Joseph S. Krueger

Evaluating tumor heterogeneity in immunohistochemistry-stained breast cancer tissue

Quantitative clinical measurement of heterogeneity in immunohistochemistry staining would be useful in evaluating patient therapeutic response and in identifying underlying issues in histopathology laboratory quality control. A heterogeneity scoring approach (HetMap) was designed to visualize a individual patients immunohistochemistry heterogeneity in the context of a patient population. HER2 semiquantitative analysis was combined with ecology diversity statistics to evaluate cell-level heterogeneity (consistency of protein expression within neighboring cells in a tumor nest) and tumor-level heterogeneity (differences of protein expression across a tumor as represented by a tissue section). This approach was evaluated on HER2 immunohistochemistry-stained breast cancer samples using 200 specimens across two different laboratories with three pathologists per laboratory, each outlining regions of tumor for scoring by automatic cell-based image analysis. HetMap was evaluated using three different scoring schemes: HER2 scoring according to American Society of Clinical Oncology and College of American Pathologists (ASCO/CAP) guidelines, H-score, and a new continuous HER2 score (HER2cont). Two definitions of heterogeneity, cell-level and tumor-level, provided useful independent measures of heterogeneity. Cases where pathologists had disagreement over reads in the area of clinical importance (+1 and +2) had statistically significantly higher levels of tumor-level heterogeneity. Cell-level heterogeneity, reported either as an average or the maximum area of heterogeneity across a slide, had low levels of dependency on the pathologist choice of region, while tumor-level heterogeneity measurements had more dependence on the pathologist choice of regions. HetMap is a measure of heterogeneity, by which pathologists, oncologists, and drug development organizations can view cell-level and tumor-level heterogeneity for a patient for a given marker in the context of an entire patient cohort. Heterogeneity analysis can be used to identify tumors with differing degrees of heterogeneity, or to highlight slides that should be rechecked for QC issues. Tumor heterogeneity plays a significant role in disconcordant reads between pathologists.

PEA-15 inhibits tumor cell invasion by binding to extracellular signal-regulated kinase 1/2.

Phosphoprotein enriched in astrocytes of 15 kDa (PEA-15) binds to extracellular signal-regulated kinase 1 and 2 (ERK1/2) mitogen-activated protein (MAP) kinases to alter ERK1/2 cellular localization and target preferences and binds to adaptors in the extrinsic cell death pathway to block apoptosis. Here, we report that PEA-15 protein expression is inversely correlated with the invasive behavior of breast cancer in an immunohistochemical analysis of a breast cancer progression tissue microarray. Short hairpin RNA-mediated inhibition of PEA-15 expression increased the invasion of PEA-15-expressing tumor cells in vitro, suggesting a causative role for PEA-15 in the inhibition of invasion. This causative role was confirmed by the finding that the enforced expression of PEA-15 in invasive tumor cells reduced invasion. The effect of PEA-15 on tumor invasion is mediated by its interaction with ERK1/2 as shown by the following: (a) PEA-15 mutants that fail to bind ERK1/2 did not inhibit invasion; (b) overexpression of ERK1 or activated MAP/ERK kinase (MEK) reversed the inhibitory effect of PEA-15; (c) when an inhibitor of ERK1/2 activation reduced invasion, PEA-15 expression did not significantly reduce invasion further. Furthermore, we find that the effect of PEA-15 on invasion seems to relate to the nuclear localization of activated ERK1/2. PEA-15 inhibits invasion by keeping ERK out of the nucleus, as a PEA-15 mutant that cannot prevent ERK nuclear localization was not able to inhibit invasion. In addition, membrane-localized ERK1, which sequesters endogenous ERK1 to prevent its nuclear localization, also inhibited invasion. These results reveal that PEA-15 regulates cancer cell invasion via its ability to bind ERK1/2 and indicate that nuclear entry of ERK1/2 is important in tumor behavior.

