Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Evidence-Based Network Approach to Recommending Targeted Cancer Therapies

Evidence-Based Network Approach to Recommending Targeted Cancer Therapies original reports abstract SPECIAL SERIES: INFORMATICS TOOLS FOR CANCER RESEARCH AND CARE Evidence-Based Network Approach to Recommending Targeted Cancer Therapies 1 2 2 2 2 Jayaram Kancherla, MS ; Shruti Rao, MS, MBA ; Krithika Bhuvaneshwar, MS ; Rebecca B. Riggins, PhD ; Robert A. Beckman, MD ; 2 1 1 Subha Madhavan, PhD ;Hector ´ Corrada Bravo, PhD ; and Simina M. Boca, PhD PURPOSE In this work, we introduce CDGnet (Cancer-Drug-Gene Network), an evidence-based network ap- proach for recommending targeted cancer therapies. CDGnet represents a user-friendly informatics tool that expands the range of targeted therapy options for patients with cancer who undergo molecular profiling by including the biologic context via pathway information. METHODS CDGnet considers biologic pathway information specifically by looking at targets or biomarkers downstream of oncogenes and is personalized for individual patients via user-inputted molecular alterations and cancer type. It integrates a number of different sources of knowledge: patient-specific inputs (molecular al- terations and cancer type), US Food and Drug Administration–approved therapies and biomarkers (curated from DailyMed), pathways for specific cancer types (from Kyoto Encyclopedia of Genes and Genomes [KEGG]), gene-drug connections (from DrugBank), and oncogene information (from KEGG). We consider 4 different evidence-based categories for therapy recommendations. Our tool is delivered via an R/Shiny Web application. For the 2 categories that use pathway information, we include an interactive Sankey visualization built on top of d3.js that also provides links to PubChem. RESULTS We present a scenario for a patient who has estrogen receptor (ER)–positive breast cancer with FGFR1 amplification. Although many therapies exist for patients with ER-positive breast cancer, FGFR1 amplifications may confer resistance to such treatments. CDGnet provides therapy recommendations, including PIK3CA, MAPK, and RAF inhibitors, by considering targets or biomarkers downstream of FGFR1. CONCLUSION CDGnet provides results in a number of easily accessible and usable forms, separating targeted cancer therapies into categories in an evidence-based manner that incorporates biologic pathway information. JCO Clin Cancer Inform 4:71-88. © 2020 by American Society of Clinical Oncology Licensed under the Creative Commons Attribution 4.0 License INTRODUCTION some tumor types, it is now routine to check for specific molecular features to decide on targeted In today’s era of cancer precision medicine, thera- treatment plans. For example, KRAS wild-type co- peutic interventions are often tailored to an individual’s lorectal cancer is generally treated with epidermal tumor molecular profile, in addition to traditional growth factor receptor (EGFR) inhibitors, estrogen considerations, including age, sex, cancer stage, and ASSOCIATED receptor (ER)–positive breast cancer with aromatase CONTENT medical and treatment histories. The term molecular inhibitors or antiestrogens such as tamoxifen or Appendix profiling is often used to refer to a test that con- fulvestrant, and human epidermal growth factor Author affiliations siders ≥1 biomarker. These biomarkers may be either and support receptor 2–positive breast cancer with monoclonal genetic characteristics or mRNA or protein expression information (if antibodies trastuzumab and pertuzumab, tyrosine values. Genetic characteristics include point muta- applicable) appear at kinase inhibitors such as neratinib, or antibody-toxin tions, insertions, deletions, duplications, gene fusions, the end of this conjugates such as trastuzumab-DM1. In many and rearrangements. They may be either germ line article. cases, if there is no US Food and Drug Adminis- (inherited and present in normal tissue) or somatic Accepted on tration (FDA)–approved targeted therapy for a spe- December 4, 2019 (present in cancer cells but not normal tissue). and published at cific tumor type, clinicians may recommend either Expression values refer to the expression of mRNA ascopubs.org/journal/ an off-label therapy that is prescribed for the pa- or protein in tumors, either in comparison with other cci on January 28, tient’s alteration in another tumor type or enrollment tumors or adjacent normal tissue. Typically, tumor 2020: DOI https://doi. molecular profilingisusedwhenapatienthas few in a precision medicine clinical trial (eg, basket, org/10.1200/CCI.19. 00097 or no standard treatment options left. However, for umbrella, or targeted therapy trial). 71 Kancherla et al CONTEXT Key Objective With the increasing use of tumor molecular profiling, it is imperative to develop approaches that consider the biologic context to better prioritize targeted therapies for patients with cancer. Knowledge Generated We introduce CDGnet (Cancer-Drug-Gene Network), a tool that integrates biologic pathway information with alterations detected from tumor molecular profiling to expand the possible range of targeted therapies and prioritize them into evidence-based categories. We demonstrate how CDGnet can be applied to a scenario where a patient with breast cancer has overexpression of both ESR1 and FGFR1, with the output representing the recommended therapies, the clinical context in which they are approved, and the links between the patient’s tumor molecular profile and the recommendations. Relevance We consider this tool to be especially valuable to clinical and translational researchers who may be interested in understanding the best course of treatment for patients with a particular tumor molecular profile. To make such decisions about off-label therapy recom- a result, patients with colorectal cancer are typically tested mendations, clinicians have to sift through vast amounts of for KRAS mutations, and EGFR inhibitors are only pre- literature and clinical databases to determine the clinical scribed to individuals without specific KRAS mutations in utility of variants identified through molecular profiling to codons 12 and 13. A comprehensive characterization of decide on the appropriate treatment option for their pa- untreated colorectal tumors estimated that 43% of non- tients. The same is true for clinical translational scientists hypermutated tumors had KRAS mutations, and these considering relevant therapeutic approaches to evaluate in mutations were generally oncogenic activating mutations, model systems or humans, using either single agents or which means that a large percentage of patients with co- combinations. In this setting, the number of possible lorectal cancer are left with few therapeutic options. Our molecular profiles that may be relevant and the number of framework and tool are seeking to remedy this issue. experimental agents create a combinatorial explosion of METHODS research possibilities among which prioritization is needed. Overview of Methods for Generating Several efforts are ongoing to capture, standardize, and Patient-Specific Networks share clinically relevant variants identified through mo- lecular diagnostic tests among several public, academic, The user inputs into CDGnet are the specific alterations 3-5 and private institutions, although challenges remain in found in a patient’s tumor and the patient’s cancer type. synthesizing evidence in a manner that is both systematic Part of the landing page is shown in Figure 2. These data 6,7 and timely. Our goal in this work is to expand the range are then integrated with biologic networks relevant to the of options for targeted therapies for patients with cancer cancer type (from the Kyoto Encyclopedia of Genes and who undergo molecular profiling by developing CDGnet Genomes [KEGG] database ), FDA-approved targeted (Cancer-Drug-Gene Network), a user-friendly, evidence- cancer therapies and indications (curated from DailyMed based approach that accounts for downstream effects therapy labels ), additional gene-drug connections in within pathways in cancer and is personalized for the in- the form of drug targets (from the DrugBank database ), dividual patient. Our tool, which uses the Shiny framework information on whether a gene is an oncogene (from KEGG). 