InSituPlex® PD-L1 multiplex immunofluorescence assay (4 markers, CD8, CD68, PD-L1, Pan CK/Sox10) was used on whole slides to identify areas of high, medium, and low PD-L1 expression. Imaging Mass Cytometry™ (IMC™, 40 markers) was performed on selected regions of interest from each slide. Advanced tissue and cell segmentation followed by multiplex cellular phenotyping and spatial analyses were performed on both whole-slide InSituPlex and IMC data. These methods combine to give a detailed readout of the location and bio-distribution of specific cell phenotypes in situ in the TME of PDAC. Four-plex imaging and analysis of whole slides gives an overview of the immune status of the section, while 40-plex imaging and analysis gives a comprehensive and multiparametric exploration of cells present. These methods combine to reveal an exceptional view of the PDAC TME at the single-cell level.

Citation Format: Andrew Quong, Jordan Nieto, Derek Quong, Amanda Esch, Kirsteen Maclean, Mark Rees, Devan Fleury, Gourab Chatterjee, Keith Wharton, Jeppe Thagaard, Fabian Schneider, Dan Winkowski, James Mansfield. Improved understanding of the biology and pathophysiology of the tumor microenvironment in PDAC samples revealed by InSituPlex, Imaging Mass Cytometry, and advanced image processing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1709.

The utilization of artificial intelligence to produce clinically relevant scores on PD-L1 immunostained non-small cell lung cancer biopsies

e14576

Background: Immunotherapies targeting complex programmed death-ligand 1 (PD-L1) interactions have been groundbreaking in the treatment of non-small cell lung cancer (NSCLC), offering new strategies for patients who are likely to respond. PD-L1 binds to the checkpoint PD-1 to regulate T cells, which has an important role in boosting the immune response. Identifying patients who may benefit from PD-L1 focused care is dependent on the interpretation of the drug’s associated companion diagnostic (CDx) immunohistochemistry (IHC) assay. PD-L1 IHC assays stratify NSCLC patients using a Tumor Proportion Score (TPS), a manual scoring paradigm which must be applied by a clinical pathologist. Application of the score is complex as it requires the reader to visually exclude stromal and tumor-infiltrating immune cells across a full sample, retaining PD-L1 expression information for true tumor cells only. Methods: In this study, we utilized the PD-L1 [28-8] immunohistochemistry (IHC) assay to stain 10 NCSLC specimens. Stained slides were then imaged using the AT2 scanner (Leica Biosystems, Buffalo Grove, IL) scanner and analyzed using the Visiopharm Image Analysis platform. Results: The intricacy of TPS application is time consuming for the pathologist and introduces opportunity for inter- and intra-reader variability. Our novel image analysis approach utilizes artificial intelligence (AI) to automatically denote tumor nests and exclude tumor-infiltrating immune cells. The remaining true tumor cells may then be quantified for PD-L1 expression across the full biopsy, producing a TPS for each sample. Conclusions: This method allows for great accuracy and time-efficiency in providing a TPS than traditional methods, which may result in improved patient care.

An AI-based, automated workflow for identification and scoring of invasive tumors in Ki-67 stained breast cancer specimens

e15108

Background: Understanding the rate of tumor cell growth in breast cancer specimens may be indicative of disease aggressiveness, a tumor characteristic which can be used to make an informed treatment decision. The nuclear protein Ki-67 is increased in cells as they prepare to divide or proliferate and is therefore widely used as a proliferation marker for tumor progression. This degree of tumor cell proliferation, or the proliferative index, is commonly detailed in pathology reports shared with the patient care team. Methods: In this study, we utilized the Ki-67 [K2] immunohistochemistry (IHC) assay to stain 10 breast cancer specimens. Stained slides were imaged using the AT2 scanner (Leica Biosystems, Buffalo Grove, IL) and analyzed using the Visiopharm Image Analysis platform. Results: Previous efforts to assess Ki-67 positivity utilizing image analysis have relied on the use of a secondary stain or manual effort by the pathologist to exclude non-invasive tumor regions. These antiquated methods are costly to the lab as they require additional materials or valuable pathologist time. Our novel image analysis approach utilizes artificial intelligence (AI) to automatically denote non-invasive verses invasive tumor regions, which can then be used to quantify the Ki-67 proliferative index. Conclusions: This valuable tool will allow for greater accuracy, cost-savings, and time efficiency when analyzing breast cancer samples compared to traditional methods.

Histological disease activity in patients with microscopic colitis is not related to clinical disease activity or long-term prognosis

BACKGROUND

Microscopic colitis (MC) is a common cause of chronic watery diarrhea. Biopsies with characteristic histological features are crucial for establishing the diagnosis. The two main subtypes are collagenous colitis (CC) and lymphocytic colitis (LC) but incomplete forms exist. The disease course remains unpredictable varying from spontaneous remission to a relapsing course.

AIM

To identify possible histological predictors of course of disease.

METHODS

Sixty patients from the European prospective MC registry (PRO-MC Collaboration) were included. Digitised histological slides stained with CD3 and Van Gieson were available for all patients. Total cell density and proportion of CD3 positive lymphocytes in lamina propria and surface epithelium were estimated by automated image analysis, and measurement of the subepithelial collagenous band was performed. Histopathological features were correlated to the number of daily stools and daily watery stools at time of endoscopy and at baseline as well as the clinical disease course (quiescent, achieved remission after treatment, relapsing or chronic active) at 1-year follow-up.

