Additional Tumor Types Research Articles

Trial in progress: A phase I/II, open-label, dose-escalation, safety and tolerability study of NC318 in subjects with advanced or metastatic solid tumors.

TPS3166 Background: Siglec-15 (S15) is a member of the Siglec family (Sialic acid-binding Immunoglobulin Lectins), a distinct subgroup of immunoglobulin (Ig) superfamily proteins involved in discriminating self from non-self-immune regulation (Macauley MS et al. 2014). Nonclinical models demonstrate S15 mediates suppression of T cell proliferation and negatively regulates T cell function. NC318 is a first-in-class monoclonal antibody that blocks S15-mediated immune suppression and prevents tumor growth by normalizing T cell function and restoring anti-tumor immunity in the tumor microenvironment (Wang J et al. 2019). The clinical hypothesis of this study is that NC318 targeting of S15 can improve anti-tumor immune response and provide benefit in multiple oncology indications. Methods: This is a multi-center, first in human, phase 1/2, open-label, non-randomized study to determine the safety and tolerability, define maximum tolerated dose and/or pharmacologically active dose, assess preliminary efficacy, and explore predictive and pharmacodynamic biomarkers of NC318 in subjects with advanced or metastatic solid tumors. Key eligibility criteria included measurable disease based on RECIST v1.1 and consent for collection of biopsies at screening and on treatment (optional for phase 1). Phase 1 used a 3+3 dose escalation design to determine the recommended phase 2 dose (RP2D) and schedule of NC318 while identifying drug related toxicities (DLTs). Phase 2 enrollment is limited to non-small cell lung, ovarian, head and neck, and triple negative breast cancer subjects with PD-L1 tumor proportion scores <50% (additional tumor types are being evaluated for inclusion). Ongoing exploratory analyses include the assessment of predictive biomarkers associated with treatment benefit, and pharmacodynamic markers associated with study drug activity (e.g. evaluation of tumor biopsies and peripheral markers of inflammation). Phase 1 enrollment began October 2018 and completed in August 2019. The RP2D was defined and the phase 2 opened to enrollment October 2019. Clinical trial information: NCT03665285 .

Preclinical antitumor efficacy of senescence-inducing chemotherapy combined with a nanoSenolytic

The induction of senescence produces a stable cell cycle arrest in cancer cells, thereby inhibiting tumor growth; however, the incomplete immune cell-mediated clearance of senescent cells may favor tumor relapse, limiting the long-term anti-tumorigenic effect of such drugs. A combination of senescence induction and the elimination of senescent cells may, therefore, represent an efficient means to inhibit tumor relapse. In this study, we explored the antitumor efficacy of a combinatory senogenic and targeted senolytic therapy in an immunocompetent orthotopic mouse model of the aggressive triple negative breast cancer subtype. Following palbociclib-induced senogenesis and senolysis by treatment with nano-encapsulated senolytic agent navitoclax, we observed inhibited tumor growth, reduced metastases, and a reduction in the systemic toxicity of navitoclax. We believe that this combination treatment approach may have relevance to other senescence-inducing chemotherapeutic drugs and additional tumor types. SignificanceWhile the application of senescence inducers represents a successful treatment strategy in breast cancer patients, some patients still relapse, perhaps due to the subsequent accumulation of senescent cells in the body that can promote tumor recurrence. We now demonstrate that a combination treatment of a senescence inducer and a senolytic nanoparticle selectively eliminates senescent cells, delays tumor growth, and reduces metastases in a mouse model of aggressive breast cancer. Collectively, our results support targeted senolysis as a new therapeutic opportunity to improve outcomes in breast cancer patients.

Evaluation of glucocorticoid-induced TNF receptor (GITR) expression in breast cancer and across multiple tumor types.