8 9 with an R backend, is available online. We incorporate Users may consider different data sources by using the pathway information specifically by looking at downstream CDGnet code directly, for example, by considering the targets of oncogenes, which are genes that are constitu- predicted oncogenes from a recent comprehensive charac- 10 16 tively activated in cancer. This is illustrated in Figure 1.If terization of The Cancer Genome Atlas (TCGA) projects. an oncogene in a biologic pathway is activated, targeting Currently, the biologic networks we consider are the genes and proteins that are found upstream may no cancer-specific pathways in KEGG, and therefore, for now, longer be effective, leading to a focus on downstream we are also restricting the cancer types to those that have targets. This includes the scenario of EGFR inhibitors for KEGG pathways. We have developed 4 different therapy KRAS wild-type colorectal tumors. The EGFR protein categories that can be prioritized for patients, given their triggers a signaling cascade in cancer, which may be specific tumor alterations, ordered from “most evidence blocked by anti-EGFR drugs; however, this is only ef- that therapy works” to “least evidence that therapy works.” fective if KRAS, which is downstream of EGFR, is not (1) FDA-approved drugs for which the input genes/proteins mutated. Otherwise, certain KRAS mutations can lead to are biomarkers for their tumor type; (2) FDA-approved a lack of effectiveness of therapies that block EGFR. As drugs for which the input genes/proteins are biomarkers 72 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies Tumor with no oncogenic Tumor with oncogenic activation of protein B activation of protein B i. No therapy ii. Targeted therapy iii. No therapy iv. Targeted therapy Drug Drug A A A A against against protein A protein A B B C C C D D D E E E Abnormal Abnormal Abnormal cancer-related cancer-related cancer-related processes processes processes FIG 1. Simplified diagram showing the reasoning behind looking at downstream targets of activated oncogenes. A simple pathway is shown that consists of 5 proteins, A, B, C, D, and E, with A activating B, B activating C, and so on, with the final activation of E leading to various abnormal cancer-related processes. (i, ii) Scenario where a tumor has no oncogenic activation of protein B. (iii, iv) Scenario where protein B has gained an oncogenic mutation that renders it constitutively active. If there is no oncogenic activation of protein B, then targeting protein A, as in (ii), may be effective in stopping cancer growth. However, if there is oncogenic activation of protein B, this means that, in particular, it is not necessary for protein A to activate protein B, so that targeting protein A is not effective for turning off the pathway. in other tumor types; (3) drugs that have as targets the input We differentiate between targets and biomarkers because genes/proteins or as biomarkers/targets other genes/ in many cases, as a result of complicated biologic in- proteins that are downstream of the input oncogenes when teractions, the target of a therapy may be different from the considering the pathway corresponding to this tumor type; biomarker used to specify the indication, such as in the case and (4) drugs that have as biomarkers/targets other genes/ of EGFR inhibitors being administered for KRAS wild-type proteins that are downstream of input oncogenes when colorectal tumors or CDK4/6 inhibitors being administered for considering the pathways corresponding to other tumor ER-positive breast tumors. The general approach is pre- types. sented in Figure 3. The options used on the landing page to obtain the different therapy categories are shown in Figure 2. In categories 3 and 4, users have the option to consider only We also provide documentation for the tool, including a step- FDA-approved targeted cancer therapies, all FDA-approved by-step analysis for the built-in patient use case scenario. therapies, or all drugs in DrugBank; this allows clinical re- searchers to consider increasing numbers of therapies only Appendix Table A1 lists FDA-approved targeted cancer as needed, as opposed to being overwhelmed with a huge therapies and indications that were obtained by consider- number of therapies from the start. We also note the differ- ing the targeted therapies listed by the National Cancer ence between categories 3 and 4; category 3 considers Institute and looking up the corresponding labels via the biologic pathway corresponding to the individual’s DailyMed. In particular, the indications and usage portion cancer type, whereas category 4 considers the pathways of the label was used to obtain the specificcancertype and corresponding to other cancer types. Given that pathways biomarker information, which is listed in the “Gene/Protein,” represent a simplification of a more complicated reality and “Data Type,” and “Alteration” columns; in the case of each tumor is unique, we found it necessary to allow for multiple biomarkers, these are listed in separate rows of possible connections between genes and proteins that may be the table. In cases where the biomarker indication is curated in cancer types different from that with which a patient unclear, the lists of FDA companion diagnostic tests were 19,20 presents, although in our experience, it is generally sufficient to also consulted. Note that although some targeted stop at category 3 therapies. therapies have specific biomarker indications, many do not. JCO Clinical Cancer Informatics 73 Kancherla et al FIG 2. Part of the landing page, which shows how users can select the cancer type and either input a tab-separated or comma-separated file or use the example data. The inset shows how under “Filter Recommended Therapies,” combinations of the first 2 checkboxes lead to the 4 different categories of therapy recommendations described in the text. Removing 1 or both of the last 2 checkboxes expands the range of therapies in categories 3 and 4 beyond US Food and Drug Administration (FDA)–approved drugs and FDA-approved targeted cancer drugs, respectively. For example, ibrutinib is a targeted therapy, administered for parsed and had identifiers converted using the KEGGREST, 22 23 a number of subtypes of leukemia/lymphoma, but not for KEGGgraph, and org.Hs.eg.db Bioconductor packages, a specific biomarker indication. If there is no biomarker respectively, and against the information input by the user, indication, this is noted as an asterisk in the table in the with the gene/protein names being normalized via the rDGIdb “Gene/Protein” column. The therapies are then cross- package, which is a wrapper for the Drug Gene Interaction referenced with DrugBank to obtain the targets for both 24,25 Database. the therapies with biomarker indications and those without To obtain the list of FDA-approved drugs, we used the data indications. The biomarkers and targets obtained in these files from the official Drugs@FDA resource. Drugs@FDA ways are checked against downstream targets from KEGG cancer-specific pathways, which were downloaded and contains several tab-separated value files that include Category 1: Patient-specific inputs Therapy recommendations FDA-approved drugs for these alterations in this Category 3: Molecular alterations cancer type (ie, the alterations represent Drugs that have as targets the input biomarkers in this genes/proteins or as biomarkers/targets other Cancer type cancer type) genes/proteins that are downstream of the input oncogenes when considering the pathway corresponding to this tumor type Category 2: FDA-approved FDA-approved drugs for Category 4: targeted therapies these alterations in other Drugs that have as biomarkers/targets other and biomarkers cancer types (ie, the genes/proteins that are downstream of input alterations represent oncogenes when considering the pathways Curation from biomarkers in other corresponding to other tumor types DailyMed cancer types) Pathways Gene-drug Oncogene for specific connections information cancer types KEGG DrugBank KEGG FIG 3. General approach for targeted therapy recommendations, including specific data sources. FDA, US Food and Drug Administration; KEGG, Kyoto Encyclopedia of Genes and Genomes. 