e14052 Background: The need for accurate pathological identification and quantitation of prostate cancer (PC) following neoadjuvant treatment with androgen deprivation therapy (ADT) and androgen receptor antagonists is increasing as PC treatment continues to evolve. In clinical practice, pathological assessment of residual tumor is a tedious and time-consuming process due to the volume of tissue from radical prostatectomy (RP). In addition, neoadjuvant treatments can greatly alter both benign and neoplastic prostate tissue morphology making the pathology assessment difficult for even specialized pathologists. Paige Prostate 1.0 is a clinical-grade artificial intelligence (AI) system for PC detection. It was trained and evaluated in over 50,000 prostate biopsy slides with validation across more than 800 institutions worldwide using multiple slide scanners. Methods: We evaluated the performance of Paige Prostate 1.0 at identifying prostatic tumor on 64 hematoxylin and eosin stained slides exhibiting neoadjuvant treatment effect from apalutamide, enzalutamide, and/or ADT. Results: Analysis of the receiver operating characteristic curve demonstrated an area under the curve of 0.96. Using the Paige Prostate 1.0 operating point, it achieved a sensitivity of 91% and a specificity of 94%, corresponding to the correct identification of challenging treated morphology in 59/64 slides using expert pathologists as the reference. False negative cases were typically represented by atypical small acinar proliferation that required expert pathological consensus confirmation. Conclusions: To our knowledge, this is the first AI based evaluation of residual disease in PC with hormone neoadjuvant therapy. Paige Prostate 1.0 effectively identified tumor despite treatment effects. Future work will include optimization of Paige Prostate 1.0 by training with RP specimens from a larger cohort of appropriate samples, as well as precise measurement of residual tumor burden to further improve its accuracy and reproducibility. Paige prostate residual disease detection 1.0 has the potential to impact emerging clinical practice at the patient level and to complement the pathological assessment of RPs in global phase 3 clinical trials, such as PROTEUS, in a standardized, reproducible, and robust way.

We analyzed potential biomarkers of response to ibrutinib plus nivolumab in biopsies from patients with diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL), and Richter's transformation (RT) from the LYM1002 phase I/IIa study, using programmed death ligand 1 (PD-L1) immunohistochemistry, whole exome sequencing (WES), and gene expression profiling (GEP). In DLBCL, PD-L1 elevation was more frequent in responders versus nonresponders (5/8 [62.5%] vs. 3/16 [18.8%]; p = 0.065; complete response 37.5% vs. 0%; p = 0.028). Overall response rates for patients with WES and GEP data, respectively, were: DLBCL (38.5% and 29.6%); FL (46.2% and 43.5%); RT (76.5% and 81.3%). In DLBCL, WES analyses demonstrated that mutations in RNF213 (40.0% vs. 6.2%; p = 0.055), KLHL14 (30.0% vs. 0%; p = 0.046), and LRP1B (30.0% vs. 6.2%; p = 0.264) were more frequent in responders. No responders had mutations in EBF1, ADAMTS20, AKAP9, TP53, MYD88, or TNFRSF14, while the frequency of these mutations in nonresponders ranged from 12.5% to 18.8%. In FL and RT, genes with different mutation frequencies in responders versus nonresponders were: BCL2 (75.0% vs. 28.6%; p = 0.047) and ROS1 (0% vs. 50.0%; p = 0.044), respectively. Per GEP, the most upregulated genes in responders were LEF1 and BTLA (overall), and CRTAM (germinal center B-cell–like DLBCL). Enriched pathways were related to immune activation in responders and resistance-associated proliferation/replication in nonresponders. This preliminary work may help to generate hypotheses regarding genetically defined subsets of DLBCL, FL, and RT patients most likely to benefit from ibrutinib plus nivolumab.