Glucocorticoid-induced TNF receptor (GITR) is an emerging immunotherapy target that is expressed at high levels on regulatory T cells. Agonistic anti-GITR antibodies have anti-tumor activity in cancer mouse models, and recent phase 1 trials have demonstrated their safe pharmacological profile. However, there is limited knowledge on the relationship between GITR expression and the tumor microenvironment. GITR protein expression was assayed by immunohistochemistry on 3992 breast cancer surgical excision specimens assembled into tissue microarrays and scored visually by a pathologist for GITR expression on tumor-infiltrating lymphocytes and on carcinoma cells. GITR expression by the malignant cells was further surveyed in gastrointestinal stromal tumor (N = 713), lung carcinoma (N = 705), pancreatic cancer (N = 486), ovarian cancer (N = 445), bladder cancer (N = 88), prostate cancer (N = 88), testicular cancer (N = 76), melanoma (N = 75), renal cell carcinoma (N = 68), epithelioid sarcoma (N = 53), and neuroendocrine tumors (N = 41). In breast cancer, GITR expression on tumor-infiltrating lymphocytes (12.4%) correlated with other immune response biomarkers (PD-L1+ on tumor cells, and PD-1+, LAG-3+, TIM-3+ lymphocytes; p < 0.001), and T-cell markers (CD8+, FOXP3+; p < 0.001). GITR+ carcinoma cells were observed in 6.0% of breast cancer cases and correlated with worse relapse-free survival (p = 0.015). Among the additional tumor types examined, cancers with GITR+ malignant cells included bladder cancer (5.7%), primary (but not metastatic) melanoma (4.5%), and ovarian cancer (3.2%); no expression was identified among examined sarcomas. To our knowledge, this is the first immunohistochemistry study to report the frequency and pattern of GITR expression in a large breast cancer cohort, or to report membranous GITR expression on malignant cells. The co-infiltration of GITR with other immune biomarkers and T-cell markers supports a potential role for anti-GITR agents in combination immunotherapies. In addition, GITR expression on carcinoma cells could imply the existence of a novel cancer immune evasion strategy worthy of further investigation.

An Integrated Meta-Analysis of Secretome and Proteome Identify Potential Biomarkers of Pancreatic Ductal Adenocarcinoma.

Pancreatic ductal adenocarcinoma (PDAC) is extremely aggressive, has an unfavorable prognosis, and there are no biomarkers for early detection of the disease or identification of individuals at high risk for morbidity or mortality. The cellular and molecular complexity of PDAC leads to inconsistences in clinical validations of many proteins that have been evaluated as prognostic biomarkers of the disease. The tumor secretome, a potential source of biomarkers in PDAC, plays a crucial role in cell proliferation and metastasis, as well as in resistance to treatments, which together contribute to a worse clinical outcome. The massive amount of proteomic data from pancreatic cancer that has been generated from previous studies can be integrated and explored to uncover secreted proteins relevant to the diagnosis and prognosis of the disease. The present study aimed to perform an integrated meta-analysis of PDAC proteome and secretome public data to identify potential biomarkers of the disease. Our meta-analysis combined mass spectrometry data obtained from two systematic reviews of the pancreatic cancer literature, which independently selected 20 studies of the secretome and 35 of the proteome. Next, we predicted the secreted proteins using seven in silico tools or databases, which identified 39 secreted proteins shared between the secretome and proteome data. Notably, the expression of 31 genes of these secretome-related proteins was upregulated in PDAC samples from The Cancer Genome Atlas (TCGA) when compared to control samples from TCGA and The Genotype-Tissue Expression (GTEx). The prognostic value of these 39 secreted proteins in predicting survival outcome was confirmed using gene expression data from four PDAC datasets (validation set). The gene expression of these secreted proteins was able to distinguish high- and low-survival patients in nine additional tumor types from TCGA, demonstrating that deregulation of these secreted proteins may also contribute to the prognosis in multiple cancers types. Finally, we compared the prognostic value of the identified secreted proteins in PDAC biomarkers studies from the literature. This analysis revealed that our gene signature performed equally well or better than the signatures from these previous studies. In conclusion, our integrated meta-analysis of PDAC proteome and secretome identified 39 secreted proteins as potential biomarkers, and the tumor gene expression profile of these proteins in patients with PDAC is associated with worse overall survival.

Pan-Tumor Pathologic Scoring of Response to PD-(L)1 Blockade.

Pathologic response assessment of tumor specimens from patients receiving systemic treatment provides an early indication of therapeutic efficacy and predicts long-term survival. Grading systems for pathologic response were first developed for chemotherapy in select tumor types. Immunotherapeutic agents have a mechanism of action distinct from chemotherapy and are being used across a broad array of tumor types. A standardized, universal scoring system for pathologic response that encompasses features characteristic for immunotherapy and spans tumor types is needed. Hematoxylin and eosin-stained slides from neoadjuvant surgical resections and on-treatment biopsies were assessed for features of immune-related pathologic response (irPR). A total of 258 specimens from patients with 11 tumor types as part of ongoing clinical trials for anti-PD-(L)1 were evaluated. An additional 98 specimens from patients receiving anti-PD-(L)1 in combination with other treatments were also reviewed, including those from three additional tumor types. Common irPR features (immune activation, cell death, tissue repair, and regression bed) were present in all tumor types reviewed, including melanoma, non-small cell lung, head and neck squamous cell, Merkel cell, and renal cell carcinoma, among others. Features were consistent across primary tumors, lymph nodes, and distant metastases. Specimens from patients treated with anti-PD-(L)1 in combination with another agent also exhibited irPR features. irPR features are consistent across tumor types and treatment settings. Standardized, pan-tumor irPR criteria (irPRC) are defined and associated specimen-handling considerations are described. Future, prospective studies are merited to validate irPRC in larger datasets and to associate pathologic features with long-term patient outcomes.