74 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies information on the submission, review, and approval pro- column represents the pathway between the altered gene/ cess for various drugs. We use the products (list of all drugs) protein and the gene/protein that is a biomarker or target; and submission (review process for all drugs) files to filter for the alteration column represents the biomarker for an FDA- drugs that are approved or tentatively approved. The Drugs@ approved indication, if this exists, in which case the tumor FDA resource contains a list of all drugs approved since 1939, for which it is approved is also listed; the predicted effect some of which may have been discontinued. As a result, we column has the value “sensitive”, if the alteration column is use the marketingstatus file to remove any discontinued not empty, or “target”, if the drug targets the protein products from the list. The R scripts to parse and filter the according to the DrugBank data. Drugs@FDA data files are available in our GitHub repository. An architecture diagram for our system is shown in Shiny App and Visualization Figure 6. We use Shiny, an R package/framework for creating interactive and standalone Web applications di- For each of the 4 categories detailed, a sortable and rectly from R. Shiny applications can run on a Web page or searchable table of therapies is output with the FDA- be embedded in RMarkdown documents to build in- approved indications; for categories 3 and 4, network vi- teractive dashboards. They use the same technology that sualizations are also shown. The table also provides the powers Web applications (ie, HTML and JavaScript) and tumor type in which a particular therapy is approved. allow users to create intuitive and interactive user interfaces Figure 4 shows a Sankey flow diagram representation that and prototypes with an R computational backend. focuses on the flow of evidence between drug-gene and gene-gene connections, enabling an intuitive visualization To support interactive Sankey charts within Shiny appli- from the molecular profile to the inferred targets and cations, we developed a Shiny Web component for vi- recommended therapies. Figure 5 shows a portion of the sualizing Sankey flow diagrams, available to download sortable and searchable corresponding table. The path as an R package. Web components are custom HTML FIG 4. Sankey flow diagram focusing on the flow of evidence between drug-gene and gene connections for a putative patient with estrogen receptor–positive breast cancer and FGFR1 overexpression, showing category 3 recommendations, namely, targets downstream of FGFR1. Therapies can be clicked to obtain a panel with PubChem information. JCO Clinical Cancer Informatics 75 Kancherla et al FIG 5. Part of the sortable, searchable table for therapies in category 3 for a putative patient with estrogen receptor–positive breast cancer and FGFR1 overexpression, showing the subset of therapies that target MAP2K1. elements that are natively extensible and reusable and a pathway connection or information on a drug when can be integrated into any framework that supports HTML. a user selects/clicks on an edge or node. Selecting an The Sankey visualization uses a custom 3-column layout edge shows the downstream pathway information used to organize nodes in the graph: molecular profile and for inference. Selecting a recommended therapy dis- FDA-approved drugs, inferred targets, and recommen- plays the structure of the drug and linked publications ded therapies; it intuitively focuses the user on the flow from PubChem, using PubChem widgets. The Sankey of evidence from input parameters to recommended visualization is built on top of d3.js, a data visuali- therapies. The Sankey visualization also contains an zation library for JavaScript to build highly customizable information panel that displays evidence related to and interactive visualizations. User interface Server User input or loaded example Downloaded data from � Mo lecular alterations � C urated therapy labels � Ca ncer type � KEGG � D rugBank R/Shiny R backend computations R/Shiny Display of therapy Therapy recommendations recommendations Tables nfpmShinyComponent R package Sankey diagrams FIG 6. Architecture diagram for our system. KEGG, Kyoto Encyclopedia of Genes and Genomes. 76 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies RESULTS can be targeted by different cancer therapies. On the basis of the level of evidence, the clinical actionability of these We will now consider the scenario of a patient who has ER- pathways can be further tested in a laboratory or clinical trial positive breast cancer. ER-positive breast cancer, generally setting. Additionally, there is a growing field of research related treated with aromatase inhibitors or antiestrogens, employs to drug-target interactions and drug repositioning using an array of mechanisms that permit escape from these 37-40 network-based models, which may in the future be in- therapies. These include amplification or upregulation tegrated with our tool. of fibroblast growth factor receptor 1 (FGFR1), which is amplified in approximately 13% of ER-positive tumors from We aim to further enhance the data that drive the CDGnet 30-32 33 TCGA and leads to ligand-independent ER activation. tool by incorporating relevant information from additional FGFR activity has also recently been shown to confer re- precision oncology efforts, tools, and resources. Users who sistance to CDK4/6 inhibitors in ER-positive breast download or connect to these resources may currently use cancer. Pan-FGFR antagonists have been combined them in the context of our approach by modifying our with endocrine therapies in prior clinical studies (eg, code. Expert-curated precision oncology databases CTKI258A2210), but the efficacy of this combination has include Clinical Interpretations of Variants in Cancer 5,41 42,43 44,45 been minimal, even in patients preselected for alterations in (CIViC), Cancer Genome Interpreter, OncoKB, 46,47 the FGFR pathway. A potential underlying explanation for Database of Evidence for Precision Oncology (DEPO), 48,49 this lack of benefit is that FGFR alterations impinge upon and Precision Medicine Knowledge Base (PMKB), and downstream signaling networks shared by many other more general resources include ClinVar. These additional receptor tyrosine kinases. Figure 4 shows CDGnet rec- sources may further strengthen the clinical annotations ommendations for a patient with breast cancer with over- and evidence related to germ line and somatic alterations in expression of both ESR1 (gene encoding ER) and FGFR1, our database and provide options between curated drug when considering only FDA-approved targeted therapies. label information and DrugBank targets. CIViC is an open- Therapy recommendations include PIK3CA, MAPK, and access, open-source, community-driven Web resource that RAF inhibitors, which may have utility in this context, along allows clinical interpretations of mutations related to can- with the standard targeted therapies prescribed for ER- cer. Cancer Genome Interpreter is an online tool that positive breast cancer. Figure 5 shows the subset of the connects genes and drugs along with their effects and corresponding table that consists of FDA-approved publication sources, not in a network format, but in a tab- MAP2K1 inhibitors, which are approved for either ABL1 ular format. OncoKB is another online precision oncology fusions in chronic myeloid leukemia, or specific BRAF knowledge base that contains information about the effects mutations in melanoma, non–small-cell lung cancer, and and treatment implications of specific cancer gene