A phase 1/2a study (LYM1002 [EudraCT 2014-005191-28]) of ibr (560 mg once daily) + nivo (3 mg/kg on a 14-day cycle) demonstrated acceptable safety and promising efficacy vs single-agent ibr in Richter’s transformation (RT), follicular lymphoma (FL), and diffuse large B-cell lymphoma (DLBCL). We examined potential biomarkers of treatment (tx) response using archived biopsy samples. GEP was used for DLBCL subtyping and to assess proportions of 22 distinct immune cells. Exome data were generated from 72 formalin-fixed paraffin-embedded samples, and sequencing analysis was used to identify mutations in genes of interest and assess somatic mutation burden. The correlations of immune cell proportions and gene variants were evaluated by investigator-assessed responses in each histology and by ongoing responses in DLBCL patients (pts; progression-free survival [PFS] > 24 months, n = 7 vs not, n = 20). In pts with available GEP and response data, overall response rates were 29.6% (8/27) for DLBCL, 43.5% (10/23) for FL and 81.3% (13/16) for RT. Proportions of CD8 and follicular helper T cells, M1 macrophages, and resting dendritic cells were higher in DLBCL responders vs nonresponders, while the proportion of regulatory T cells was decreased. These subsets did not differ by response in FL although an increase of CD4 memory resting T-cells was noted in responders. The trend toward increased follicular helper T cell and resting dendritic cell proportions in DLBCL pts was also associated with longer survival (PFS > 24 months) vs not. Gene variant data and responder status were available for 26 pts with DLBCL, 26 with FL, and 17 with RT. Comparison between responders and nonresponders showed that DLBCL pts with RNF213 (4/10 [40.0%] vs 1/16 [6.2%]) and KLHL14 (3/10 [30.0%] vs 0/16) mutations were more likely to respond to ibr + nivo. Conversely, nonresponders were associated with variants in EBF1, ADAMTS20, AKAP9, SOCS1, TP53, and genes in BCR pathways such as TNFRSF14, MYD88, and NFKB1B. BCL2 mutation in FL (9/12 [75.0%] vs 4/14 [28.6%]) and ROS1 mutation in RT (0/13 vs 2/4 [50.0%]) were associated with response; both are involved in the NF-kB pathway. In DLBCL, the most frequent gene mutations were RNF213, NBPF1, and BCL2 in pts who had PFS > 24 months (3/7 [42.9%] each), and KMT2D (8/20 [40.0%]) and CSMD3 (8/20 [40.0%]) in pts who did not. Somatic mutation burden was lower in responders vs nonresponders, especially in germinal center B-cell-DLBCL, and in DLBCL pts with PFS > 24 months vs not. In conclusion, we report gene variations among DLBCL, FL, and RT pts associated with response or durable PFS with ibr + nivo. While ibr inhibits Bruton’s tyrosine kinase-dependent pathways, we identify alternative gene pathway variants that may impact tx outcomes. Immune cell infiltration into the microenvironment relates to differential tx response with this immune combination and is histology dependent. Citation Format: Brendan Hodkinson, Michael Schaffer, Joshua Brody, Wojciech Jurczak, Cecilia Carpio, Dina Ben-Yehuda, Irit Avivi, Rao Saleem, Muhit Özcan, John Alvarez, Rob Ceulemans, Nele Fourneau, Sriram Balasubramanian, Anas Younes. Phase 1/2a LYM1002 study of ibrutinib (ibr) + nivolumab (nivo): Exome and gene expression profiling (GEP) analyses by histology and responder status [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4024.

Background Preclinical studies have shown synergistic antitumour effects between ibrutinib and immune-checkpoint blockade. The aim of this study was to assess the safety and activity of ibrutinib in combination with nivolumab in patients with relapsed or refractory B-cell malignant diseases. Methods We did a two-part, open-label, phase 1/2a study at 21 hospitals in Australia, Israel, Poland, Spain, Turkey, and the USA. The primary objective of part A (dose escalation) was to assess the safety of daily oral ibrutinib (420 mg or 560 mg) in combination with intravenous nivolumab (3 mg/kg every 2 weeks) to ascertain a recommended phase 2 dose in patients with relapsed or refractory high-risk chronic lymphocytic leukaemia or small lymphocytic lymphoma (del17p or del11q), follicular lymphoma, or diffuse large B-cell lymphoma. Dose optimisation was investigated using a modified toxicity probability interval design. The primary objective of the part B expansion phase was to establish the preliminary activity (the proportion of patients who achieved an overall response) of the combination of ibrutinib and nivolumab in four cohorts: relapsed or refractory high-risk chronic lymphocytic leukaemia or small lymphocytic lymphoma (del17p or del11q), follicular lymphoma, diffuse large B-cell lymphoma, and Richter's transformation. All participants who received at least one dose of treatment were included in the primary analysis and analyses were done by disease cohort. Findings Between March 12, 2015, and April 11, 2017, 144 patients were enrolled in the study. Three patients died before receiving study treatment; thus, 141 patients were included in the analysis, 14 in part A and 127 in part B. One dose-limiting toxicity (grade 3 hyperbilirubinaemia) was reported at the 420 mg dose in the diffuse large B-cell lymphoma cohort, which resolved after 5 days. The combination of ibrutinib and nivolumab led to overall responses in 22 (61%) of 36 patients with high-risk chronic lymphocytic leukaemia or small lymphocytic lymphoma, 13 (33%) of 40 patients with follicular lymphoma, 16 (36%) of 45 patients with diffuse large B-cell lymphoma, and 13 (65%) of 20 patients with Richter's transformation. The most common all-grade adverse events were diarrhoea (47 [33%] of 141 patients), neutropenia (44 [31%]), and fatigue (37 [26%]). 11 (8%) of 141 patients had adverse events leading to death; none were reported as drug-related. The most common grade 3-4 adverse events were neutropenia (40 [28%] of 141 patients) and anaemia (32 [23%]). The incidence of grade 3-4 neutropenia ranged from eight (18%) of 45 patients with diffuse large B-cell lymphoma to 19 (53%) of 36 patients with chronic lymphocytic leukaemia or small lymphocytic lymphoma; incidence of grade 3-4 anaemia ranged from five (13%) of 40 patients with follicular lymphoma to seven (35%) of 20 patients with Richter's transformation. The most common serious adverse events included anaemia (six [4%] of 141 patients) and pneumonia (five [4%]). The most common grade 3-4 immune-related adverse events were rash (11 [8%] of 141 patients) and increased alanine aminotransferase (three [2%]). Interpretation The combination of ibrutinib and nivolumab had an acceptable safety profile and preliminary activity was similar to that reported with single-agent ibrutinib in chronic lymphocytic leukaemia or small lymphocytic lymphoma, follicular lymphoma, and diffuse large B-cell lymphoma. The clinical response in patients with Richter's transformation was promising and supports further clinical assessment. Copyright (C) 2019 Elsevier Ltd. All rights reserved.