RESULTS

Neither total cell density in lamina propria, proportion of CD3 positive lymphocytes in lamina propria or surface epithelium, or thickness of collagenous band showed significant correlation to the number of daily stools or daily watery stools at any point of time. None of the assessed histological parameters at initial diagnosis were able to predict clinical disease course at 1-year follow-up.

CONCLUSIONS

Our data indicate that the evaluated histological parameters were neither markers of disease activity at the time of diagnosis nor predictors of disease course.

Automated Quantification of sTIL Density with H&E-Based Digital Image Analysis Has Prognostic Potential in Triple-Negative Breast Cancers

Triple-negative breast cancer (TNBC) is an aggressive and difficult-to-treat cancer type that represents approximately 15% of all breast cancers. Recently, stromal tumor-infiltrating lymphocytes (sTIL) resurfaced as a strong prognostic biomarker for overall survival (OS) for TNBC patients. Manual assessment has innate limitations that hinder clinical adoption, and the International Immuno-Oncology Biomarker Working Group (TIL-WG) has therefore envisioned that computational assessment of sTIL could overcome these limitations and recommended that any algorithm should follow the manual guidelines where appropriate. However, no existing studies capture all the concepts of the guideline or have shown the same prognostic evidence as manual assessment. In this study, we present a fully automated digital image analysis pipeline and demonstrate that our hematoxylin and eosin (H&E)-based pipeline can provide a quantitative and interpretable score that correlates with the manual pathologist-derived sTIL status, and importantly, can stratify a retrospective cohort into two significant distinct prognostic groups. We found our score to be prognostic for OS (HR: 0.81 CI: 0.72-0.92 p = 0.001) independent of age, tumor size, nodal status, and tumor type in statistical modeling. While prior studies have followed fragments of the TIL-WG guideline, our approach is the first to follow all complex aspects, where appropriate, supporting the TIL-WG vision of computational assessment of sTIL in the future clinical setting.

Pitfalls in assessing stromal tumor infiltrating lymphocytes (sTILs) in breast cancer

Stromal tumor-infiltrating lymphocytes (sTILs) are important prognostic and predictive biomarkers in triple-negative (TNBC) and HER2-positive breast cancer. Incorporating sTILs into clinical practice necessitates reproducible assessment. Previously developed standardized scoring guidelines have been widely embraced by the clinical and research communities. We evaluated sources of variability in sTIL assessment by pathologists in three previous sTIL ring studies. We identify common challenges and evaluate impact of discrepancies on outcome estimates in early TNBC using a newly-developed prognostic tool. Discordant sTIL assessment is driven by heterogeneity in lymphocyte distribution. Additional factors include: technical slide-related issues; scoring outside the tumor boundary; tumors with minimal assessable stroma; including lymphocytes associated with other structures; and including other inflammatory cells. Small variations in sTIL assessment modestly alter risk estimation in early TNBC but have the potential to affect treatment selection if cutpoints are employed. Scoring and averaging multiple areas, as well as use of reference images, improve consistency of sTIL evaluation. Moreover, to assist in avoiding the pitfalls identified in this analysis, we developed an educational resource available at www.tilsinbreastcancer.org/pitfalls.

Report on computational assessment of Tumor Infiltrating Lymphocytes from the International Immuno-Oncology Biomarker Working Group

Assessment of tumor-infiltrating lymphocytes (TILs) is increasingly recognized as an integral part of the prognostic workflow in triple-negative (TNBC) and HER2-positive breast cancer, as well as many other solid tumors. This recognition has come about thanks to standardized visual reporting guidelines, which helped to reduce inter-reader variability. Now, there are ripe opportunities to employ computational methods that extract spatio-morphologic predictive features, enabling computer-aided diagnostics. We detail the benefits of computational TILs assessment, the readiness of TILs scoring for computational assessment, and outline considerations for overcoming key barriers to clinical translation in this arena. Specifically, we discuss: 1. ensuring computational workflows closely capture visual guidelines and standards; 2. challenges and thoughts standards for assessment of algorithms including training, preanalytical, analytical, and clinical validation; 3. perspectives on how to realize the potential of machine learning models and to overcome the perceptual and practical limits of visual scoring.

From Detection of Individual Metastases to Classification of Lymph Node Status at the Patient Level: The CAMELYON17 Challenge

Automated detection of cancer metastases in lymph nodes has the potential to improve the assessment of prognosis for patients. To enable fair comparison between the algorithms for this purpose, we set up the CAMELYON17 challenge in conjunction with the IEEE International Symposium on Biomedical Imaging 2017 Conference in Melbourne. Over 300 participants registered on the challenge website, of which 23 teams submitted a total of 37 algorithms before the initial deadline. Participants were provided with 899 whole-slide images (WSIs) for developing their algorithms. The developed algorithms were evaluated based on the test set encompassing 100 patients and 500 WSIs. The evaluation metric used was a quadratic weighted Cohen's kappa. We discuss the algorithmic details of the 10 best pre-conference and two post-conference submissions. All these participants used convolutional neural networks in combination with pre- and postprocessing steps. Algorithms differed mostly in neural network architecture, training strategy, and pre- and postprocessing methodology. Overall, the kappa metric ranged from 0.89 to -0.13 across all submissions. The best results were obtained with pre-trained architectures such as ResNet. Confusion matrix analysis revealed that all participants struggled with reliably identifying isolated tumor cells, the smallest type of metastasis, with detection rates below 40%. Qualitative inspection of the results of the top participants showed categories of false positives, such as nerves or contamination, which could be targeted for further optimization. Last, we show that simple combinations of the top algorithms result in higher kappa metric values than any algorithm individually, with 0.93 for the best combination.