alter- anaplastic thyroid cancer. ations. DEPO contains druggable variant information such as drug therapies, evidence levels (FDA approved, clinical DISCUSSION trials, case reports, and preclinical), and cancer types for We developed the CDGnet tool using an approach that intended treatments. PMKB provides information about considers biologic pathways and connections among clinical cancer variants and interpretations. We are also genes, proteins, and drugs to prioritize targeted therapies using a simplified model for incorporating pathway in- for patients with cancer. Our approach integrates many formation via the consideration of targets that are down- disparate sources of knowledge and provides results in an stream of oncogenes; there are scenarios we do not capture easily accessible and usable format. With our tool, users are where upstream targeting can also be useful, for example, able to quickly obtain information on the FDA-approved 51-53 in the case of a feedback loop. We will incorporate more therapies (category 1) and potential off-label therapies complex information in future iterations of our tool. Our tool (category 2) associated with a patient’s molecular profile. partly relies on manual curation of information for FDA- Our definitions of categories 1 and 2 in CDGnet are in approved targeted therapies and thus has challenges alignment with the tier I and II evidence level classifications similar to those of other tools in this space, including the recommended by the Association for Molecular Pathology, time- and labor-intensive nature of this process. However, American College of Medical Genetics and Genomics, the KEGG and DrugBank components only need to be ASCO, and College of American Pathologists. However, downloaded and reprocessed through our existing code CDGnet categories 3 and 4 are unique to our evidence- when updates are desired. based network approach and enable users to evaluate additional targeted therapy options based on an individual’s Consortia such as the Clinical Genome Resource Somatic 3,54 tumor profile. It is important to note that the targeted therapy Cancer Working Group and the Global Alliance for recommendations in categories 3 and 4 have lower evidence Genomics and Health Variant Interpretation for Cancer levels and may or may not have proven clinical significance Consortium have ongoing efforts to standardize and in ongoing clinical trials. However, by examining the down- harmonize the expert-curated data in these different stream targets of candidate biomarkers, clinical researchers knowledge bases, with the goal of enhancing the in- can derive key insights into potential biologic pathways that teroperability among these databases. We will align the JCO Clinical Cancer Informatics 77 Kancherla et al future development of CDGnet with the guidelines and a personalized tool may eventually expand the range of consensus frameworks developed by these consortia. options of targeted therapies for patients with cancer in the CDGnet can also serve as an informative tool for oncolo- clinical setting, a key goal of precision oncology. gists, molecular pathologists, and genomic scientists who We currently consider clinical or translational researchers routinely participate in molecular tumor board discussions. to be the primary target user group for our tool. For in- 56,57 Tools similar to CDGnet include PreMedKB and the stance, if they are interested in a particular combination of Drug Gene Interaction Network. PreMedKB is an in- molecular alterations for a specific cancer type and gen- tegrated precision medicine knowledgebase for interpret- erally find the recommendations to be for drugs prescribed ing relationships among diseases, genes, variants, and in a different cancer type, they may decide to pursue formal drugs. The Drug Gene Interaction Network is a commercial studies of drug repurposing, which is made easier by tool offered by Seqome (Tramore, Ireland) that builds drug- knowing whether they are considering an FDA-approved gene interaction networks to predict clinical response from targeted drug, FDA-approved drug, or non–FDA-approved multiomics data sets. The advantage of CDGnet over these drug. Our eventual goal is to allow for the use of this tool by tools is that our approach allows users to input specific clinicians, especially for the care of patients with advanced- alterations found in a patient’s tumor and cancer type and stage disease for whom the immediate FDA-approved outputs therapy options ordered based on priority. Such therapy choices have been exhausted. AFFILIATIONS AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF University of Maryland, College Park, MD INTEREST Georgetown University, Washington, DC The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless Preprint version available on bioRxiv. otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the CORRESPONDING AUTHOR subject matter of this manuscript. For more information about ASCO’s Simina M. Boca, PhD, Georgetown University Medical Center, 2115 conflict of interest policy, please refer to www.asco.org/rwc or ascopubs. Wisconsin Ave, Suite 110, Washington, DC 20007; e-mail: smb310@ org/cci/author-center. georgetown.edu. Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments). EQUAL CONTRIBUTION J.K. and S.R. contributed equally to this work. Shruti Rao Research Funding: Symphony (Inst) SUPPORT Robert A. Beckman Supported by National Institutes of Health grant No. R21CA220398 and Leadership: Onco-Mind supplements No. R21CA220398-02 (S.M.B.), U01HG007437 (S.M.), Stock and Other Ownership Interests: Johnson & Johnson U24CA237719 (S.M.), and P30CA051008 (S.M.B., via pilot award); Consulting or Advisory Role: AstraZeneca, Zymeworks, Vertex, Department of Defense Breast Cancer Research Program award No. EMDSerono, CStone. W81XWH-17-1-0615 (R.B.R.); National Institute of General Medical Patents, Royalties, Other Intellectual Property: Two patents for dynamic Sciences grant No. R01GM114267 (H.C.B.); and National Institute of precision medicine, a novel approach to precision medicine, have been Mental Health grant No. R24MH114815 (H.C.B.). granted in Japan and Taiwan and transferred to Onco-Mind; these patents are pending in the United States and European Union. Subha Madhavan AUTHOR CONTRIBUTIONS Leadership: Perthera Conception and design: Robert A. Beckman, Subha Madhavan, Hector Stock and Other Ownership Interests: Perthera Corrada Bravo, Simina M. Boca Consulting or Advisory Role: Perthera Financial support: Rebecca B. Riggins, Subha Madhavan, Hector Corrado Research Funding: Teewinot Life Sciences (Inst) Bravo, Simina M. Boca Administrative support: Simina M. Boca Hector Corrada Bravo Collection and assembly of data: Jayaram Kancherla, Shruti Rao, Simina Consulting or Advisory Role: Genentech/Roche M. Boca Travel, Accommodations, Expenses: Genentech/Roche Data analysis and interpretation: Krithika Bhuvaneshwar, Rebecca B. Simina M. Boca Riggins, Robert A. Beckman, Subha Madhavan, Simina M. Boca Research Funding: Symphogen (Inst) Manuscript writing: All authors Final approval of manuscript: All authors Accountable for all aspects of the work: All authors No other potential conflicts of interest were reported. 