Background: The cell of origin (COO) in diffuse large B-cell lymphoma (DLBCL) has prognostic importance. While the COO was originally classified into germinal center B-cell (GCB) or activated B-cell (ABC) subtypes by microarray analysis, routine use was not practical. Immunohistochemistry (IHC)–based methods are widely used with varying results due to lack of standardization. Several classification methods have been developed recently; understanding the concordance between these and existing methods is essential to their practical application. Therefore, we evaluated concordance between 3 commercial assays: a standardized Hans-based IHC method and 2 gene expression profiling (GEP) methods and compared these to the accepted microarray classification.Methods: 137 DLBCL-confirmed tumor samples were evaluated using a standardized Hans-based IHC method for GCB or non-GCB subtype, by a published microarray-based assay, a digital gene expression-based Lymphoma Subtyping Test (LST) and a next-generation sequencing-based assay (EdgeSeq COO). Subtype calls from the 3 GEP methods were harmonized to “GCB” or “non-GCB” and assessed for concordance.Results: Concordance between the Hans-based IHC assay and the microarray-based assay, LST assay and EdgeSeq assay was 79.6% (N=137), 80.0% (N=125) and 78.2% (N=64), respectively; the positive percent agreement (PPA) in non-GCB was 88.9%, 87.0% and 78.1%, respectively. Concordance for the Hans-based IHC assay versus GEP methods was especially high for direct GCB calls (91.0%, 88.3% and 78.1% for microarray, LST and EdgeSeq COO methods, respectively). The newly developed GEP assays performed well against the microarray GEP method, against which they were calibrated (concordance 93.7% and 87.5% and PPA 94.3% and 92.9%, respectively, for LST and EdgeSeq COO).Conclusion: These results demonstrated good consistency between various platforms for stratification of DLBCL into COO subtype classifications. Application of a standardized Hans-based IHC assay offers a robust, rapid and easily accessible platform to classify DLBCL into prognostically important subtypes.

Abstract CD3 bispecific antibodies target both CD3 on T cells, and a tumor-specific antigen on cancer cells to harness the ability of cytotoxic T cells to eradicate solid tumors. Preclinical modeling of potential clinical leads in animal models can be challenging due to the need to have both a human immune system and a tumor model to study. Here we describe using various humanized mouse models in immune-compromised NSG mice as well as the use of surrogate antibodies in syngeneic models in mice with a competent immune system. To reconstitute the human immune system, female NSG mice were inoculated either intravenously with PBMCs or intraperitoneally with T cells that were expanded and activated in vitro. Engraftment of human T cells was evaluated in PBMC and T-cell humanized NSG mice in peripheral blood. In the absence of treatment, PBMC humanized mice had a greater reconstitution of T cells in the peripheral blood (~10-40%) compared to T-cell humanized mice (~3-10%). An antibody-specific expansion of T cells was observed in the T-cell humanized model in response to CD3 bispecific treatment. The slower engraftment of effector T cells in the T-cell humanized model correlated with slower onset to GVHD (graft versus host disease), allowing for extended evaluation of antitumor responses. Bispecific CD3 redirection antibodies elicited antitumor efficacy and T-cell infiltration in various human xenografts in both T-cell and PBMC-humanized mouse models. Additionally, mouse surrogate bispecific antibodies were generated that bind CD3 on mouse T cells. CT26 murine mouse colon carcinoma syngeneic tumors were transfected with a human cancer antigen for use in immune-competent Balb/c mice. Treatment with a surrogate bispecific molecule resulted in significant tumor growth inhibition (greater than 60%) and significant increase in life span. To confirm the mechanism of action, histologic analyses of T-cell infiltration into tumors was performed as well as T-cell cytokine analysis. These data showed that the level of T-cell infiltration and cytokine production correlated well with in vivo efficacy, demonstrating a T cell-mediated elimination of antigen-presenting tumor cells. In summary, both humanized mouse models and syngeneic mouse models using surrogate antibodies were successfully employed to study antitumor effects of CD3 bispecific antibodies. Citation Format: Bethany Mattson, Krista Menard, Damon Hamon, Darlene Pizutti, Emily Chen, Margarita Romero, Kristen Chevalier, Karla Wiehagen, Mark Richter, Gerald Chu, Brenda Hertzog, Anna Hughes, John Alvarez, Raluca Verona, Colleen Kane, Sheri Moores, Sylvie Laquerre, Joseph Erhardt, Kathryn Packman. Preclinical assessment of CD3 bispecific antibody efficacy: A comparison of humanized mouse models bearing xenografts and syngeneic mouse models using surrogate antibodies [abstract]. In: Proceedings of the AACR Special Conference: Advances in Modeling Cancer in Mice: Technology, Biology, and Beyond; 2017 Sep 24-27; Orlando, Florida. Philadelphia (PA): AACR; Cancer Res 2018;78(10 Suppl):Abstract nr A11.

