Abstracts & Presentations

The pursuit of effective cancer treatment is greatly challenged by the heterogeneity of the tumor microenvironment. Traditional histopathology, which relies on H&E staining, is inadequate in differentiating between patients who will benefit from treatment and those who will not. This leads to unnecessary treatment toxicities due to overtreatment or suboptimal treatment outcomes because of undertreatment, highlighting a critical unmet clinical need to accurately identify treatment responders and spare non-responders from the adverse side effects of unnecessary therapy.

Triple negative breast cancer (TNBC), a highly aggressive and heterogeneous breast cancer subtype, is particularly challenging to treat. Neoadjuvant chemotherapy, administered before surgery, is the standard treatment for TNBC and yields a mere 40% response rate. With the recent addition of immunotherapy to treatment regimens, pCR rates have further improved by 10-15%. Thus, about 50% of patients experience severe side effects from chemotherapy without any therapeutic benefits.

There is a growing interest in leveraging artificial intelligence, including deep learning and machine learning techniques, to predict treatment response from whole slide images of H&E-stained tissue biopsies. In this talk, I present how machine learning can unlock the power of a biopsy H&E image by analyzing the tumor microenvironment to predict the response of TNBC tumors to neoadjuvant chemotherapy with high accuracy. We are also assembling cohorts treated with a combination of chemotherapy and immunotherapy, reflecting the current standard of care in TNBC, and will be training our algorithms on these cohorts. Our ongoing work focuses on enhancing the prediction models by integrating spatial proteomics data, allowing for an even more nuanced understanding of the tumor microenvironment and potentially leading to the development of algorithms that predict molecular biomarkers directly from H&E slides. This innovative approach harnesses digital pathology and spatial biology, signifying a leap forward into precision oncology and offering hope for better, more targeted cancer treatment strategies.
Back to Top of Page

Over the past decades our understanding of the biology of melanocytic neoplasms has changed significantly. This is mainly due the characterization of underlying driver mutations in benign nevi and melanoma and the recent demonstration of fusion genes in Spitz melanocytic tumors and novel, recently described entities, such as TRIM11-rearranged tumors. These findings have led to more confident diagnosis of melanocytic neoplasms in which the judicious use of immunohistochemistry plays an important role. In this short overview the main immunohistochemical markers and their correct application will be discussed with emphasis on diagnostic pitfalls. Melanocytic markers can be broadly grouped into those highlighting melanocytic differentiation (S100, SOX10, Melan A/Mart-1, HMB45) and those of prognostic value (PRAME, p16, Ki-67). Other markers are used to detect underlying mutations or gene fusions (BRAF-V600E, NRAS Q61R, ALK, NTRK, ROS). PRAME has gained particular popularity, as it is thought to help in separating benign form malignant melanocytic proliferations. There are however significant limitations and associated pitfalls with the use of PRAME.
Back to Top of Page

Epithelioid vascular neoplasms continue to pose a significant diagnostic challenge. They range from outright malignant to benign disease but there is significant morphologic overlap. In addition to subtle histopathologic features, immunohistochemistry plays an important part in the correct diagnosis. The immunohistochemical markers CD31, CD34 and more recently ERG are important to confirm the endothelial cell differentiation of these tumors and exclude non-vascular mimics. As we have gained a better understanding of the underlying molecular divers in vascular tumors over the past decade, immunohistochemistry can also be used in the stratification of epithelioid vascular tumors. The main entities included are epithelioid hemangioma, cutaneous epithelioid angiomatous nodule, pseudomyogenic hemangioendothelioma, epithelioid hemangioendothelioma and epithelioid angiosarcoma. The salient histopathologic features and immunohistochemical profile will be discussed in the context of the clinical presentation and behavior and underlying molecular aberrations.
Back to Top of Page

The digital age is upon us, whether we like it or not. Its presence in the daily work practices of cellular pathology laboratories is being increasingly seen, for example in the use of digitised images to enable reporting remotely across pathology networks (telepathology), or for expert opinion referral, or to improve internal workflows within departments.

Hand-in-hand with this we are beginning to see the introduction of digital image analysis (DIA) applications as aids to clinical decision making and as tools to improve standardisation and reproducibility of diagnostic tests.

External quality assessment programmes designed to examine the technical quality of tests undertaken in cellular pathology laboratories, such as those offered by UK National Quality Assessment for Immunocytochemistry and In-Situ Hybridisation (UK NEQAS ICC & ISH) have traditionally used manual visual assessment of the participants’ submitted stained slides to qualitatively separate the good from the poor.