78 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies REFERENCES 1. Amado RG, Wolf M, Peeters M, et al: Wild-type KRAS is required for panitumumab efficacy in patients with metastatic colorectal cancer. J Clin Oncol 26:1626-1634, 2008 2. Waks AG, Winer EP: Breast cancer treatment: A review. JAMA 321:288-300, 2019 3. Ritter DI, Roychowdhury S, Roy A, et al: Somatic cancer variant curation and harmonization through consensus minimum variant level data. Genome Med 8:117, 2016 4. Dolman L, Page A, Babb L, et al: ClinGen advancing genomic data-sharing standards as a GA4GH driver project. Hum Mutat 39:1686-1689, 2018 5. Griffith M, Spies NC, Krysiak K, et al: CIViC is a community knowledgebase for expert crowdsourcing the clinical interpretation of variants in cancer. Nat Genet 49:170-174, 2017 6. Boca SM, Panagiotou OA, Rao S, et al: Future of evidence synthesis in precision oncology: Between systematic reviews and biocuration. JCO Precis Oncol 10.1200/PO.17.00175 7. Rao S, Beckman RA, Riazi S, et al: Quantification and expert evaluation of evidence for chemopredictive biomarkers to personalize cancer treatment. Oncotarget 8:37923-37934, 2017 8. Chang W, Cheng J, Allaire JJ, et al: Shiny: Web application framework for R. https://rdrr.io/cran/shiny/ 9. CDGnet: Therapy recommendations using biological networks. http://epiviz.cbcb.umd.edu/shiny/CDGnet/ 10. Vogelstein B, Kinzler KW: Cancer genes and the pathways they control. Nat Med 10:789-799, 2004 11. Cancer Genome Atlas Network: Comprehensive molecular characterization of human colon and rectal cancer. Nature 487:330-337, 2012 12. Kanehisa M, Furumichi M, Tanabe M, et al: KEGG: New perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res 45:D353-D361, 2017 13. US National Library of Medicine: DailyMed. https://dailymed.nlm.nih.gov/dailymed/ 14. Wishart DS, Feunang YD, Guo AC, et al: DrugBank 5.0: A major update to the DrugBank database for 2018. Nucleic Acids Res 46:D1074-D1082, 2018 15. CDGnet: Repository for code related to the CDGnet tool for prioritizing targeted cancer therapies. https://github.com/SiminaB/CDGnet/ 16. Bailey MH, Tokheim C, Porta-Pardo E, et al: Comprehensive characterization of cancer driver genes and mutations. Cell 173: 371-385.e18, 2018 [Erratum: Cell 174:1034-1035, 2018] 17. CDGnet: CDGnet documentation. https://github.com/SiminaB/CDGnet/blob/master/CDGnet_documentation.pdf 18. National Cancer Institute: Targeted cancer therapies. https://www.cancer.gov/about-cancer/treatment/types/targeted-therapies/targeted-therapies-fact-sheet 19. US Food and Drug Administration: List of cleared or approved companion diagnostic devices (in vitro and imaging tools). https://www.fda.gov/medicaldevices/ productsandmedicalprocedures/invitrodiagnostics/ucm301431.htm 20. US Food and Drug Administration: Nucleic acid based tests. https://www.fda.gov/medical-devices/vitro-diagnostics/nucleic-acid-based-tests 21. Tenenbaum D: KEGGREST: Client-side REST access to KEGG. https://rdrr.io/bioc/KEGGREST/ 22. Zhang JD, Wiemann S: KEGGgraph: A graph approach to KEGG PATHWAY in R and Bioconductor. Bioinformatics 25:1470-1471, 2009 23. Carlson M: Genome wide annotation for human. http://bioconductor.org/packages/release/data/annotation/html/org.Hs.eg.db.html 24. Thurnherr T, Singer F, Stekhoven DJ, et al: Genomic variant annotation workflow for clinical applications. F1000 Res 5:1963, 2016 25. Wagner AH, Coffman AC, Ainscough BJ, et al: DGIdb 2.0: Mining clinically relevant drug-gene interactions. Nucleic Acids Res 44:D1036-D1044, 2016 26. US Food and Drug Administration: Drugs@FDA data files. http://www.fda.gov/Drugs/InformationOnDrugs/ucm079750.htm 27. Shiny Component: API for interactive network visualization of gene-drug and gene-gene interactions. https://github.com/jkanche/nfpmShinyComponent 28. Kim S, Chen J, Cheng T, et al: PubChem 2019 update: Improved access to chemical data. Nucleic Acids Res 47:D1102-D1109, 2019 29. Bostock M, Ogievetsky V, Heer J: D : Data-driven documents. IEEE Trans Vis Comput Graph 17:2301-2309, 2011 30. Gao J, Aksoy BA, Dogrusoz U, et al: Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 6:pl1, 2013 31. Cerami E, Gao J, Dogrusoz U, et al: The cBio cancer genomics portal: An open platform for exploring multidimensional cancer genomics data. Cancer Discov 2:401-404, 2012 32. Ciriello G, Gatza ML, Beck AH, et al: Comprehensive molecular portraits of invasive lobular breast cancer. Cell 163:506-519, 2015 33. Formisano L, Stauffer KM, Young CD, et al: Association of FGFR1 with ERαLmaintains ligand-independent ER transcription and mediates resistance to estrogen deprivation in ER+ breast cancer. Clin Cancer Res 23:6138-6150, 2017 [Erratum: Clin Cancer Res 25:1433, 2019] 34. Formisano L, Lu Y, Servetto A, et al: Aberrant FGFR signaling mediates resistance to CDK4/6 inhibitors in ER+ breast cancer. Nat Commun 10:1373, 2019 35. Perez-Garcia J, Muñoz-Couselo E, Soberino J, et al: Targeting FGFR pathway in breast cancer. Breast 37:126-133, 2018 36. Li MM, Datto M, Duncavage EJ, et al: Standards and guidelines for the interpretation and reporting of sequence variants in cancer: A joint consensus recommendation of the Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists. J Mol Diagn 19:4-23, 37. Zhang W, Chien J, Yong J, et al: Network-based machine learning and graph theory algorithms for precision oncology. NPJ Precis Oncol 1:25, 2017 38. Jiang XL, Martinez-Ledesma E, Morcos F: Revealing protein networks and gene-drug connectivity in cancer from direct information. Sci Rep 7:3739, 2017 39. Zhang Y, Tao C: Network analysis of cancer-focused association network reveals distinct network association patterns. Cancer Inform 13:45-51, 2014 (suppl 3) [Erratum: Cancer Inform 13:89, 2014 (suppl 3)] 40. Cheng F, Liu C, Jiang J, et al: Prediction of drug-target interactions and drug repositioning via network-based inference. PLOS Comput Biol 8:e1002503, 2012 41. CIViC: Clinical Interpretations of Variants in Cancer. https://civicdb.org/ 42. Tamborero D, Rubio-Perez C, Deu-Pons J, et al: Cancer Genome Interpreter annotates the biological and clinical relevance of tumor alterations. Genome Med 10:25, 2018 43. Cancer Genome Interpreter: Cancer Biomarkers database. https://www.cancergenomeinterpreter.org/biomarkers 44. Chakravarty D, Gao J, Phillips SM, et al: OncoKB: A precision oncology knowledge base. JCO Precis Oncol 10.1200/PO.17.00011 45. OncoKB: Precision oncology knowledge base. https://oncokb.org/ 46. Sun SQ, Mashl RJ, Sengupta S, et al: Database of evidence for precision oncology portal. Bioinformatics 34:4315-4317, 2018 47. DEPO: Database of Evidence for Precision Oncology. http://depo-dinglab.ddns.net/ 48. Huang L, Fernandes H, Zia H, et al: The cancer precision medicine knowledge base for structured clinical-grade mutations and interpretations. J Am Med Inform Assoc 24:513-519, 2017 49. PMKB: Precision Medicine Knowledge Base. https://pmkb.weill.cornell.edu/ JCO Clinical Cancer Informatics 79 Kancherla et al 50. Landrum MJ, Lee JM, Benson M, et al: ClinVar: Public archive of interpretations of clinically relevant variants. Nucleic Acids Res 44:D862-D868, 2016 51. Prahallad A, Sun C, Huang S, et al: Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR. Nature 483:100-103, 52. Solit DB, Janne PA: Translational medicine: Primed for resistance. Nature 483:44-45, 2012 53. Chandarlapaty S, Sawai A, Scaltriti M, et al: AKT inhibition relieves feedback suppression of receptor tyrosine kinase expression and activity. Cancer Cell 19:58-71, 2011 54. Madhavan S, Ritter D, Micheel C, et al: ClinGen Cancer Somatic Working Group: Standardizing and democratizing access to cancer molecular diagnostic data to drive translational research. Pac Symp Biocomput 23:247-258, 2018 55. Wagner AH, Walsh B, Mayfield G, et al: A harmonized meta-knowledgebase of clinical interpretations of cancer genomic variants. https://www.biorxiv.org/ content/10.1101/366856v1 56. PreMedKB: Precision Medicine Knowledgebase. http://www.fudan-pgx.org/premedkb/index.html#/home 57. Yu Y, Wang Y, Xia Z, et al: PreMedKB: An integrated precision medicine knowledgebase for interpreting relationships between diseases, genes, variants and drugs. Nucleic Acids Res 47:D1090-D1101, 2019 58. Seqome: Drug Gene Interaction Network. http://seqome.com/drug-gene-network/ nn n 80 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies APPENDIX TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source ER Abemaciclib (Verzenio) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Abemaciclib (Verzenio) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Abiraterone Prostate cancer Target NCI/FDA acetate (Zytiga) * Acalabrutinib (Calquence) Mantle cell lymphoma Target NCI/FDA HER2 Ado-trastuzumab emtansine (Kadcyla) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression EGFR Afatinib dimaleate (Gilotrif) NSCLC Sensitive Mutation Exon 19 deletion NCI/FDA EGFR Afatinib dimaleate (Gilotrif) NSCLC Sensitive Mutation L858R NCI/FDA * Afatinib dimaleate (Gilotrif) NSCLC Target NCI/FDA ALK Alectinib (Alecensa) NSCLC Sensitive Mutation Rearrangement NCI/FDA * Alemtuzumab (Campath) B-cell chronic Target NCI/FDA lymphocytic leukemia * Alitretinoin (Panretin) Kaposi sarcoma Target NCI/FDA ER Anastrozole (Arimidex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Anastrozole (Arimidex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Apalutamide (Erleada) Prostate cancer Target NCI/FDA * Atezolizumab (Tecentriq) NSCLC Target NCI/FDA PD-L1 Atezolizumab (Tecentriq) Urothelial carcinoma Sensitive Gene or protein Overexpression NCI/FDA expression * Atezolizumab (Tecentriq) Urothelial carcinoma Target NCI/FDA * Avelumab (Bavencio) Urothelial carcinoma Target NCI/FDA * Avelumab (Bavencio) Merkel cell carcinoma Target NCI/FDA * Axicabtagene ciloleucel (Yescarta) Large B-cell lymphoma Target NCI/FDA * Axitinib (Inlyta) Kidney cancer Target NCI/FDA * Belinostat (Beleodaq) Peripheral T-cell Target NCI/FDA lymphoma * Bevacizumab (Avastin) Glioblastoma Target NCI/FDA * Bevacizumab (Avastin) Cervical cancer Target NCI/FDA * bevacizumab (Avastin) Colorectal cancer Target NCI/FDA * Bevacizumab (Avastin) Fallopian tube cancer Target NCI/FDA * Bevacizumab (Avastin) Renal cell carcinoma Target NCI/FDA * Bevacizumab (Avastin) NSCLC Target NCI/FDA * Bevacizumab (Avastin) Ovarian cancer Target NCI/FDA * Bevacizumab (Avastin) Primary peritoneal Target NCI/FDA cancer * Bexarotene (Targretin) Cutaneous T-cell Target NCI/FDA lymphoma * Blinatumomab (Blincyto) B-cell precursor acute Target NCI/FDA lymphoblastic leukemia * Bortezomib (Velcade) Mantle cell lymphoma Target NCI/FDA * Bortezomib (Velcade) Multiple myeloma Target NCI/FDA (Continued on following page) JCO Clinical Cancer Informatics 81 Kancherla et al TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source ABL/BCR Bosutinib (Bosulif) Chronic myeloid Sensitive Mutation Fusion NCI/FDA leukemia * Brentuximab vedotin (Adcetris) Classic Hodgkin Target NCI/FDA lymphoma * Brentuximab vedotin (Adcetris) Systemic anaplastic Target NCI/FDA large-cell lymphoma * Brentuximab vedotin (Adcetris) Primary cutaneous Target NCI/FDA anaplastic large-cell lymphoma CD30 Brentuximab vedotin (Adcetris) Mycosis fungoides Sensitive Gene or protein Overexpression NCI/FDA expression ALK Brigatinib (Alunbrig) NSCLC Sensitive Mutation Rearrangement NCI/FDA * Cabazitaxel (Jevtana) Prostate cancer Target NCI/FDA * Cabozantinib (Cabometyx) Renal cell carcinoma Target NCI/FDA * Cabozantinib (Cometriq) Medullary thyroid cancer Target NCI/FDA * Carfilzomib (Kyprolis) Multiple myeloma Target NCI/FDA ALK Ceritinib (LDK378/Zykadia) NSCLC Sensitive Mutation Rearrangement NCI/FDA KRAS Cetuximab (Erbitux) Colorectal cancer Sensitive Mutation WT NCI/FDA * Cetuximab (Erbitux) Head and neck Target NCI/FDA squamous cell carcinoma BRAF Cobimetinib (Cotellic) Melanoma Sensitive Mutation V600E NCI/FDA BRAF Cobimetinib (Cotellic) Melanoma Sensitive Mutation V600K NCI/FDA * Copanlisib hydrochloride (Aliqopa) Follicular Target NCI/FDA lymphoma ALK Crizotinib NSCLC Sensitive Mutation Rearrangement NCI/FDA (Xalkori) ROS1 Crizotinib NSCLC Sensitive Mutation Rearrangement NCI/FDA (Xalkori) BRAF Dabrafenib (Tafinlar) NSCLC Sensitive Mutation V600E NCI/FDA BRAF Dabrafenib (Tafinlar) Melanoma Sensitive Mutation V600E NCI/FDA BRAF Dabrafenib (Tafinlar) Melanoma Sensitive Mutation V600K NCI/FDA BRAF Dabrafenib (Tafinlar) Anaplastic thyroid cancer Sensitive Mutation V600E NCI/FDA * Daratumumab (Darzalex) Multiple myeloma Target NCI/FDA ABL/BCR Dasatinib Chronic myeloid Sensitive Mutation Fusion NCI/FDA (Sprycel) leukemia ABL/BCR Dasatinib Acute lymphoblastic Sensitive Mutation Fusion NCI/FDA (Sprycel) leukemia * Denileukin Cutaneous T-cell Target NCI/FDA diftitox (Ontak) lymphoma * Denosumab Giant-cell tumor of the Target NCI/FDA (Xgeva) bone * Dinutuximab (Unituxin) Neuroblastoma Target NCI/FDA * Durvalumab (Imfinzi) Urothelial carcinoma Target NCI/FDA * Durvalumab (Imfinzi) NSCLC Target NCI/FDA * Elotuzumab (Empliciti) Multiple myeloma Target NCI/FDA IDH2 Enasidenib AML Sensitive Mutation R140Q NCI/FDA mesylate (Idhifa) IDH2 Enasidenib AML Sensitive Mutation R140L NCI/FDA mesylate (Idhifa) (Continued on following page) 82 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source IDH2 Enasidenib AML Sensitive Mutation R140G NCI/FDA mesylate (Idhifa) IDH2 Enasidenib mesylate (Idhifa) AML Sensitive Mutation R140W NCI/FDA IDH2 Enasidenib mesylate (Idhifa) AML Sensitive Mutation R172K NCI/FDA IDH2 Enasidenib AML Sensitive Mutation R172M NCI/FDA mesylate (Idhifa) IDH2 Enasidenib AML Sensitive Mutation R172G NCI/FDA mesylate (Idhifa) IDH2 Enasidenib AML Sensitive Mutation R172S NCI/FDA mesylate (Idhifa) IDH2 Enasidenib AML Sensitive Mutation R172W NCI/FDA mesylate (Idhifa) * Enzalutamide (Xtandi) Prostate cancer Target NCI/FDA EGFR Erlotinib NSCLC Sensitive Mutation Exon 19 deletion NCI/FDA (Tarceva) EGFR Erlotinib NSCLC Sensitive Mutation L858R NCI/FDA (Tarceva) * Erlotinib Pancreatic Target NCI/FDA (Tarceva) cancer ER Everolimus Breast cancer Sensitive Gene or protein Overexpression NCI/FDA (Afinitor) expression PR Everolimus Breast cancer Sensitive Gene or protein Overexpression NCI/FDA (Afinitor) expression * Everolimus Renal cell carcinoma Target NCI/FDA (Afinitor) * Everolimus GI neuroendocrine tumor Target NCI/FDA (Afinitor) * Everolimus Pulmonary Target NCI/FDA (Afinitor) neuroendocrine tumor * Everolimus Pancreatic Target NCI/FDA (Afinitor) neuroendocrine tumor * Everolimus Tuberous sclerosis Target NCI/FDA (Afinitor) complex–associated renal angiomyolipoma * Everolimus Tuberous sclerosis Target NCI/FDA (Afinitor) complex–associated subependymal giant cell astrocytoma ER Exemestane (Aromasin) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression ER Fulvestrant (Faslodex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Fulvestrant (Faslodex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression EGFR Gefitinib (Iressa) NSCLC Sensitive Mutation Exon 19 deletion NCI/FDA EGFR Gefitinib (Iressa) NSCLC Sensitive Mutation L858R NCI/FDA CD33 Gemtuzumab ozogamicin (Mylotarg) AML Sensitive Gene or protein Overexpression NCI/FDA expression * Ibritumomab tiuxetan (Zevalin) Follicular B-cell Target NCI/FDA non-Hodgkin lymphoma * Ibritumomab tiuxetan (Zevalin) Follicular non-Hodgkin Target NCI/FDA lymphoma (Continued on following page) JCO Clinical Cancer Informatics 83 Kancherla et al TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source * Ibrutinib Mantle cell lymphoma Target NCI/FDA (Imbruvica) * Ibrutinib Marginal zone lymphoma Target NCI/FDA (Imbruvica) * Ibrutinib Chronic lymphocytic Target NCI/FDA (Imbruvica) leukemia * Ibrutinib Waldenstrom Target NCI/FDA (Imbruvica) macroglobulinemia * Idelalisib (Zydelig) Chronic lymphocytic Target NCI/FDA leukemia * Idelalisib (Zydelig) Follicular B-cell Target NCI/FDA non-Hodgkin lymphoma * Idelalisib (Zydelig) Small lymphocytic Target NCI/FDA lymphoma * Imatinib mesylate (Gleevec) Dermatofibrosarcoma Target NCI/FDA protuberans c-Kit Imatinib mesylate (Gleevec) GI stromal tumor Sensitive Gene or protein Overexpression NCI/FDA expression ABL/BCR Imatinib mesylate (Gleevec) Chronic myeloid Sensitive Mutation Fusion NCI/FDA leukemia ABL/BCR Imatinib mesylate (Gleevec) Acute lymphoblastic Sensitive Mutation Fusion NCI/FDA leukemia PDGFRB Imatinib mesylate (Gleevec) Myelodysplastic/ Sensitive Mutation Rearrangement NCI/FDA myeloproliferative disorders c-Kit Imatinib mesylate (Gleevec) Systemic mastocytosis Resistant Mutation D816V NCI/FDA c-Kit Imatinib mesylate (Gleevec) Systemic mastocytosis Sensitive Mutation Unknown status NCI/FDA * Inotuzumab ozogamicin (Besponsa) B-cell precursor acute Target NCI/FDA lymphoblastic leukemia * Ipilimumab (Yervoy) Renal cell carcinoma Target NCI/FDA * Ipilimumab (Yervoy) Melanoma Target NCI/FDA MSI-H Ipilimumab (Yervoy) Colorectal cancer Sensitive MSI-H Positive NCI/FDA dMMR Ipilimumab (Yervoy) Colorectal cancer Sensitive dMMR Positive NCI/FDA * Ixazomib citrate (Ninlaro) Multiple myeloma Target NCI/FDA * Lanreotide acetate (Somatuline