BACKGROUND:Plasmablasts and plasma cells play a key role in many autoimmune diseases, such as rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). This study was undertaken to evaluate the potential of targeting CD38 as a plasma cell/plasmablast depletion mechanism by daratumumab in the treatment of patients with RA and SLE.METHODS:RNA-sequencing analysis of synovial biopsies from various stages of RA disease progression, flow cytometry analysis of peripheral blood mononuclear cells (PBMC) from patients with RA or SLE and healthy donors, immunohistochemistry assessment (IHC) of synovial biopsies from patients with early RA, and ex vivo immune cell depletion assays using daratumumab (an anti-CD38 monoclonal antibody) were used to assess CD38 as a therapeutic target.RESULTS:We demonstrated that the plasma cell/plasmablast-related genes CD38, XBP1, IRF4, PRDM1, IGJ and TNFSF13B are significantly up-regulated in synovial biopsies from patients with arthralgia, undifferentiated arthritis (UA), early RA and established RA as compared to healthy controls and control patients with osteoarthritis. In addition, the highest CD38 expression was observed on plasma cells and plasmablasts compared to natural killer (NK) cells, classical dendritic cells (DCs), plasmacytoid DCs (pDCs) and T cells, in blood from healthy controls and patients with SLE and RA. Furthermore, IHC showed CD38 staining in the same region as CD3 and CD138 staining in synovial tissue biopsies from patients with early RA. Most importantly, our data show for the first time that daratumumab effectively depletes plasma cells/plasmablasts in PBMC from patients with SLE and RA in a dose-dependent manner ex vivo.CONCLUSION:These results indicate that CD38 may be a potential target for RA disease interception and daratumumab should be evaluated clinically for the treatment of both RA and SLE.

Abstract Background: Though the techniques to interrogate the appearance of a biomarker in tissue sections have greatly advanced, there are limitations as to how representative an analysis of a tissue section is compared to the entire diseased tissue. Depending on the heterogeneous expression level of a biomarker, tissue sampling can result in different interpretations of the biomarker’s appearance, and hence could potentially lead to a false therapeutic intervention. Hypothesis: Digital image analysis has demonstrated tremendous value in quantifying many features related to biomarker distribution and expression in biological tissues. The information can be collected for various indications and biomarkers and a phenotypic signature can be established that describes a biomarker representation across indications. Moreover, the assessment of new samples can be compared to the established phenotypic signature and a confidence score applied in support to the determined endpoint. Approach: For a proof of concept, 6 prostate cancer samples were processed and a single section was collected after every 100microns. A total of 7 sections per sample were stained for the lymphocyte marker CD3, and the number of positive target cells were determined in the tumor and tumor microenvironment using tissue Image Analysis (tIA™). To assess how indicative the evaluation of a single tissue section would be for the entire tumor, the heterogeneity level was determined on the section level as well as by random grid analysis on each individual section. Both criteria were utilized to define an indication and biomarker specific confidence interval and heterogeneity score. Conclusion: The combination of IHC and tIA is a powerful tool to convert complex data into meaningful interpretations. tIA is also a capable tool to catalogue valuable information about the biomarker’s expression pattern across different disease stages and hence could be used to evaluate how representative a single biomarker evaluation is in the grand scheme. Ultimately, we demonstrated a technique that can be applied to any biomarker and would assist in guiding therapeutic decisions. Citation Format: Carsten Schnatwinkel, Daniel Rudmann, Famke Aeffner, Jasmeet Bajwa, Natalie Hutnick, Michael Sharp, Gerry Chu, JD Alvarez. Providing confidence around computational tissue analysis using heterogeneity assessments [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 1710. doi:10.1158/1538-7445.AM2017-1710

7561 Background: CD38 is a type II transmembrane glycoprotein expressed on normal lymphoid and myeloid cells and can be highly expressed in hematologic malignancies. An IHC prototype assay was developed to detect CD38 expression in formalin-fixed, paraffin-embedded (FFPE) tissue specimens from three relapsed or refractory non-Hodgkin’s lymphoma (NHL) subtypes: diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL) and mantle cell lymphoma (MCL) for selection of patients for treatment with daratumumab in a Phase 2 clinical trial (NCT02413489). Methods: This assay is based on EnVision FLEX IHC technology using CD38, clone DAK-CD38, primary antibody that has been developed and manufactured by Agilent Technologies. The assay staining protocol was developed for Dako PT Link and Autostainer Link 48. Specificity of DAK-CD38 staining was demonstrated by Western blot analysis of cancer cell lysates, as well as IHC on both normal and cancer tissue. Assay precision and robustness were evaluated using commercially procured FFPE DLBCL, FL and MCL specimens. Results: CD38 IHC DAK-CD38 detected a broad range of CD38 expression in DLBCL, FL, and MCL specimens. FFPE specimens derived from cancer cell lines exhibited a range of IHC staining intensity and confirmed the reported expression of CD38 in the scientific literature. All precision and robustness results met acceptance criteria for both IHC intensity and percent positive cells staining. Conclusions: Our studies demonstrate that the Dako CD38 IHC DAK-CD38 assay is sensitive, specific, precise, and robust for the detection of CD38 expression in DLBCL, FL and MCL. This assay may also have the potential to be used for the detection of CD38 in additional tumor types.