Traditionally qualitative assessment has served us well, and it continues to do so where the questions being addressed by the testing have a non-quantitative answer i.e. when we are using our primary antibodies as diagnostic biomarkers. But increasingly we are being asked not only if a particular antigen is present, but also how much of it is present. This is true particularly in the field of precision medicine where ICC tests are being used as predictive for targeted drug therapies. DIA is particularly suited to answering questions involving quantification and its introduction is seeing rapid progress.

At UK NEQAS, we have been developing and incorporating DIA into a number of our programmes looking at companion diagnostics for a number of years, and in this talk I will draw on our now extensive experience to illustrate how they are improving the services we are providing to our participants.
Back to Top of Page

Demonstration of a protein called Ki-67, that is present in all phases of the cell-cycle except G0 using a monoclonal antibody (MIB1) was first described by Gerdes and his co-workers in Kiel, Germany in1984. The use of antibodies to Ki-67 continues to this day to be the most widely used method for identifying proliferating cells in normal and diseased formalin-fixed paraffin-embedded tissue.

We will consider some of the evidence supporting the use of Ki-67 assessment in breast cancer, specifically with regard to its use as a prognostic indicator of disease recurrence in oestrogen-receptor-positive (ER-positive) early breast cancer. Dowsett and colleagues were amongst the first to publish on short-term changes in Ki-67 expression during neoadjuvant endocrine treatment. They were able to show that proliferation was decreased by endocrine treatment in the majority of patients (~85%), and that change was positively associated with ER and progesterone receptor (PgR) levels. Importantly, in a later study of the same trial they found that higher Ki-67 expression after 2 weeks of endocrine therapy was statistically significantly associated with lower recurrence-free survival. These observations led to the much larger POETIC trial, which recently reported confirmatory results.

Despite undisputed utility, problems in the standardisation and reproducibility of Ki-67 measurement, especially when assessed across single cut-points have hampered its widespread integration into routine clinical practice. The International Ki67 in Breast Cancer Working Group (IKBCWG) was established to ‘devise a strategy to harmonize Ki67 analysis and increase scoring concordance’ and has published a number of papers on its findings. As a result of the group’s efforts there is now available a method for manual scoring of Ki-67 expression that shows very good inter-rater reliability (Interclass Correlation Coefficient (ICC) = >0.8) when used in breast cancer samples similar to those seen in routine clinical practice.

We will consider the evidence for an alternative approach, which integrates Ki-67 together with IHC-derived information on ER and PgR expression levels and a number of well-established clinicopathologic parameters into a single prognostic score (termed IHC4+C) has been shown to produce results which are robust and reliable. It has been analytically and clinically validated in a number of studies. In a comparative study using IKBCWG data it achieved almost perfect agreement (ICC = 0.99).
Back to Top of Page

The post-analytical phase is the final phase of the total test pipeline and involves the application of the following processes to test results:
• evaluation;
• validation;
• release.
We will consider the post-analytical phase in the context of precision medicine. We will discover the role cellular pathology has played, and in particular how immunohistochemistry-based companion diagnostics have enabled individualised patient care.
Back to Top of Page

If you spend enough time around Immunohistochemistry stains, you can’t help but notice some very weird things. Is it real staining, is it some random occurrence or is it just another unexplained mystery? In this presentation we will delve into some weird events from the Princess Alexandra Hospital and try to work out if there is a plausible explanation for them. It may even be that we have to reassess what we thought was fact and start believing in the fictitious.
Back to Top of Page

As a participant in the RCPAQAP IHC module, it can be difficult to understand what is involved in assessing your slides. When a result comes back to the laboratory, opinions may range from “I totally agree with the expert comments and score”, to “Were they looking at the right slide?” This presentation will attempt to clear up a few myths about assessments as well as include the audience in a mock assessment to get a better understanding of the process. Digital images will be used from the most recent RCPA QAP IHC assessment.
Back to Top of Page

Ki-67 is a nuclear protein serendipitously discovered by monoclonal antibody selection in the early 1980s.  While it has been applied for decades in the context of breast cancer as a putative prognostic and, more recently, predictive,  biomarker, even after all this time there is incomplete agreement as to the validity of the immunohistochemical assays employed for Ki-67 assessment, given possible effects of the disparate methodologies employed and possible confounding pre-analytical, analytical and interpretive variables.  In this talk, the history of Ki-67 and the problems, particularly with the analytical and interpretive variables, are highlighted through a selective review of the published literature.  The contributions of the International Ki-67 Breast Cancer Working Group are highlighted, and in particular, the recommendations made by this group are reviewed.  The potential of Ki-67 as a biomarker for breast cancer has not yet been fully realized. Still, an understanding of the power and limitations of the methods of Ki-67 assessment is important if this biomarker can realize its potential.
Back to Top of Page