Abstract A008: PDX-based screen to evaluate and compare approved CDK4/6 inhibitors in breast cancer and determine palbociclib efficacy in additional tumor types

Abstract Background: FDA-approved CDK 4/6 inhibitors (CDK4/6i) including palbociclib, abemaciclib, and ribociclib have demonstrated clinical benefit in treating hormone receptor-positive (HR+) breast cancer. However, whether there is differential sensitivity to each therapy in vivo is unknown. In addition, whether these inhibitors demonstrate activity towards other tumor types is unclear. To better understand potential differences in CDK4/6i activity, we compared activity of each in a panel of 100 HR+/- breast cancer PDX models. To identify additional tumor types sensitive to CDK4/6i, we screened single agent palbociclib in an additional panel of 350 PDX models representing solid and hematologic malignancies. FFPE samples were collected from control and treatment groups in all studies at endpoint, and %T/C values in each model were calculated and in breast studies compared. Methods: START PDX models were established and validated as previously described and further characterized using IHC, WES, RNAseq, and drug sensitivity studies. For each CDK4/6i study, models were implanted SC in immune-deficient mice and studies initiated at 200-300 mm3 (n=1-3/grp). CDK4/6i agents were administered orally once daily at 50-75 mg/kg for a minimum of twenty-eight days or until study completion. Models tested with palbociclib included breast, lung, ovary, pancreas, head/neck, renal, gastric, uterine, colorectal, melanoma, and various hematologic malignancies. Study endpoints included tumor volume and time from treatment initiation with %T/C values and tumor regression reported at study completion; a %T/C of ≤ 20% versus control was considered sensitive. Partial tumor regression (PR) (%PR=1-99%) and complete tumor regression (CR) (%CR=100%) values versus Day 0 tumor volume of treated groups were also reported. Results: HR+/- breast models demonstrated variable sensitivity to the three CDK4/6i with HR+ models established from chemo-naïve patients reporting greatest tumor growth inhibition (TGI). Two of four HR+ ESR1 mutant models reported modest sensitivity to CDK4/6i (ST941Y537S and ST2535D538G), while ST2177Y537S was sensitive and ST2056Y537S refractory to the three therapies. Evaluation of palbociclib in additional indications reported notable TGI versus control in several indications with the following percentage of tested models reporting sensitivity: Lung: 10%, Ovary: 15%, Pancreas: 30%, Head/Neck: 35%, Renal: 45%, Gastric: 60%, Uterine: 20%, Colorectal: 45%, Melanoma: 40% and Hematologic: 15%. In addition, several indications reported tumor regressions including one CR (lung: ST1748) and partial responses in ovary, pancreas, head/neck, renal, colorectal, and melanoma. Conclusion: We evaluated and compared three FDA-approved CDK4/6i therapies in a panel of HR+/- breast PDX models and identified models with differential responses. In addition, we screened single agent palbociclib in a panel of 350 PDX models representing solid and hematologic malignancies and identified several sensitive models in multiple indications, including some with partial or complete tumor regressions. Citation Format: Lizette Gamez, Tomoyuki Mashimo, Alyssa Moriarty, Kyriakos Papadopoulos, Drew Rasco, Amita Patnaik, Amy Lang, Ronald Drengler, Lon Smith, Michael J Wick. PDX-based screen to evaluate and compare approved CDK4/6 inhibitors in breast cancer and determine palbociclib efficacy in additional tumor types [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics; 2019 Oct 26-30; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2019;18(12 Suppl):Abstract nr A008. doi:10.1158/1535-7163.TARG-19-A008

NIMG-20. EVALUATION OF HEAD SEGMENTATION QUALITY FOR TREATMENT PLANNING OF TUMOR TREATING FIELDS IN BRAIN TUMORS

Abstract Tumor treating fields (TTFields) is an FDA approved therapy for the treatment of glioblastoma multiform (GBM), malignant pleural mesothelioma (MPM), and currently being investigated for additional tumor types. TTFields are delivered to the tumor through the placement of transducer arrays (TAs) placed on the patient’s shaved scalp. The positions of the TAs are associated with treatment outcomes via simulations of the electric fields. Therefore, we are currently developing a method for recommending optimal placement of TAs. A key step to achieve this goal is to correctly segment the head into tissues of similar electrical properties. Visual inspection of segmentation quality is invaluable but time-consuming. Automatic quality assessment can assist in automatic refinement of the segmentation parameters, suggest flaw points to the user and indicate if the segmented method is of sufficient accuracy for TTFields simulation. As a first step in this direction, we identified a set of features that are relevant to atlas-based segmentation and show that these are significantly correlated (p &lt; 0.05) with a similarity measure between gold-standard and automatically computed segmentations. Furthermore, we incorporated these features in a decision tree regressor to predict the similarity of the gold-standard and computed segmentations of 20 TTFields patients using a leave-one-out approach. The predicted similarity measures were highly correlated with the actual ones (average absolute difference 3% (SD = 3%); r = 0.92, p &lt; 0.001). We conclude that automatic quality estimation of segmentations is feasible by incorporating segmentation-relevant features with statistical and machine learning methods, such as decision tree regressor.