Background: Preclinical studies have demonstrated synergistic antitumor activity when immune checkpoint inhibitors targeting the programmed cell death protein-1 (PD-1)/PD ligand-1 (PD-L1) axis were combined with the Bruton's tyrosine kinase inhibitor ibrutinib. To test this hypothesis, we conducted a phase 1/2a study evaluating the safety and efficacy of ibrutinib in combination with nivolumab in patients (pts) with relapsed or refractory (R/R) high-risk chronic lymphocytic leukemia (CLL), small lymphocytic lymphoma (SLL), follicular lymphoma (FL), diffuse-large B-cell lymphoma (DLBCL), and Richter's transformation (RT).

There is growing recognition that immunotherapy is likely to significantly improve health outcomes for cancer patients in the coming years. Currently, while a subset of patients experience substantial clinical benefit in response to different immunotherapeutic approaches, the majority of patients do not but are still exposed to the significant drug toxicities. Therefore, a growing need for the development and clinical use of predictive biomarkers exists in the field of cancer immunotherapy. Predictive cancer biomarkers can be used to identify the patients who are or who are not likely to derive benefit from specific therapeutic approaches. In order to be applicable in a clinical setting, predictive biomarkers must be carefully shepherded through a step-wise, highly regulated developmental process. Volume I of this two-volume document focused on the pre-analytical and analytical phases of the biomarker development process, by providing background, examples and “good practice” recommendations. In the current Volume II, the focus is on the clinical validation, validation of clinical utility and regulatory considerations for biomarker development. Together, this two volume series is meant to provide guidance on the entire biomarker development process, with a particular focus on the unique aspects of developing immune-based biomarkers. Specifically, knowledge about the challenges to clinical validation of predictive biomarkers, which has been gained from numerous successes and failures in other contexts, will be reviewed together with statistical methodological issues related to bias and overfitting. The different trial designs used for the clinical validation of biomarkers will also be discussed, as the selection of clinical metrics and endpoints becomes critical to establish the clinical utility of the biomarker during the clinical validation phase of the biomarker development. Finally, the regulatory aspects of submission of biomarker assays to the U.S. Food and Drug Administration as well as regulatory considerations in the European Union will be covered.

Immunotherapies have emerged as one of the most promising approaches to treat patients with cancer. Recently, there have been many clinical successes using checkpoint receptor blockade, including T cell inhibitory receptors such as cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) and programmed cell death-1 (PD-1). Despite demonstrated successes in a variety of malignancies, responses only typically occur in a minority of patients in any given histology. Additionally, treatment is associated with inflammatory toxicity and high cost. Therefore, determining which patients would derive clinical benefit from immunotherapy is a compelling clinical question. Although numerous candidate biomarkers have been described, there are currently three FDA-approved assays based on PD-1 ligand expression (PD-L1) that have been clinically validated to identify patients who are more likely to benefit from a single-agent anti-PD-1/PD-L1 therapy. Because of the complexity of the immune response and tumor biology, it is unlikely that a single biomarker will be sufficient to predict clinical outcomes in response to immune-targeted therapy. Rather, the integration of multiple tumor and immune response parameters, such as protein expression, genomics, and transcriptomics, may be necessary for accurate prediction of clinical benefit. Before a candidate biomarker and/or new technology can be used in a clinical setting, several steps are necessary to demonstrate its clinical validity. Although regulatory guidelines provide general roadmaps for the validation process, their applicability to biomarkers in the cancer immunotherapy field is somewhat limited. Thus, Working Group 1 (WG1) of the Society for Immunotherapy of Cancer (SITC) Immune Biomarkers Task Force convened to address this need. In this two volume series, we discuss pre-analytical and analytical (Volume I) as well as clinical and regulatory (Volume II) aspects of the validation process as applied to predictive biomarkers for cancer immunotherapy. To illustrate the requirements for validation, we discuss examples of biomarker assays that have shown preliminary evidence of an association with clinical benefit from immunotherapeutic interventions. The scope includes only those assays and technologies that have established a certain level of validation for clinical use (fit-for-purpose). Recommendations to meet challenges and strategies to guide the choice of analytical and clinical validation design for specific assays are also provided.

Abstract 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 H&E slide, and/or adjacent slides. Using the aligned FACTS data and our proprietary MultivariateMap approach, we integrated the patterns of each marker based on immune cell type function and their location relative to each other and the tumor epithelial cells. In this study, we demonstrated how spatial integration of immune cell markers in the context of whole tissues can be applied to the diagnostic setting. By creating a comprehensive landscape of the immune system state in the tissue biopsies, we were able to identify crucial patterns which represent function and role in immune system biology. These approaches provide a robust platform for immuno-oncology applications by providing information on the state of the immune system in cancer using approaches implementable in the clinic. The use of these approaches will benefit further understanding of cancer pathology, and can directly lead to the development of diagnostic tests with clinical utility. Citation Format: Joseph S. Krueger, Nathan Martin, Famke Aeffner, Anthony Milici, John Alvarez, Micheal Sharp. Quantitative analysis of multiple subtypes of immune system cells in cancer tissues. [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 C109.

Meeting abstracts The state of the immune system is reflected, in part, by the cell populations present in individual tissues which reflect the tumor microenvironment (TME). Several studies have suggested that understanding the TME constituents is useful for predicting drug response outcomes.