From its inception as a clinical tool in the 1980s, applications of immunohistochemistry have focussed on the identification of organ-specific markers that might help in the identification of the primary site of carcinomas presenting as metastases.  In this overview talk, I will focus on the most important of these markers, providing a panoramic view of the moving landscape over the past few decades of immunohistochemical markers of lung, GI, and breast carcinomas.  Drawing on my experience and the published literature, I will identify the persistent limitations of many of these markers and problems with the published literature.  I will eschew the use of algorithms or flow charts, which I don’t endorse as a method of analysis, and instead outline how they can best and most intelligently be employed.
Back to Top of Page

Dual immunohistochemical staining is being increasingly used in diagnostic pathology where two different antigens can be detected in a tissue sample. The dual staining can be sequential or simultaneous. Sequential double staining is the preferred method when epitopes are in different cellular compartments or different cell types (e.g. nuclei and cytoplasm). Simultaneous double staining is the technique when the antibodies that are used are from different hosts. There is an increasing demand to perform more ancillary techniques such as biomarker testing using immunohistochemical stains and molecular assays for tumour classification, prognostication and targeted therapy. Preserving valuable tissue is hence of paramount importance. Dual staining has the advantage of saving time, reagents and tissue. There are potential savings through consumables, scientists, and pathologists’ time. It also brings with it a significant impact on environmental sustainability. The combination of dual stains performed over two years in our laboratory will be discussed.
Back to Top of Page

Immunohistochemistry is a powerful tool with numerous applications in research and diagnostics, it also comes with its challenges and limitations that need to be carefully considered and addressed.
Immunohistochemistry (IHC) can be both a friend and a foe depending on the context and application. Examples of cases in both categories will be discussed and associated pitfalls will be addressed.
Back to Top of Page

Interpretation of medical renal biopsies generally requires histological examination by light microscopy (LM), immunohistochemistry, and electron microscopy (EM). Immunofluorescence on frozen tissue is the gold standard immunohistochemical technique for evaluation of immune deposits in the kidney. This technique is an integral part of final renal diagnosis. Apart for this immunohistochemistry has been used in medical renal biopsies both as diagnostic as well as research. The routine diagnostic IHC stains are SV40 for polyoma virus, C4d and PLA2R. This helps nephrologists to provide right therapy to their patients and subclassify the disease. There is significant amount of work happens at research level to understand the underlying pathogenesis of renal disease which can be preventable or targetable treated.
Back to Top of Page

Immunohistochemistry has become an indispensable ancillary study in the identification and classification of tumours. It has got diagnostic and therapeutic implications. The diagnostic accuracy has significantly improved because of continues discoveries of tissue-specific biomarkers and development of effective panels. It has become a standard of practice to include results of therapeutic IHC in pathology reports. IHC usage in diagnostics is invaluable; however, the genetic and therapeutic significance of biomarker immunostaining has become equally relevant. In the emerging era of personalized medicine, IHC continues to serve a valuable function, complementing and enhancing other molecular techniques.
Back to Top of Page

The use of control material is essential to ensure the integrity, reliability and reproducibility of all stains in immunohistochemistry. Without it, a reporting pathologist will never be able to have full confidence that a test has worked and that the cells demonstrated are the intended target.

In my time in the IHC department at Monash Health, the role of this control material has not changed, but the way in which it is sourced, used, documented and stored has changed enormously.

This presentation will cover some of the key aspects of that time, and will include:
Sourcing, selection and storage
• What is a good choice for future control material, where to get it from and how to make the most of it into the future.
Documentation
• How much information is required to ensure traceability and what are the responsibilities of the testing institution.
IPA processing changes
• Making large scale changes require a complete shift in materials, but this cannot be achieved overnight.
Multi-tissue controls
• Ideally, one control block will be suitable for all antibodies, but that is never the case.
• How we manage our resources is a key factor in control longevity
Fresh Vs Pre-cut
• Minimising the number of controls that must be cut at the time of sectioning is key to resource management and can affect through-put.
The presentation will also discuss requirements for dual-stain control material, immunofluorescence control material and storage and a NATA assessment finding.
Back to Top of Page

The samples that are received in any histology laboratory all share one key attribute.

They are all irreplaceable.

This one factor cannot be overstated. From a fragment of colonic mucosa to a skin ellipse, or a brain biopsy or paediatric tumor core, we must do our best to ensure that the tissue sample is well cared for and appropriately handled to allow us, the medical scientist, to be part of the team that provides the best outcome for the patient.