Existing and Emerging Biomarkers for Immune Checkpoint Immunotherapy in Solid Tumors.

In the last few years, immunotherapy has transformed the way we treat solid tumors, including melanoma, lung, head neck, breast, renal, and bladder cancers. Durable responses and long-term survival benefit has been experienced by many cancer patients, with favorable toxicity profiles of immunotherapeutic agents relative to chemotherapy. Cures have become possible in some patients with metastatic disease. Additional approvals of immunotherapy drugs and in combination with other agents are anticipated in the near future. Multiple additional immunotherapy drugs are in earlier stages of clinical development, and their testing in additional tumor types is under way. Despite considerable early success and relatively fewer side effects, the majority of cancer patients do not respond to checkpoint inhibitors. Additionally, while the drugs are generally well tolerated, there is still the potential for significant, unpredictable and even fatal toxicity with these agents. Improved biomarkers may help to better select patients who are more likely to respond to these drugs. Two key biologically important predictive tissue biomarkers, specifically, PD-L1 and mismatch repair deficiency, have been FDA-approved in conjunction with the checkpoint inhibitor, pembrolizumab. Tumor mutation burden, another promising biomarker, is emerging in several tumor types, and may also soon receive approval. Finally, several other tissue and liquid biomarkers are emerging that could help guide single-agent immunotherapy and in combination with other agents. Of these, one promising investigational biomarker is alteration or deficiency in DNA damage response (DDR) pathways, with altered DDR observed in a broad spectrum of tumors. Here, we provide a critical overview of current, emerging, and investigational biomarkers in the context of response to immunotherapy in solid tumors.

SMARCA2-deficiency confers sensitivity to targeted inhibition of SMARCA4 in esophageal squamous cell carcinoma cell lines

SMARCA4/BRG1 and SMARCA2/BRM, the two mutually exclusive catalytic subunits of the BAF complex, display a well-established synthetic lethal relationship in SMARCA4-deficient cancers. Using CRISPR-Cas9 screening, we identify SMARCA4 as a novel dependency in SMARCA2-deficient esophageal squamous cell carcinoma (ESCC) models, reciprocal to the known synthetic lethal interaction. Restoration of SMARCA2 expression alleviates the dependency on SMARCA4, while engineered loss of SMARCA2 renders ESCC models vulnerable to concomitant depletion of SMARCA4. Dependency on SMARCA4 is linked to its ATPase activity, but not to bromodomain function. We highlight the relevance of SMARCA4 as a drug target in esophageal cancer using an engineered ESCC cell model harboring a SMARCA4 allele amenable to targeted proteolysis and identify SMARCA4-dependent cell models with low or absent SMARCA2 expression from additional tumor types. These findings expand the concept of SMARCA2/SMARCA4 paralog dependency and suggest that pharmacological inhibition of SMARCA4 represents a novel therapeutic opportunity for SMARCA2-deficient cancers.

Mechanisms of Lymphoma Clearance Induced by High-Dose Alkylating Agents.

The extraordinary activity of high-dose cyclophosphamide against some high-grade lymphomas was described nearly 60 years ago. Here we address mechanisms that mediate cyclophosphamide activity in bona fide human double-hit lymphoma. We show that antibody resistance within the bone marrow (BM) is not present upon early engraftment but develops during lymphoma progression. This resistance required a high tumor:macrophage ratio, was recapitulated in spleen by partial macrophage depletion, and was overcome by multiple, high-dose alkylating agents. Cyclophosphamide induced endoplasmic reticulum (ER) stress in BM-resident lymphoma cells in vivo that resulted in ATF4-mediated paracrine secretion of VEGFA, massive macrophage infiltration, and clearance of alemtuzumab-opsonized cells. BM macrophages isolated after cyclophosphamide treatment had increased phagocytic capacity that was reversed by VEGFA blockade or SYK inhibition. Single-cell RNA sequencing of these macrophages identified a "super-phagocytic" subset that expressed CD36/FCGR4. Together, these findings define a novel mechanism through which high-dose alkylating agents promote macrophage-dependent lymphoma clearance. SIGNIFICANCE: mAbs are effective against only a small subset of cancers. Herein, we recapitulate compartment-specific antibody resistance and define an ER stress-dependent mechanism induced by high-dose alkylating agents that promotes phagocytosis of opsonized tumor cells. This approach induces synergistic effects with mAbs and merits testing across additional tumor types.See related commentary by Duval and De Palma, p. 834.This article is highlighted in the In This Issue feature, p. 813.