The Gold Standard Paradox in Digital Image Analysis: Manual Versus Automated Scoring as Ground Truth

CONTEXT - Novel therapeutics often target complex cellular mechanisms. Increasingly, quantitative methods like digital tissue image analysis (tIA) are required to evaluate correspondingly complex biomarkers to elucidate subtle phenotypes that can inform treatment decisions with these targeted therapies. These tIA systems need a gold standard, or reference method, to establish analytical validity. Conventional, subjective histopathologic scores assigned by an experienced pathologist are the gold standard in anatomic pathology and are an attractive reference method. The pathologists score can establish the ground truth to assess a tIA solutions analytical performance. The paradox of this validation strategy, however, is that tIA is often used to assist pathologists to score complex biomarkers because it is more objective and reproducible than manual evaluation alone by overcoming known biases in a humans visual evaluation of tissue, and because it can generate endpoints that cannot be generated by a human observer. OBJECTIVE - To discuss common visual and cognitive traps known in traditional pathology-based scoring paradigms that may impact characterization of tIA-assisted scoring accuracy, sensitivity, and specificity. DATA SOURCES - This manuscript reviews the current literature from the past decades available for traditional subjective pathology scoring paradigms and known cognitive and visual traps relevant to these scoring paradigms. CONCLUSIONS - Awareness of the gold standard paradox is necessary when using traditional pathologist scores to analytically validate a tIA tool because image analysis is used specifically to overcome known sources of bias in visual assessment of tissue sections.

AngioMap is a novel image analysis algorithm for assessment of plasma cell distribution within bone marrow vascular niche.

The ability to characterize distribution of neoplastic hematopoietic cells and their progenitors in their native microenvironment is emerging as an important challenge and potential therapeutic target in many disease areas, including multiple myeloma. In multiple myeloma, bone marrow (BM) angiogenesis is typically increased and microvessel density is a known indicator of poor prognosis. However, the difficulty of consistently measuring 3D vessels from 2D cut sections has previously limited the study of spatial distribution of plasma cells (PC) and their interaction with BM microenvironment. The aim of the study is to report a novel method to study myeloma cells spatial distribution within their hematopoietic niche context using readily available tissue sections and standard histology approaches. We utilized a novel whole-tissue image analysis approach to identify vessels, and then applied computational grown regions extended out from each vessel at 15, 35, 55, 75, and 100 &mgr;m to identify the spatial distribution of PC on CD34/CD138 double-stained core biopsy slides. Percent PC to total cells (TC) was significantly higher at <15 &mgr;m distance compared with those at 16 to 35, 36 to 55, 56 to 75, and 76 to 100 &mgr;m distance (P=0.0001). Similarly, PC/TC at <35 &mgr;m region was significantly higher compared with 36 to 55 (P=0.0001), 56 to 75 (P⩽0.0001), and 76 to 100 (P=0.0002) &mgr;m distances. The mean PC/TC differences in the spatial gradient of 36 to 55, 56 to 75, and 76 to 100 &mgr;m distance regions were not significant. Our findings suggest possible preferential advantage to neoplastic PC in the proximity of blood vessels compared with other hematopoietic marrow cells. We demonstrate the feasibility of analyzing the spatial distribution of PC, and possibly other hematopoietic/stem cells in their microenvironment, as characterized by the distance to vessels in BM using a novel image analysis approach.

Abstract 4582: Evaluating benefits of PD-L1 image analysis for the clinical setting

Tissue-based investigations can prove challenging due to complex tissue architecture and heterogeneous biomarker expression, visual and cognitive “traps” that affect interpretive precision, and subjective assessments that affect reproducibility. A major concern is that these challenges could increase the risk of failure for therapeutic/diagnostic co-development and clinical use, as the biomarker measurements continue to increase in complexity and require increasingly precise diagnostic cut-points. Image analysis tools have been developed to overcome some of the challenges for conventional anatomic pathology practices, capitalizing on the objectivity and computational power of a digital platform. A computer, however, lacks the cognitive ability and experience of a human to interpret tissue architecture and context. Flagship Biosciences’ computational Tissue Analysis (cTA™) platform integrates the power of our tissue Image Analysis (tIA™) technology with the contextual experience of an anatomic pathologist to produce robust, precise, quantitative results that demonstrate biomarker content in the tissue context. Flagship Biosciences envisions the integration of our cTA™ technology into a computer-aided clinical pathology workflow as a method to improve the precision of scoring for even some of the most challenging tissue-based biomarker measurements. In a proof-of-concept study, we evaluated the performance of manual versus digital scoring approaches in a cohort of non-small cell lung carcinoma (NSCLC) samples stained with the IHC protocol for the PD-L1 PharmDx 28-8 complementary diagnostic. A comparison of the 2 modalities demonstrated that in nearly all cases, the within sample standard deviation of the cTA™ digital score results was less than the manual score (median inter-pathologist %CVs were reduced from 124.9% to 7.8% and intra-pathologists from 65.4% to 7.6% for manual and digital scores, respectively). As an additional exploratory examination, the effect of heterogeneity on PD-L1 interpretation was also investigated. Pathologists evaluated the same whole tissue slides within 5 high powered fields (HPFs) using both manual and cTA™ -derived scoring. Results demonstrated that the use of cTA™ provides improves agreement between HPF and whole slide assessments (absolute difference between the manual scores from HPF to whole slide were larger than the absolute differences for the digitally derived scores, at 3.14% and 8.27%, respectively). Taken together, these studies demonstrate that the use of cTA™ can significantly reduce variability in PD-L1 scoring, as compared to a manual scoring approach. Citation Format: Staci Kearney, Joshua Black, Famke Aeffner, Joshua Black, Luke Pratte, Joseph Krueger. Evaluating benefits of PD-L1 image analysis for the clinical setting [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4582. doi:10.1158/1538-7445.AM2017-4582