Abstract Mesothelin is a 40 kDa secreted glycoprotein expressed in normal mesothelial cells and over-expressed in several histological types of tumors. Detection of mesothelin by immunohistochemistry (IHC) may assist in the diagnosis of mesothelioma. Mesotheliomas are positive for mesothelin staining, but carcinomas of the lung may also be positive for this marker. Mesothelin positivity in adenocarcinoma of the lung reportedly ranges from 22% to 71% of cases depending upon the specific subtype. Mesothelin positivity in squamous cell carcinomas (SCC) of lung is reported to be 16% in non-keratinizing and 31% in keratinizing subtypes. We performed mesothelin IHC on a cohort of 50 lung carcinoma samples (16 adenocarcinomas and 34 SCC). Mesothelin expression was observed in 10 out of 16 (62.5%) lung adenocarcinoma samples, of which 8 showed greater than 10% of tumor cells with positive membrane staining of any intensity. Mesothelin expression was observed in 19 out of 34 (55.9%) lung SCC samples, of which only 3 samples showed greater than 10% of tumor cells with positive membrane staining of any intensity. These findings are in concordance with previous reports which show a higher prevalence of mesothelin protein expression in lung adenocarcinoma than in lung SCC. To correlate RNA expression with protein expression, we performed gene expression profiling on a subcohort of these lung cancer specimens. Out of 50 lung carcinoma samples, 28 samples (10 adenocarcinomas and 18 SCCs) provided adequate RNA yield for gene expression profiling for mesothelin. A total mean relative gene expression (mRGE) value of 13.64 with a standard deviation (SD) of ±3.71 was obtained for the adenocarcinoma samples and a mRGE value of 14.78 with a SD of 2.64 was obtained for the SCC samples. We next arbitrarily assigned samples with greater than 10% mesothelin stained cells as “positive” and the remainder as “negative”. Six negative adenocarcinoma samples yielded a mRGE value of 12.98 with a SD of ±3.84, and four positive adenocarcinoma samples yielded a mRGE value of 14.63 with SD of ±3.82. Sixteen negative SCC samples resulted in a mRGE 14.63 with SD of ± 2.75 and two positive SCC samples yielded a mRGE of 15.98 with a SD of ±1.37. There was no apparent correlation between mRGE values and IHC positivity. We also correlated gene expression with p53 mutation status. Sixteen wild type samples, composed of equal number of adenocarcinoma and SSC had a total mRGE of 14.29 with SD of ±3.32. Seven samples with p53 mutations had a total mRGE of 16.09 with SD ±1.93. These data indicate that mesothelin gene expression is not associated with p53 mutation status. Future studies with an increased number of samples may yield significant associations between protein expression, gene expression and mutation status. Citation Format: Jackson Wong, Dana Gaffney, Michael Sharp, Brenda Hertzog, Jayaprakash Karkera, Suso Platero, John Alvarez. Profiling mesothelin protein expression by immunohistochemistry and gene expression in adenocarcinoma and squamous cell carcinoma of lung. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 4666. doi:10.1158/1538-7445.AM2014-4666

Abstract Immunocompromised mice engrafted with human hematopoietic stem cells (HSC) or peripheral blood mononuclear cells (PBMC) can be promising models of human immune system compartments, useful for studying diseases such as diabetes, lupus, GVHD, and cancer. We have utilized the PBMC-humanized mouse as a platform for xenograft models. To investigate the effect of human antibody (Ab) treatment on immune cells and xenografts using this platform, we examined the infiltrate of human immune cells in A431 xenografts with treatment of an effector function-enhanced anti-tissue factor VIIA (TFVIIA) Ab. Female NSG mice were intravenously engrafted with human PBMC, and subcutaneously implanted two weeks later with A431 human SCC cells. One week post tumor cell implant and thereafter twice weekly, mice received treatment with anti-TFVIIA Ab. Peripheral blood, bone marrow and spleen were analyzed by FACS and terminal tumor samples were analyzed by FACS and immunohistochemistry (IHC). FACS analysis of blood from Ab treated and untreated tumor-bearing mice collected from baseline to termination, showed an overall increase in the frequency of human CD45+, CD3+, CD8+, and CD4+ cells, while the frequency of CD56+ cells remained stable. In comparison, non-tumor bearing mice had increased CD56+ cells and decreased CD3+ cells. FACS analysis of tumors showed a robust increase in the frequency of CD56+ cells and a decrease in the frequency of CD3+ cells with treatment of anti-TF Ab. With treatment, the frequency of CD8+ cells was markedly lower and the frequency of CD4+ cells unchanged. Histological analysis was performed on A431 FFPE xenografts from both the treated and non-treated groups. Hematoxylin and eosin (H&E) staining of the xenografts confirmed the presence of inflammatory cells in the tumors and also showed greater PBMC infiltration of the tumors in the presence of the anti-TFVIIA Ab. Xenografts treated with anti-TFVIIA Ab exhibited greater infiltration of human CD3+, CD8+, and CD57+ cells. There was an increase in cleaved caspase-3 (CC3)+ cells in Ab-treated xenografts, indicating an increase in apoptosis with anti-TFVIIA Ab treatment. In conclusion, anti-TFVIIA Ab treatment causes changes in immune cell populations in xenografts. While the mechanism is unknown, one possibility is that NK migration into the tumor coupled with increased apoptosis with treatment indicates ADCC activity by the NK cells. This would signify some preservation of immune cell function in the PBMC engraftment model, and would enhance studying the effects of experimental antibodies in a more humanized setting. Citation Format: Hillary Millar-Quinn, Brenda Hertzog, Rebecca Hanson, Jeffrey Nemeth, John Alvarez. Human immune cell infiltration of tumors in a PBMC-humanized NSG mouse xenograft model changes with treatment of an anti-tissue factor antibody. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1655. doi:10.1158/1538-7445.AM2014-1655