Abstract 397: SMARCA4 is a selective vulnerability in SMARCA2-deficient esophageal squamous cell carcinoma cell lines

Abstract Esophageal squamous cell carcinoma (ESCC) is a highly prevalent tumor type with poor prognosis and limited treatment options. In this study, we aimed to discover potential novel drug targets using a pooled epigenome sgRNA viability screen in a panel of ESCC cell models. We identify the BAF chromatin remodeling helicase subunit SMARCA4/BRG1 as a novel dependency in a subset of ESCC cell lines. In line with the established synthetic lethal interaction of SMARCA4 and SMARCA2/BRM, the two mutual exclusive catalytic subunits of the BAF complex, SMARCA4-dependent cell lines display low or absent expression of SMARCA2. In rescue studies using SMARCA4 variants, we demonstrate that SMARCA4-dependency is linked to its ATPase/helicase activity, but not to bromodomain function. Similarly, ectopic expression of wild-type and bromodomain-mutant SMARCA2, but not a helicase-dead variant, rescues from loss of SMARCA4 in SMARCA2-low ESCC cells. SMARCA2-proficient ESCC cell models are rendered SMARCA4-dependent upon CRISPR/Cas9-mediated knock-out of SMARCA2. We further identify SMARCA4-dependent cell models from additional tumor types (colon, ovarian and pancreas carcinoma) with low or absent SMARCA2 expression. These findings expand the concept of SMARCA2/SMARCA4 paralog dependency and indicate pharmacological inhibition of SMARCA4 as a novel opportunity for targeted therapy of SMARCA2-deficient cancers. Citation Format: Katharina Ehrenhöfer-Wölfer, Teresa Puchner, Silvia Blaha-Ostermann, Alexandra Hörmann, Wolfgang Sommergruber, Norbert Schweifer, Thomas Zichner, Andreas Schlattl, Ralph A. Neumüller, Manfred Koegl, Mark P. Petronczki, Mark Pearson, Simon Wöhrle. SMARCA4 is a selective vulnerability in SMARCA2-deficient esophageal squamous cell carcinoma cell lines [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 397.

Abstract 2129: CPTAC: Biospecimen accrual for proteogenomics

Abstract The Clinical Proteomics Tumor Analysis Consortium (CPTAC) is a National Cancer Institute initiative that seeks to uncover the molecular basis of cancer using a proteogenomic approach to study prospective cancer specimens from treatment naïve patients. Leidos Biomed provides an infrastructure for supporting the collection of high quality biospecimens specific to proteogenomics and the corresponding clinical data, in addition to project and subcontract management for the program. The proportion of biospecimens so far from racial and ethnic minority patients are underrepresented, which preclude equitable research across all patient groups for cancer treatment. CPTAC program seeks diversity in the accrual cohort and an effort is ongoing to encourage the participation of tissue source sites serving racial and ethnic minorities that can contribute high quality biospecimens towards the collection. CPTAC applies the understanding of the molecular basis of cancer to identify biomarker candidates. CPTAC Phase II completed in 2016, collecting over 350 cases from breast, colon and ovarian patients. CPTAC Phase III that is in progress began to collect and analyze 200 cases from each of ten additional cancers. Analysis is close to completion for uterine corpus endometrial carcinoma, kidney clear cell renal cell carcinoma and lung adenocarcinoma while accruals are ongoing for 7 other cancer types. The study entails collection and pathology evaluation of biospecimens, high-quality clinical data and images from clinical sites from around the world. A biorepository evaluates and processes the biospecimens, sending nucleic acids to a sequencing center and tissues to proteomics groups. Data are combined and analyzed by analysis and translational centers. Genomic data are made available to the research community through the NCI Genomic Data Commons(GDC). Proteomic data are made available through CPTAC’s Data Coordinating Center(DCC). Imaging data are made available through The Cancer Imaging Archive (TCIA). The program is expanding and initiating biospecimen collection for proteogenomic analysis of ten additional rare tumor types. Citation Format: Mathangi Thiagarajan. CPTAC: Biospecimen accrual for proteogenomics [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 2129.