Abstract 2225: Image analysis-based PD-L1 companion and complementary diagnostics

The PD-1 pathway, comprised of the immune cell co-receptor Programmed Death 1 (PD-1) and its ligands PD-L1 and PD-L2, mediates local immunosuppression in the tumor microenvironment. Immune checkpoint modulators are designed to block the local immunosuppression caused by this pathway. The FDA approved anti-PD-1 antibody therapies Opdivo® (nivolumab; Bristol-Myers Squibb) and Keytruda® (prembrolizumab; Merck & Co) rely on PD-L1 immunohistochemistry (IHC) in vitro diagnostic (IVD) tests to determine the PD-L1 status in patients in non-small cell lung cancer (NSCLC), in order to predict response to these drugs. The current complementary diagnostic for Opdivo® (Dako 28-8 PharmDx®) relies on a pathologist scoring paradigm which considers any patient with ≥1% positive tumor cells an optimal candidate for Opdivo® treatment. However, overall survival (OS) is further increased when patients have ≥5% or ≥10% PD-L1 positive tumor cells. This scoring approach is vastly different than the PD-L1 scoring approach used in the Keytruda® companion diagnostic (Dako 22C3 PharmDx®), which utilizes a ≥50% positive tumor cells value to predict a positive Overall Response Rate (ORR; OS not yet determined). Thus, the 28-8 test for Opdivo® utilizes a more precise approach than the 22C3 test for Keytruda®, and requires a more calibrated scoring approach. This calibrated approach for Opdivo® requires the difficult challenge of pathologists reliably distinguishing membrane staining to define the fine gradations of 1%, 5% and 10% PD-L1 positive neoplastic cells. To best meet this challenge, we developed a digital Tissue Image Analysis (TIA) solution which enabled accurate, unbiased quantification of PD-L1 on a cell-by-cell basis to classify the percentage positive tumor cells in patients with high granularity. Using Flagship9s proprietary CellMapTM algorithm, we evaluated 40 formalin-fixed paraffin-embedded (FFPE) NS-NSCLC samples which were stained using the Dako 28-8 PharmDx® PD-L1 IHC test. The TIA strategy digitally separated tumor cells from other cell types, and quantified membrane staining intensity according to a consistent threshold. The performance of the resulting IHC-TIA assay was evaluated in the context of a CLIA validation study performed by Flagship. The results demonstrated equivalency to the manually scored IVD reference standard; however, the TIA scoring of this assay provided consistent, unbiased, and more detailed scoring of PD-L1 stained tissues for determining the patients with ≥1, ≥5, and ≥10% PD-L1 positive tumor cells with greater confidence than a manual scoring approach. Moving forward, these TIA tools can be utilized to assess PD-L1 positive cell frequencies with greater reliability and granularity to identify optimal treatment cutpoints for these and other PD-L1 IHC tests used to predict response to PD-L1/PD-1 inhibitors. Citation Format: Joseph S. Krueger, Roberto Gianani, Brooke Hirsch, Stefan Pieterse, Famke Aeffner, David Young. Image analysis-based PD-L1 companion and complementary diagnostics. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2225.

Abstract 4162: Identifying T lymphocytes in IHC-stained tissues independently of CD3+ staining using morphometric features extracted by image analysis