So many aspects of the health care model come together to provide this in the pre-analytical phase of the histology laboratory.

This presentation will discuss some aspects of pre-analytical care of the histology tissue sample including:

Fixation and specimen handling
• What can go wrong with a small fragment of tissue
Registration
• The importance of a correctly identified sample
Documentation of errors.
• How we at Monash Health record deficiencies and remedial actions
There will be a key focus on the needs of the immunohistochemistry department and the specific requirements that arise from IHC staining and interpretation.

It will end with a timely case study from our laboratory in which a tissue sample was compromised in the pre-analytical phase, how this was recognized and what actions were taken to overcome the problems that arose.
Back to Top of Page

The color intensity of immunohistochemical (IHC) slides strongly depends upon staining. Even well-controlled workflows may lead to color intensity differences between stain batches. Furthermore, in-depth workflow regularization between different laboratories is arduous. Small differences in staining intensity can significantly affect the interpretation of IHC slides, particularly with membrane stains. Scanners also produce color differences in Whole Slide Images (WSI). These variables affect an automated image analysis for IHC assessment. Therefore, stain evaluation by WSI needs to be standardized for accurate assessment. We developed methodologies for staining and color calibration of scanners for WSI-based IHC assessment using IHC calibration slides. In this study, we used the HER2 IHC stain on breast cancer cases because intensity is critical for pathologist interpretation. Design: We used IHC calibration slides that include a microbeads-based calibrator and four breast cancer excisions with HER2 IHC scores of 0, 1+, 2+, and 3+. WSI of these slides are used to quantify the staining intensities and to correct for the variations in staining. The slides are also used to scanner calibration, as the quantified intensities reflect the characteristics of each scanner. Color calibrations were performed using a color chart slide. We obtained 6 slides for each IHC calibration slide by staining on 6 different dates. All slides were digitized by two scanners, Nano Zoomer S60 (Hamamatsu　Photonics K.K., Japan) at 40x (0.23um/pixel) and PANNORAMIC 250 Flash III (3DHISTECH Ltd., Hungary) at 20x (0.18um/pixel). Accuracy was assessed after performing different combinations of staining and scanner calibration. Results: The staining calibration provided accurate results. However, there are still failures in the result without scanner calibration. When both staining and scanner calibration was applied, result showed a concordance of 100% with pathologist’s assessment. Both the color chart and the IHC calibrator effectively worked for the scanner calibration. Conclusions: Our method helps with the stain and color calibration and standardization for assessment using WSI of HER2 IHC. In the future, we will continue to improve the calibration slides to establish efficient and precise testing methods. Further validation will be conducted by using clinical slides and slides stained at different institutions.
Back to Top of Page

Accurate assessment of HER2 overexpression by immunohistochemistry (IHC) and gene amplification by in situ hybridization (ISH) is critical for targeted therapy in breast cancer patients. HER2 Dual ISH requires time-consuming quantification and is subjective. Computational analysis methods offer practical solutions to help pathologists with quantitative and qualitative studies and provide more objective and consistent results. We have previously optimized a scanning protocol for Dual ISH . In this study, we use this protocol to evaluate two different AI-integrated systems to confirm the reliability of the results. 26 de-identified invasive breast carcinoma cases with prior HER2 IHC and FISH manual results were evaluated. Each case was classified into ASCO/CAP ISH Group 1 to 5. H&E, IHC, and Dual ISH slides were scanned by ROCHE DP200 at 0.23 µm/pixel and 3D HISTECH Panoramic p250 at 0.17 µm/pixel 3 z-stack. Selected ROIs in WSIs were analyzed by two systems (Navify DP and Shimaris). Shimaris is an in-house standalone application that can use any scanner’s data, and the size and number of ROIs are adjustable. Navify DP is a cloud-based image management system with automated assessments such as HER2 IHC and Dual ISH using one fixed ROI. The concordance of HER2 gene status between scanner-dependent and system-dependent were determined and compared with manual FISH results. Manual Dual ISH results counted on WSIs match the results counted microscopically. HER2 determination results on WSIs by AI-integrated applications show 100% concordance with manual FISH results. The Navify DP system gives results very fast, one fixed size ROI is enough to reach 20 nuclei. ISH Group concordance with manual FISH were under 60% in all results. 20 nuclei were counted for each case, except borderline cases (40 nuclei). Both systems worked well, are user-friendly and time-saving for HER2 assessment. Based on the results of our another previous study which gave the best results with 80 nuclei, the agreement rate in ISH groups will be improved by counting more nuclei such as 80.
Back to Top of Page