Abstract 1672: Clinical validation of a genomics-based classifier to predict tissue of origin from targeted tumor sequencing

Abstract Introduction Various classes of genomic alterations can now be robustly detected from targeted Next Generation Sequencing (NGS)-based tumor profiling assays. Despite recent advances in genome-directed cancer therapies and lineage-agnostic basket clinical trials, tissue of origin remains a critical determinant of tumor biology and therapeutic sensitivity. Therefore, efficiently harnessing the mutational information to predict tissue of origin from NGS data can further aid diagnosis and treatment. Methods We have developed a novel random forest machine learning classifier that infers tissue of origin from the mutational features of each tumor NGS profile. We trained and validated the model using prospective sequencing data from &amp;gt;30,000 patients profiled by MSK-IMPACT, a custom FDA-authorized clinical sequencing assay. An initial version of this classifier trained on a smaller cohort, presented previously, has been further optimized to incorporate additional tumor types, genomic features, mutational signatures, and accessibility features for clinicians. Altogether, the model evaluates a comprehensive range of mutational features to generate predictions. In addition, we developed a framework for the prospective clinical implementation of our method that allows for extension to the expanding MSK-IMPACT cohort and utilization at the point of care. Results Overall, we predicted the correct cancer type in 74% of cases, with nearly half of cases predicted with high confidence (&amp;gt;95%). In order to make this tool accessible to pathologists for real-time diagnostic and treatment decisions, we also have implemented APIs that transmit our classifier predictions to the cBioPortal for Cancer Genomics and the MPath console from MSK Molecular Diagnostics Service. With the additional samples and genomic features, our model predictions have been improved and are considered during clinical review and sign-out. This practice has brought about critical diagnostic changes and orthogonal validation in several cases. In one case, a patient referred to MSK with a misdiagnosis of metastatic breast cancer to the bladder underwent MSK-IMPACT sequencing for her primary breast tumor and bladder lesion. The model predictions, confirmed by orthogonal testing, classified two lesions as independent primaries of breast and bladder origins, leading to significant changes in treatment regimens. Conclusion Our work delineates the framework of utilizing and optimizing machine learning models on NGS-based sequencing data to aid diagnosis and treatment. The same framework can be applied to cell-free DNA sequencing, which will capture tumor spatial heterogeneity and improve classifier performance. These results indicate that leveraging machine learning to predict tissue of origin complements conventional histologic review to provide integrated diagnoses, often with critical therapeutic implications. Citation Format: Youyun Zheng, Alexander V. Penson, Niedzica Camacho, Evan Biederstedt, Ahmet Zehir, Cyriac Kandoth, Aijazuddin Syed, Anna M. Varghese, Hikmat A. Al-Ahmadie, Nikolaus Schultz, Marc Ladanyi, David B. Solit, David S. Klimstra, David Hyman, Barry Taylor, Michael F. Berger. Clinical validation of a genomics-based classifier to predict tissue of origin from targeted tumor sequencing [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 1672.

The cancer cell proteome and transcriptome predicts sensitivity to targeted and cytotoxic drugs.

Tumors of different molecular subtypes can show strongly deviating responses to drug treatment, making stratification of patients based on molecular markers an important part of cancer therapy. Pharmacogenomic studies have led to the discovery of selected genomic markers (e.g., BRAFV600E), whereas transcriptomic and proteomic markers so far have been largely absent in clinical use, thus constituting a potentially valuable resource for further substratification of patients. To systematically assess the explanatory power of different -omics data types, we assembled a panel of 49 melanoma cell lines, including genomic, transcriptomic, proteomic, and pharmacological data, showing that drug sensitivity models trained on transcriptomic or proteomic data outperform genomic-based models for most drugs. These results were confirmed in eight additional tumor types using published datasets. Furthermore, we show that drug sensitivity models can be transferred between tumor types, although after correcting for training sample size, transferred models perform worse than within-tumor-type predictions. Our results suggest that transcriptomic/proteomic signals may be alternative biomarker candidates for the stratification of patients without known genomic markers.

Association of an inflammatory gene signature with CD8 expression by immunohistochemistry (IHC) in multiple tumor types.