Abstract Interleukin-6 (IL-6) is an important growth factor for estrogen receptor-alpha (ERα) positive breast cancer. High IL-6 serum levels are associated with poor prognosis in ERα+ breast cancer patients, and IL-6 related polymorphisms that lead to elevated IL-6 expression are associated with decreased overall survival (PMID 17261184, 12771987). Preclinical data suggest that in contrast to ERα negative breast cancer cell lines, ERα positive cell lines rarely produce autocrine IL-6, and are therefore dependent on paracrine IL6 produced by the microenvironment. Furthermore, when ERα positive cell lines are exposed to paracrine IL-6, phosphorylation of STAT3-Y705, increased growth rates and more aggressive tumor phenotypes are observed (PMID 17586727, 18974155, 19581928). We utilized 3D co-cultures and heterotypic xenograft models to investigate the ability of siltuximab, an anti-IL-6 antibody in clinical development, to attenuate paracrine IL-6 effects across a panel of 8 ERα positive breast cancer cell lines and heterotypic xenograft models. From a panel of 8 ERα postive breast cancer lines, 75% responded to recombinant human IL-6 (hIL-6) protein by phosphorylation of STAT3 (Ty705), but not AKT, MEK1/2, or ERK1/2. Siltuximab treatment blunted pSTAT3 induction in all IL-6 responsive ERα positive breast cancer lines but failed to reduce pSTAT3 phosphorylation in three ERα negative cell lines that produced autocrine IL-6. The 3D tumor growth assay demonstrated accelerated growth rates for ERα positive breast cancer lines in the presence of IL-6 or human mesenchymal stem cells (hMSC), which returned to baseline with siltuximab treatment. In addition, when hIL-6 was supplemented in vivo, the ERα positive tumor cell line, MCF-7 engrafted without the need for estrogen supplementation and tumor growthblunted with siltuximab treatment. When tumors were allowed to establish before treatment, siltuximab was able to induce tumor regressions in all treated animals (10/10). Tumor regression was associated with decreased mitotic counts in tumors. In addition to these experiments, we investigated STAT3 activation in a panel of primary patient-derived ERα positive breast cancer samples. 43% of samples demonstrated activated STAT3 as determined by Y705 phosphorylation. These same ERα positive breast cancer patients displayed an activated IL-6 network as assessed by an a priori 32-gene IL-6 network gene expression profile signature and was also associated with increased IL-6 serum levels. This activated IL-6 gene signature was also associated with ERα positive tumors and increased metastasis as assessed by positive node status in four independent breast cancer data sets. Taken together, these data suggest a key role for IL-6 in estrogen independent ERα positive breast cancer progression. Citation Format: Amy Axel, Tineke Casneuf, Peter King, JD Alvarez, Brett Hall, Kate Sasser. The role of IL-6 in ERα+ breast cancer and potential use for Siltuximab, an anti-IL-6 antibody, in ERα+ breast cancer treatment. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3530. doi:10.1158/1538-7445.AM2013-3530

Abstract Background: The molecular apocrine (MA) subtype of breast cancer is identified by gene expression profiling. MA tumors are estrogen receptor (ER) negative and progesterone receptor (PR) negative, but still express estrogen responsive genes. The androgen receptor (AR) pathway may be driving growth in these tumors because androgen responsive genes are expressed in tumors with the MA gene signature. The MA gene signature is identified in approximately 10% of triple negative breast cancer (TNBC) and may predict patients with tumors responsive to agents that inhibit the AR pathway. AR protein expression, measured by immunohistochemistry (IHC), may be a surrogate for the MA gene signature, but to date, a careful comparison of gene expression profiles and AR protein expression has not been conducted. In this study, cohorts of TNBCs were assessed for the MA gene signature and these results were compared with AR IHC expression and with a novel gene expression assay that may predict tumors with the MA gene signature. Methods: Formalin fixed, paraffin-embedded (FFPE) TNBC samples were commercially obtained. ER, PR and HER2 status of these samples was confirmed by IHC. AR expression was detected by IHC using two different antibody clones. Both staining intensity and percent positive cells were recorded for each sample. Gene expression data was collected from a cohort of TNBC FFPE samples using cDNA-mediated Annealing, Selection, extension, and Ligation (DASL) technology. A 2-gene classifier of the MA gene expression signature was derived by interrogating publically available gene expression data from ER-negative breast cancers. A reverse-transcriptase polymerase chain reaction (RT-PCR) assay to detect the 2-gene classifier was developed. Cell lines predicted to have the MA gene signature by the 2-gene assay were tested for sensitivity to R-1881 in vitro. Results: Using computational approaches and publically available datasets, we confirmed the validity of the MA gene signature and estimated the prevalence to be between 12% and 37% in ER-negative breast tumors. The 2-gene classifier was 100% specific in determining MA tumors in a training set using gene expression data as a standard. In a validation set, the 2-gene assay was 66% correlative with AR IHC positivity when the IHC cut-off was set at 10% positive tumor cells. Cell lines predicted to express the MA gene signature by the 2-gene classifier proliferated in response to androgen. This effect was blocked by Flutamide. Conclusions: These results indicate that AR IHC using a 10% cut-off may not completely correlate with the MA gene signature. Further refinement of AR IHC scoring criteria may produce greater specificity. Cell proliferation data suggests the 2-gene assay can predict tumors that will proliferate in response to androgen. Work is ongoing to determine the correlation between the 2-gene assay results, AR IHC and DASL gene expression data to fully understand the predictability of this assay. Understanding this correlation may allow use of simple clinical assays to accurately select patients responsive to agents that block AR signaling. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P5-01-09.