Assessments of leukocyte populations in the context of cancer tissues are typically determined by staining for leukocyte subtype markers in formalin fixed tissues. This requires the identification, categorization, and localization of multiple leukocytes within tissue context utilizing multiple markers. Meeting these demands are challenging due to technical constraints on the number of individual markers which can be visualized or scored in a single slide, and the complexity of staining observed. The use of multispectral imaging with fluorescent multiplexed markers has better enabled the assessment of multiple markers in a single tissue section. However, these wet chemistry and image capture technologies are very complex, and can only be successfully implemented with significant laboratory infrastructure, specialty equipment, and experienced resources. For these reasons, this approach is not widely implementable in the clinical pathology laboratory setting, where such tests are oriented to a companion diagnostic utility. Approaches which rely on widely adopted chromogenic immunohistochemistry (IHC) staining are preferred in the clinical laboratory setting, but the number of assayable markers is limited to 1-3 unique markers in a single tissue. In order to create a useful approach that could be implemented in existing clinical laboratory workflow, Flagship Biosciences has developed an approach for deriving the complex endpoints often necessary in immuno-oncology studies which rely on 1-3 chromagenic stains and computer interpretation of the tissue using only hematoxylin counterstain to identify T-lymphocytes. In a proof-of-concept study, we utilized our Tissue Image Analysis (TIA) tools to identify morphometric parameters which could identify T-lymphocytes, independent of staining for the T-lymphocyte marker CD3. A cohort of non-small cell lung cancer (NSCLC) tissues was stained by CD3 IHC, and both CD3 and isotype-stained tissues were analyzed with Flagship9s CellMap™ software to capture the morphometric and staining features of cells in the tissues. The morphometric features which characterized CD3+ cells were used to approximate the T-lymphocyte population frequency in the isotype-stained tissues. This T-lymphocyte classification scheme was defined based on hematoxylin staining alone, and accuracy of T-lymphocyte classification was verified by CD3 staining. Based on this study, the method described herein could be utilized to reasonably estimate the frequency of T-lymphocyte subsets (e.g. CD4+, CD8+, etc.) or different marker-positive leukocyte (e.g. macrophages) subsets relative to the total T-lymphocyte population without an additional T-lymphocyte marker such as CD3. The approach could, therefore, provide an added dimension of analysis for tissues stained by IHC without adding complexity to the wet assay by necessitating a marker for T-lymphocytes. Citation Format: Nathan T. Martin, Joshua Black, Famke Aeffner, Logan Cerkovnik, Jasmeet Bajwa, Staci J. Kearney, Crystal Pulliam, A.J. Milici, Joseph Krueger. Identifying T lymphocytes in IHC-stained tissues independently of CD3+ staining using morphometric features extracted by image analysis. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4162.

Abstract C108: Quantifying PD-L1 spatial distribution signatures for patient selection approaches

Inhibitors of inflammatory checkpoints, such as PD-L1 inhibitors, have demonstrated great promise in preclinical and clinical studies. This therapeutic paradigm focuses on controlling natural inflammatory checkpoints to stimulate an elevated inflammatory response against the tumor to increase anti-tumor inflammatory cell infiltrates in the tumor microenvironment or decrease inflammatory suppressor infiltrates. The proteins which control these processes can be found in the tumor cells, cells in the tumor micro-environment (TME), or in both locales. Positive cells are often assessed in a qualitative or semi-quantitative manner using immunohistochemistry and evaluation of a limited number of representative microscopy fields across a particular tissue compartment (tumor vs stroma) or the whole tissue area. However, the locale of the inflammatory suppressors such as PD-L1 may be more revealing than estimating the tumor-wide dispersion of an inflammatory cell type. Unfortunately, the intricate spatial relationships and the often complex distribution of inflammatory cells in tissues pose significant challenges for a meaningful evaluation. We have developed an approach which can quantify these spatial relationship in a contextual, biologically meaningful score. Immunohistochemistry staining for PD-L1 in whole lung cancer tissue sections was performed, and our CellMap software was used to assess inflammatory cell distribution in the whole tissue sections. PD-L1 positive cells were quantified relative to: 1) the total number of cells in the tumor and stromal tissue compartments, and 2) the number of cells within a distance from the tumor/stroma interface. Interestingly, several unique PD-L1 distribution patterns relative to the tumor/stroma interface were observed in the sample cohort analyzed. Quantifying the distribution of PD-L1 positive cells as a function of distance from the tumor/stroma interface revealed distribution signatures, which could be used to differentiate between samples. In contrast, this differentiation of the same samples was not possible when PD-L1 cells were assessed relative to the total number of cells. This study provided a novel method for assessing inflammatory cell type spatial distribution relative to a tissue feature, the tumor/stroma interface. The data suggested that unique spatial patterns of inflammatory cell type distribution could be used to uniquely stratify patients compared to existing quantitative methods. Taken together, this proof-of-concept study demonstrates a unique quantitative assessment of inflammatory cell infiltrates in tumors that could be used to gain new insights into inflammatory cell type distributions and interactions in tumors, inflammatory cell spatial responses to oncology therapies, and novel patient selection criteria for traditional and immuno-oncology therapeutics. Citation Format: Joseph S. Krueger, Nathan Martin, Anthony Milici, Famke Aeffner. Quantifying PD-L1 spatial distribution signatures for patient selection approaches. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr C108.