2593 Background: A multiparameter tumor inflammation assay based on gene expression profiling (TIA-GEP) can extend the utility of IHC to interrogate the tumor microenvironment (TME). Using CD8 expression assessed by IHC (CD8-IHC) as a surrogate for inflammation, statistical modelling was used to develop a specific gene signature on the TIA-GEP panel to predict CD8-IHC. The correlation between TIA-GEP and CD8-IHC and the prevalence of inflammation were explored across multiple tumor types. Methods: Levels of inflammation were measured by CD8-IHC and TIA-GEP on 1778 procured samples across 12 tumor types. Quality control metrics involved sample input quality, technical errors, and inter-run variability. Generalized linear models were used to identify an inflammation score that predicts the CD8-IHC score in melanoma and SCCHN tissue. The predictive accuracy of this signature was also examined in 10 additional tumor types. Results: Assessment of TME inflammation by CD8-IHC was consistent with that observed by TIA-GEP in multiple tumor types. The range of inflammation varied across different tumor types, with relatively lower inflammation range and scores in SCLC, ovarian, and prostate cancers, and higher values in NSCLC, melanoma, SCCHN, and gastric cancers. R2 x 100 values reflecting percent variation in CD8-IHC associated with TIA-GEP ranged from 62.4% to 79.2% ( P &lt; 0.0001) for all tumor types except prostate cancer (32.5%). Low correlation in prostate cancer may be a result of low prevalence of inflammation by CD8-IHC. Estimated linear regression slopes between CD8-IHC and TIA-GEP ranged from 0.74 in SCLC to 1.27 in gastric cancer. Conclusions: The results suggest that the inflammation signature is a robust potential diagnostic tool predicting inflammation in the TME. The inflammation signature not only correlates with CD8-IHC for multiple tumor types, but also leverages the alternative benefits associated with TIA-GEP, which include information related to tumor inflammation-associated biomarkers and flexibility in exploring the value of other genomic signatures.

Impact of in vitro driven expression signatures of CD133 stem cell marker and tumor stroma on clinical outcomes in gastric cancers

BackgroundThe CD133 transmembrane protein is a well-recognized stem cell marker that has been used to isolate putative cancer stem cell populations from gastric cancers (GCs). However, the molecular features or biomarkers underlying CD133 are largely unknown in GCs.MethodsWe performed gene expression profiling of CD133+ and CD133- cells sorted by flow cytometry from three GC cell lines to identify the CD133 expression signatures of GC. The CD133 expression signatures were investigated across publicly available expression profiles of multiple tumor types including GC and also for their relationship with patient survival.ResultsThe CD133 signature genes defined as 177 upregulated genes and 129 downregulated genes in CD133+ cells compared to CD133- cells were enriched with genes involving the cell cycle and cytoskeleton, implying that cancer stem cells with unlimited self-renewal play cancer-initiating roles. The CD133 expression signatures in GC expression profiles were positively correlated with those of brain tumors expressing CD133 and human embryonic stem cells, emphasizing the transcriptional similarities across stem cell-related expression signatures. We also found that these stem cell expression signatures were inversely correlated with those representing tumor infiltrating immune and stromal cells. Additionally, high CD133 expression signatures were found in intestinal subtypes and low tumor stage GCs as well as in those with microsatellite instabilities and high mutation burdens. As examined across 20 additional tumor types, both the expression signatures representing CD133 and stromal cells were unfavorable prognostic features; however, their impact were variable across tumor types.ConclusionsThe transcriptional activities of CD133 and those of stromal cells representing the activity of stem cells and level of epithelial-to-mesenchymal transition, respectively, may be inversely correlated with each other across multiple tumor types including GC. This relationship may be a confounding factor and should therefore be considered when evaluating the clinical relevance of stem cell-related markers.

Immunotherapy utilizing the combination of natural killer\u2013 and antibody dependent cellular cytotoxicity (ADCC)\u2013mediating agents with poly (ADP-ribose) polymerase (PARP) inhibition

BackgroundPoly (ADP-ribose) polymerase inhibitors (PARPi) prevent single-stranded DNA repair. Olaparib is a PARPi approved for the treatment of BRCA mutant ovarian and breast carcinoma. Emerging clinical data suggest a benefit of combining olaparib with immunotherapy in prostate cancer patients both with and without somatic BRCA mutations.MethodsWe examined if olaparib, when combined with IgG1 antibody-dependent cellular cytotoxicity (ADCC)-mediating monoclonal antibodies (mAbs) cetuximab (anti-EGFR), or avelumab (anti-PD-L1), would increase tumor cell sensitivity to killing by natural killer (NK) cells independently of BRCA status or mAb target upregulation. BRCA mutant and BRCA wildtype (WT) prostate carcinoma cell lines were pretreated with olaparib and then exposed to NK cells in the presence or absence of cetuximab or avelumab.ResultsNK-mediated killing was significantly increased in both cell lines and was further increased using the ADCC-mediating mAbs. Pre-exposure of NK cells to recombinant IL-15/IL-15Rα further increased the lysis of olaparib treated tumor cells. In addition, olaparib treated tumor cells were killed to a significantly greater degree by engineered high-affinity NK cells (haNK). We show here for the first time that (a) olaparib significantly increased tumor cell sensitivity to NK killing and ADCC in both BRCA WT and BRCA mutant prostate carcinoma cells, independent of PD-L1 or EGFR modulation; (b) mechanistically, treatment with olaparib upregulated death receptor TRAIL-R2; and (c) olaparib significantly enhanced NK killing of additional tumor types, including breast, non-small cell lung carcinoma, and chordoma.ConclusionsThese studies support the combined use of NK- and ADCC-mediating agents with correctly timed PARP inhibition.