Patients with hormone refractory prostate cancer (HRPC) have very few clinical options. A bottle neck in development of new therapies is the poor understanding of the molecular basis of hormone independency. A pathway thought to be important in this progression is FGFR. Amplification of FGFR1 has been observed in breast cancer and has led to the hypothesis that it is involved in Tamoxifen resistance (Turner et al., 2010). Here we evaluated the copy number variations (CNV) of the gene loci of FGFR1 (8p12) and FGFR2 (10q26) by dual color FISH in FFPE tissue of recurrent PC patients (n=28), non-recurrent PC patients (n=15), PC patients undergoing palliative RPE (n=15), patients with benign prostate hyperplasia (n=17) and tumor free tissue samples (n=12). We analyzed CNV between different clinical patient subgroups. Two types of CNV for the gene loci of FGFR1 and FGFR2 were observed. Either the gene dose of the target gene locus occurred to be amplified or the gene dose was elevated due to polysomy of the respective chromosome. Furthermore patients with CNV of FGFR1 or FGFR2 exhibited in 40-50% intra tumor heterogeneity by presenting normal cells and cells with CNV. FGFR1 gene dose elevations (either due to gene locus amplification or polysomy of chromosome 8) were observed in 7% (1/15) of patients without recurrence but in 49% (17/35) of the patients with recurrence. Also, in patients undergoing palliative RPE 75% (12/16) of the patients exhibited FGFR1 gene dose elevation. In the case of FGFR2, gene dose elevations (gene locus amplification or polysomy of chromosome 10) were observed in none of the patients without recurrence but in 13% of the patients with recurrence and in 20% of patients undergoing palliative RPE. In the control groups such variations were not found. Interestingly, the gene dose variations differed in the prostate cancer patients depending on their hormone sensitivity. 47% (7/15) of hormone refractory patients and 30% (6/20) of the hormone sensitive patients showed elevated FGFR1 gene dose, for FGFR2, 30% and 20% of the respective subgroups were affected. In summary we conclude that this is first evidence that CNV of FGFR1 and FGFR2 might have implications for hormone resistance in PC. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 3211. doi:10.1158/1538-7445.AM2011-3211

Fibroblast growth factor receptors (FGFRs) are widely expressed in multiple organ systems and are involved in growth and angiogenesis. Mutations, translocations and amplifications in FGFR genes have been observed in several types of cancer. The purpose of this study was to identify biomarkers that may predict sensitivity and may change upon FGFR inhibition. 240 tumor cell lines from different types of cancer were treated with a small molecule FGFR inhibitor and IC50 values were calculated. Mutations, amplifications, deletions, gene expression analysis, and network analysis were used to find biomarkers that were sensitive or resistant to the inhibitor. Xenografts were used to identify pharmacodynamics biomarkers. Tumors from cancer patients were analyzed for FGFR gene amplifications using Fluorescent in situ hybridization. FGFR1 mRNA overexpression and FGFR2 amplification were identified as sensitive biomarkers while Kras mutation was associated with resistance. Multiple sensitive cell lines (n = 2–4) were found in cells derived from breast, lung, gastric, kidney, lymphoma, sarcoma, melanoma, and endometrial cancer. No sensitive cell lines were found in colorectal, pancreatic, leukemia, myeloma, or ovarian cancer. In human breast cancer cell lines, amplifications were found for FGFR1 (23%), FGFR2 (11%), and FGFR4 (16%). Of the 26 breast cancer cell lines analyzed for FGFR1 amplification, 5 had moderate amplification (4–10 copies) and 1 was highly amplified (>10 copies). In non-small cell lung cancer cell lines, amplifications were found for FGFR1 (37%) and FGFR2 (20%). In xenografts, pS6 and pMAPK levels changed upon compound treatment. FGFR1 and FGFR2 amplifications were also found in prostate (15% and 15%, respectively) tumors from patients. This study identified biomarkers for an FGFR small molecule inhibitor. These results may provide a rationale for patient selection and patient dosing when using a FGFR inhibitor. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2011 Nov 12-16; San Francisco, CA. Philadelphia (PA): AACR; Mol Cancer Ther 2011;10(11 Suppl):Abstract nr C12.