Abstract C109: Quantitative analysis of multiple subtypes of immune system cells in cancer tissues

Current cancer biology acknowledges the key role of the immune system in tumor biology, and promise for the modulation of immune system in cancer treatment. The composition of the inflammatory cell populations in tissues is reflective of the overall state of the Tumor Micro-Environment (TME), and the identification of distinct inflammatory cell types may hold prognostic or predictive value. Immunohistochemistry allows for reliable identification of the cell constituents to facilitate analysis of the TME while remaining in the tissue context. Establishing a quantitative paradigm for inflammatory cell types and subtype profiling requires unbiased and automated whole-tissue based quantitation methods, which are capable of spatial integration of multiple inflammatory cell markers across the whole tissue. While single slide fluorescent multiplex approaches can address this need, the use of difficult-to-implement wet assay strategies involving multiplexing 6-8 fluorescent markers on the same tissue section are difficult to implement in a global clinical diagnostic lab setting. To answer this need, we combined novel advents in Tissue Image Analysis (TIA) to integrate spatial expression of serial-section stained whole tissue clinical lung cancer specimens. In this proof-of-principle study,we were able to superimpose specific locations of individual cell types onto 6 serial sections and evaluate different inflammatory cell types. We used serial sections of clinical lung specimens stained for six immune phenotypic markers (CD68, CD4, CD8, CD33, FoxP3, and CD11b) to illustrate a repertoire of inflammatory cell types. Our proprietary CellMap algorithm was utilized to identify, enumerate, and determine the precise location of individual inflammatory cells in tissues on cell-by-cell basis in the tumor microenvironment (TME). Our proprietary FACTS (Feature Analysis on Consecutive Tissue Sections) approach was used to integrate the spatial expression of individual markers onto a reference HE 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr C109.

Abstract 3391: Companion diagnostic strategies specific to antibody therapies

Proceedings: AACR 106th Annual Meeting 2015; April 18-22, 2015; Philadelphia, PA One premise of antibody-drug conjugates (ADC) is that the bound mAb-antigen complex on the cell surface will internalize and be metabolized by lysosomal proteases to release the free drug. Thus, the efficacy of an ADC is dependent not only on the presence of cell surface antigens, but also an active system of receptor turnover and receptor-mediated endocytosis. Thus, a predictive assay for patient response would ideally account for both the degree of cell surface expression of the target, as well as cytoplasmic presence of the target to quantify a surrogate for receptor turnover and internalization. Immunohistochemistry based assays (IHC) are best suited to address these questions, as it is the only method which provides the ability to measure both membrane and cytoplasm expression of the target simultaneously within archival FFPE biopsies. However, the biological mechanisms behind receptor internalization and turnover have not been elucidated for novel therapeutic targets. In most cases, an IHC assay is utilized to evaluate these measures, without prior advance knowledge of how these measures are suitable for patient selection. Unanticipated difficulties in tissue interpretation, such as low apparent expression of the target, occlusion of membrane staining by cytoplasmic staining, or heterogeneity in staining often lead to failure in determining a correct patient stratification approach. In order to investigate patient selection strategies for ADCs, we have invented several proprietary approaches for measuring critical properties of the therapeutic target on the cell surface or inside the cell which can be used to understand and predict efficacy to an ADC using FFPE biopsies. These quantitative pathology approaches are based on image analysis approaches which been designed specifically for ADC CDx programs to develop a pathology based scoring system which can be predictive of ADC response: 1) Accurately quantifying low levels of cell surface target expression; 2) Defining cell surface target expression independent of cytoplasmic expression; 3) Overcoming staining heterogeneity; and 4) Determining the correct staining thresholds for quantification. These image analysis based approaches can be used to define and evaluate a scoring approach, train pathologists, assess objective performance, and best determine a cutpoint approach using statistical approaches. These image analysis based tools can be used to create a manual scoring paradigm for an IHC assay or can be incorporated into a medical device directly to support the PMA effort. Incorporation of these novel tools will enable ADC developers to create efficacy and patient stratification paradigms which incorporate the critical biological endpoints unique to ADCs. Citation Format: Joseph S. Krueger, David Young, Holger Lange, Steve Potts. Companion diagnostic strategies specific to antibody therapies. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3391. doi:10.1158/1538-7445.AM2015-3391