Synthesis and SAR of novel capsazepine analogs with significant anti-cancer effects in multiple cancer types

We previously demonstrated that capsazepine (CPZ), a synthetic transient receptor potential Vanilloid subtype 1 (TRPV1) antagonist, has significant anti-cancer effects in vivo. The purpose of this study was to develop more potent analogs based upon CPZ pharmacophore and structure-activity relationships (SAR) across analogs. We generated 30 novel compounds and screened for their anti-proliferative effects in cultured HeLa cervical cancer cells. Cell viability assays identified multiple compounds with IC50s < 15 μM and one compound, 29 with an IC50 < 5 μM; six fold more potent than CPZ. We validated the anti-proliferative efficacy of two lead compounds, 17 and 29, in vivo using HeLa-derived xenografts in athymic nude mice. Both analogs significantly reduced tumor volumes by day 8 compared to control treated animals (p < 0.001) with no observable adverse effects. Calcium imaging determined that compound 17 activates TRPV1 whereas 29 neither activates nor inhibits TRPV1; indicating a unique mechanism-of-action that does not involve TRPV1 signaling. Cell viability assays using a panel of additional tumor types including oral squamous cell carcinoma, non-small cell lung cancer (NSCLC), breast cancer, and prostate cancer cell lines (HSC-3, H460, MDA-231, and PC-3 respectively) demonstrated that both lead compounds were efficacious against every cancer type tested. Compounds 29 displayed IC50s of 1-2.5 μM in HSC-3and PC-3cells. Thus, we propose that these novel CPZ analogs may serve as efficacious therapeutic agents against multiple tumor types that warrant further development for clinical application.

Heightened CXCR4 and CXCL12 expression in NF1-associated neurofibromas.

Neurofibromatosis type 1 (NF1) is a common autosomal dominantly inherited disorder that affects both the skin and the nervous system. NF1 occurs due to the mutations in the NF1 gene. Neurofibromas are the most common Schwann cell-based tumors in NF1 patients, which are mainly categorized into dermal and plexiform neurofibromas. Studies on different tumor types demonstrate that CXCR4 expression increases in tumor tissues and is linked to metastasis and cancer progression. In the present study, we aimed to analyze the gene expression of CXCR4, and its ligand CXCL12, in human neurofibromas. Eight NF1 patients aged between 5 and 37 (2 males, 6 females) were selected. The patient group comprised 1 plexiform neurofibroma, 1 pheochromocytoma, and 6 dermal neurofibromas. Following pathological examination and diagnosis, tumors were co-stained with antibodies against Schwann cell marker S100 and target molecule CXCR4. CXCR4 expression in Schwann cell-based tumors was detected at the protein level. RNA isolated from the same tumors was used for RT-PCR-based studies to measure the quantitative expression of CXCR4 and CXCL12. CXCR4 gene expression increased 3- to 120-fold and CXCL12 gene expression increased 33- to 512-fold in all human Schwann cell-based tumors. In order to validate the role of CXCR4 and its relationship with CXCL12 in NF1, future studies should be performed with additional tumors and different tumor types.

Updated Molecular Testing Guideline for theSelection of Lung Cancer Patients for Treatment With Targeted Tyrosine Kinase Inhibitors: Guideline From the College of American Pathologists, theInternational Association for the Study of Lung Cancer, and theAssociation for Molecular Pathology.

In 2013, an evidence-based guideline was published by the College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology to set standards for the molecular analysis of lung cancers to guide treatment decisions with targeted inhibitors. New evidence has prompted an evaluation of additional laboratory technologies, targetable genes, patient populations, and tumor types for testing. To systematically review and update the 2013 guideline to affirm its validity; to assess the evidence of new genetic discoveries, technologies, and therapies; and to issue an evidence-based update. The College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology convened an expert panel to develop an evidence-based guideline to help define the key questions and literature search terms, review abstracts and full articles, and draft recommendations. Eighteen new recommendations were drafted. The panel also updated 3 recommendations from the 2013 guideline. The 2013 guideline was largely reaffirmed with updated recommendations to allow testing of cytology samples, require improved assay sensitivity, and recommend against the use of immunohistochemistry for EGFR testing. Key new recommendations include ROS1 testing for all adenocarcinoma patients; the inclusion of additional genes (ERBB2, MET, BRAF, KRAS, and RET) for laboratories that perform next-generation sequencing panels; immunohistochemistry as an alternative to fluorescence in situ hybridization for ALK and/or ROS1 testing; use of 5% sensitivity assays for EGFR T790M mutations in patients with secondary resistance to EGFR inhibitors; and the use of cell-free DNA to "rule in" targetable mutations when tissue is limited or hard to obtain.