Depot) Gastroenteropancreatic Target NCI/FDA neuroendocrine tumor HER2 Lapatinib (Tykerb) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Lenvatinib mesylate (Lenvima) Renal cell carcinoma Target NCI/FDA * Lenvatinib mesylate (Lenvima) Differentiated thyroid Target NCI/FDA cancer ER Letrozole (Femara) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Letrozole (Femara) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Lutetium Lu-dotatate (Lutathera) Gastroenteropancreatic Target NCI/FDA neuroendocrine tumor FLT3 Midostaurin (Rydapt) AML Sensitive Mutation Internal tandem NCI/FDA duplication FLT3 Midostaurin (Rydapt) AML Sensitive Mutation D835X NCI/FDA (Continued on following page) 84 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source FLT3 Midostaurin (Rydapt) AML Sensitive Mutation I836X NCI/FDA * Midostaurin (Rydapt) Mast cell leukemia Target NCI/FDA * Midostaurin (Rydapt) Systemic mastocytosis Target NCI/FDA * Necitumumab (Portrazza) NSCLC Target NCI/FDA HER2 Neratinib maleate (Nerlynx) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression ABL/BCR Nilotinib (Tasigna) Chronic myeloid Sensitive Mutation Fusion NCI/FDA leukemia * Niraparib tosylate monohydrate (Zejula) Fallopian tube cancer Target NCI/FDA * Niraparib tosylate monohydrate (Zejula) Ovarian cancer Target NCI/FDA * Niraparib tosylate monohydrate (Zejula) Primary peritoneal Target NCI/FDA cancer MSI-H Nivolumab (Opdivo) Colorectal cancer Sensitive MSI-H Positive NCI/FDA dMMR Nivolumab (Opdivo) Colorectal cancer Sensitive dMMR Positive NCI/FDA * Nivolumab (Opdivo) Head and neck Target NCI/FDA squamous cell carcinoma * Nivolumab (Opdivo) Renal cell carcinoma Target NCI/FDA * Nivolumab (Opdivo) Hepatocellular Target NCI/FDA carcinoma * Nivolumab (Opdivo) NSCLC Target NCI/FDA * Nivolumab (Opdivo) Classic Hodgkin Target NCI/FDA lymphoma * Nivolumab (Opdivo) Melanoma Target NCI/FDA * Nivolumab (Opdivo) Urothelial carcinoma Target NCI/FDA * Obinutuzumab (Gazyva) Chronic lymphocytic Target NCI/FDA leukemia * Obinutuzumab (Gazyva) Follicular lymphoma Target NCI/FDA * Ofatumumab (Arzerra) Chronic lymphocytic Target NCI/FDA leukemia BRCA1 Olaparib (Lynparza) Breast cancer Sensitive Mutation Deleterious NCI/FDA BRCA2 Olaparib (Lynparza) Breast cancer Sensitive g.mutation Deleterious NCI/FDA * Olaparib (Lynparza) Fallopian tube cancer Target NCI/FDA BRCA1 Olaparib (Lynparza) Ovarian cancer Sensitive g.mutation Deleterious NCI/FDA BRCA2 Olaparib (Lynparza) Ovarian cancer Sensitive g.mutation Deleterious NCI/FDA * Olaparib (Lynparza) Ovarian cancer Target NCI/FDA * Olaparib (Lynparza) Primary peritoneal Target NCI/FDA cancer * Olaratumab (Lartruvo) Soft tissue sarcoma Target NCI/FDA EGFR Osimertinib (Tagrisso) NSCLC Sensitive Mutation Exon 19 deletion NCI/FDA EGFR Osimertinib (Tagrisso) NSCLC Sensitive Mutation L858R NCI/FDA EGFR Osimertinib (Tagrisso) NSCLC Sensitive Mutation T790M NCI/FDA ER Palbociclib (Ibrance) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Palbociclib (Ibrance) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression KRAS Panitumumab (Vectibix) Colorectal cancer Sensitive Mutation WT NCI/FDA (Continued on following page) JCO Clinical Cancer Informatics 85 Kancherla et al TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source NRAS Panitumumab (Vectibix) Colorectal cancer Sensitive Mutation WT NCI/FDA * Panobinostat (Farydak) Multiple myeloma Target NCI/FDA * Pazopanib (Votrient) Renal cell carcinoma Target NCI/FDA * Pazopanib (Votrient) Soft tissue sarcoma Target NCI/FDA PD-L1 Pembrolizumab (Keytruda) Urothelial carcinoma Sensitive Gene or protein Overexpression NCI/FDA expression PD-L1 Pembrolizumab (Keytruda) Cervical cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Pembrolizumab (Keytruda) Head and neck Target NCI/FDA squamous cell carcinoma * Pembrolizumab (Keytruda) NSCLC Target NCI/FDA PD-L1 Pembrolizumab (Keytruda) NSCLC Sensitive Gene or protein Overexpression NCI/FDA expression * Pembrolizumab (Keytruda) Classic Hodgkin Target NCI/FDA lymphoma * Pembrolizumab (Keytruda) Primary mediastinal large Target NCI/FDA B-cell lymphoma MSI-H Pembrolizumab (Keytruda) Colorectal cancer Sensitive MSI-H Positive NCI/FDA dMMR pembrolizumab (Keytruda) Colorectal cancer Sensitive dMMR Positive NCI/FDA MSI-H Pembrolizumab (Keytruda) Solid tumors Sensitive MSI-H Positive NCI/FDA dMMR Pembrolizumab (Keytruda) Solid tumors Sensitive dMMR Positive NCI/FDA * Pembrolizumab (Keytruda) Melanoma Target NCI/FDA PD-L1 Pembrolizumab (Keytruda) Gastric cancer Sensitive Gene or protein Overexpression NCI/FDA expression HER2 Pertuzumab (Perjeta) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Ponatinib hydrochloride (Iclusig) Chronic myeloid Target NCI/FDA leukemia ABL/BCR Ponatinib hydrochloride (Iclusig) Chronic myeloid Sensitive Mutation T315I NCI/FDA leukemia ABL/BCR Ponatinib hydrochloride (Iclusig) AML Sensitive Mutation Fusion NCI/FDA ABL/BCR Ponatinib hydrochloride (Iclusig) AML Sensitive Mutation T315I NCI/FDA * Pralatrexate (Folotyn) Peripheral T-cell Target NCI/FDA lymphoma * Radium 223 dichloride (Xofigo) Prostate cancer Target NCI/FDA * Ramucirumab (Cyramza) Colorectal cancer Target NCI/FDA * Ramucirumab (Cyramza) Gastric cancer Target NCI/FDA * Ramucirumab (Cyramza) NSCLC Target NCI/FDA * Regorafenib (Stivarga) Colorectal cancer Target NCI/FDA * Regorafenib (Stivarga) GI stromal tumor Target NCI/FDA * Regorafenib (Stivarga) Hepatocellular Target NCI/FDA carcinoma ER Ribociclib (Kisqali) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Ribociclib (Kisqali) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression CD20 Rituximab (Rituxan) Chronic lymphocytic Sensitive Gene or protein Overexpression NCI/FDA leukemia expression (Continued on following page) 86 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source CD20 Rituximab (Rituxan) Non-Hodgkin lymphoma Sensitive Gene or protein Overexpression NCI/FDA expression * Rituximab and hyaluronidase human Chronic lymphocytic Target NCI/FDA (Rituxan Hycela) leukemia * Rituximab and hyaluronidase human DLBCL Target NCI/FDA (Rituxan Hycela) * Rituximab and hyaluronidase human Follicular lymphoma Target NCI/FDA (Rituxan Hycela) * Romidepsin (Istodax) Cutaneous T-cell Target NCI/FDA lymphoma * Romidepsin (Istodax) Peripheral T-cell Target NCI/FDA lymphoma * Rucaparib camsylate (Rubraca) Fallopian tube cancer Target NCI/FDA BRCA1 Rucaparib camsylate (Rubraca) Fallopian tube cancer Sensitive Mutation Deleterious NCI/FDA BRCA2 Rucaparib camsylate (Rubraca) Fallopian tube cancer Sensitive Mutation Deleterious NCI/FDA * Rucaparib camsylate (Rubraca) Ovarian cancer Target NCI/FDA BRCA1 Rucaparib camsylate (Rubraca) Ovarian cancer Sensitive Mutation Deleterious NCI/FDA BRCA2 Rucaparib camsylate (Rubraca) Ovarian cancer Sensitive Mutation Deleterious NCI/FDA * Rucaparib camsylate (Rubraca) Primary peritoneal Target NCI/FDA cancer BRCA1 Rucaparib camsylate (Rubraca) Primary peritoneal Sensitive Mutation Deleterious NCI/FDA cancer BRCA2 Rucaparib camsylate (Rubraca) Primary peritoneal Sensitive Mutation Deleterious NCI/FDA cancer * Ruxolitinib phosphate (Jakafi) Myelofibrosis Target NCI/FDA * Ruxolitinib phosphate (Jakafi) Polycythemia vera Target NCI/FDA * Siltuximab (Sylvant) Multicentric Castleman Target NCI/FDA disease * Sonidegib (Odomzo) Basal cell carcinoma Target NCI/FDA * Sorafenib (Nexavar) Renal cell carcinoma Target NCI/FDA * Sorafenib (Nexavar) Hepatocellular Target NCI/FDA carcinoma * Sorafenib (Nexavar) Thyroid carcinoma Target NCI/FDA c-Kit Sunitinib (Sutent) GI stromal tumor Sensitive Gene or protein Overexpression NCI/FDA expression * Sunitinib (Sutent) Renal cell carcinoma Target NCI/FDA * Sunitinib (Sutent) Pancreatic Target NCI/FDA neuroendocrine tumor ER Tamoxifen (Nolvadex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Tamoxifen (Nolvadex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Temsirolimus (Torisel) Renal cell carcinoma Target NCI/FDA * Tisagenlecleucel (Kymriah) B-cell acute Target NCI/FDA lymphoblastic leukemia * Tisagenlecleucel (Kymriah) DLBCL Target NCI/FDA ER Toremifene (Fareston) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression BRAF Trametinib (Mekinist) NSCLC Sensitive Mutation V600E NCI/FDA BRAF Trametinib (Mekinist) Melanoma Sensitive Mutation V600E NCI/FDA (Continued on following page) JCO Clinical Cancer Informatics 87 Kancherla et al TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source BRAF Trametinib (Mekinist) Melanoma Sensitive Mutation V600K NCI/FDA BRAF Trametinib (Mekinist) Anaplastic thyroid cancer Sensitive Mutation V600E NCI/FDA HER2 Trastuzumab (Herceptin) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression HER2 Trastuzumab (Herceptin) Gastric cancer Sensitive Gene or protein Overexpression NCI/FDA expression RARA Tretinoin (Vesanoid) Acute promyelocytic Sensitive Mutation t(15;19) NCI/FDA leukemia * Vandetanib (Caprelsa) Medullary thyroid cancer Target NCI/FDA BRAF Vemurafenib (Zelboraf) Erdheim-Chester disease Sensitive Mutation V600X NCI/FDA BRAF Vemurafenib (Zelboraf) Melanoma Sensitive Mutation V600E NCI/FDA * Venetoclax (Venclexta) Chronic lymphocytic Target NCI/FDA leukemia * Venetoclax (Venclexta) Small lymphocytic Target NCI/FDA lymphoma * Vismodegib (Erivedge) Basal cell carcinoma Target NCI/FDA * Vorinostat (Zolinza) Cutaneous T-cell Target NCI/FDA lymphoma * Ziv-aflibercept (Zaltrap) Colorectal cancer Target NCI/FDA NOTE. Last updated on July 12, 2018. Asterisks denote no biomarker indication. Abbreviations: AML, acute myeloid leukemia; DLBCL, diffuse large B-cell lymphoma; dMMR, deficient mismatch repair; EGFR, epidermal growth factor receptor; ER, estrogen receptor; FDA, US Food and Drug Administration; HER2, human epidermal growth factor receptor 2; MSI-H, microsatellite instability high; NCI, National Cancer Institute; NSCLC, non–small-cell lung cancer; PD-L1, programmed death 1 ligand; PR, progesterone receptor; WT, wild type. 88 © 2020 by American Society of Clinical Oncology http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png JCO Clinical Cancer Informatics Wolters Kluwer Health

Loading next page...
 
/lp/wolters-kluwer-health/evidence-based-network-approach-to-recommending-targeted-cancer-EcZGgf5wag

References (124)

Publisher
Wolters Kluwer Health
Copyright
(C) 2020 American Society of Clinical Oncology
ISSN
2473-4276
DOI
10.1200/CCI.19.00097
Publisher site
See Article on Publisher Site

Abstract

original reports abstract SPECIAL SERIES: INFORMATICS TOOLS FOR CANCER RESEARCH AND CARE Evidence-Based Network Approach to Recommending Targeted Cancer Therapies 1 2 2 2 2 Jayaram Kancherla, MS ; Shruti Rao, MS, MBA ; Krithika Bhuvaneshwar, MS ; Rebecca B. Riggins, PhD ; Robert A. Beckman, MD ; 2 1 1 Subha Madhavan, PhD ;Hector ´ Corrada Bravo, PhD ; and Simina M. Boca, PhD PURPOSE In this work, we introduce CDGnet (Cancer-Drug-Gene Network), an evidence-based network ap- proach for recommending targeted cancer therapies. CDGnet represents a user-friendly informatics tool that expands the range of targeted therapy options for patients with cancer who undergo molecular profiling by including the biologic context via pathway information. METHODS CDGnet considers biologic pathway information specifically by looking at targets or biomarkers downstream of oncogenes and is personalized for individual patients via user-inputted molecular alterations and cancer type. It integrates a number of different sources of knowledge: patient-specific inputs (molecular al- terations and cancer type), US Food and Drug Administration–approved therapies and biomarkers (curated from DailyMed), pathways for specific cancer types (from Kyoto Encyclopedia of Genes and Genomes [KEGG]), gene-drug connections (from DrugBank), and oncogene information (from KEGG). We consider 4 different evidence-based categories for therapy recommendations. Our tool is delivered via an R/Shiny Web application. For the 2 categories that use pathway information, we include an interactive Sankey visualization built on top of d3.js that also provides links to PubChem. RESULTS We present a scenario for a patient who has estrogen receptor (ER)–positive breast cancer with FGFR1 amplification. Although many therapies exist for patients with ER-positive breast cancer, FGFR1 amplifications may confer resistance to such treatments. CDGnet provides therapy recommendations, including PIK3CA, MAPK, and RAF inhibitors, by considering targets or biomarkers downstream of FGFR1. CONCLUSION CDGnet provides results in a number of easily accessible and usable forms, separating targeted cancer therapies into categories in an evidence-based manner that incorporates biologic pathway information. JCO Clin Cancer Inform 4:71-88. © 2020 by American Society of Clinical Oncology Licensed under the Creative Commons Attribution 4.0 License INTRODUCTION some tumor types, it is now routine to check for specific molecular features to decide on targeted In today’s era of cancer precision medicine, thera- treatment plans. For example, KRAS wild-type co- peutic interventions are often tailored to an individual’s lorectal cancer is generally treated with epidermal tumor molecular profile, in addition to traditional growth factor receptor (EGFR) inhibitors, estrogen considerations, including age, sex, cancer stage, and ASSOCIATED receptor (ER)–positive breast cancer with aromatase CONTENT medical and treatment histories. The term molecular inhibitors or antiestrogens such as tamoxifen or Appendix profiling is often used to refer to a test that con- fulvestrant, and human epidermal growth factor Author affiliations siders ≥1 biomarker. These biomarkers may be either and support receptor 2–positive breast cancer with monoclonal genetic characteristics or mRNA or protein expression information (if antibodies trastuzumab and pertuzumab, tyrosine values. Genetic characteristics include point muta- applicable) appear at kinase inhibitors such as neratinib, or antibody-toxin tions, insertions, deletions, duplications, gene fusions, the end of this conjugates such as trastuzumab-DM1. In many and rearrangements. They may be either germ line article. cases, if there is no US Food and Drug Adminis- (inherited and present in normal tissue) or somatic Accepted on tration (FDA)–approved targeted therapy for a spe- December 4, 2019 (present in cancer cells but not normal tissue). and published at cific tumor type, clinicians may recommend either Expression values refer to the expression of mRNA ascopubs.org/journal/ an off-label therapy that is prescribed for the pa- or protein in tumors, either in comparison with other cci on January 28, tient’s alteration in another tumor type or enrollment tumors or adjacent normal tissue. Typically, tumor 2020: DOI https://doi. molecular profilingisusedwhenapatienthas few in a precision medicine clinical trial (eg, basket, org/10.1200/CCI.19. 00097 or no standard treatment options left. However, for umbrella, or targeted therapy trial). 71 Kancherla et al CONTEXT Key Objective With the increasing use of tumor molecular profiling, it is imperative to develop approaches that consider the biologic context to better prioritize targeted therapies for patients with cancer. Knowledge Generated We introduce CDGnet (Cancer-Drug-Gene Network), a tool that integrates biologic pathway information with alterations detected from tumor molecular profiling to expand the possible range of targeted therapies and prioritize them into evidence-based categories. We demonstrate how CDGnet can be applied to a scenario where a patient with breast cancer has overexpression of both ESR1 and FGFR1, with the output representing the recommended therapies, the clinical context in which they are approved, and the links between the patient’s tumor molecular profile and the recommendations. Relevance We consider this tool to be especially valuable to clinical and translational researchers who may be interested in understanding the best course of treatment for patients with a particular tumor molecular profile. To make such decisions about off-label therapy recom- a result, patients with colorectal cancer are typically tested mendations, clinicians have to sift through vast amounts of for KRAS mutations, and EGFR inhibitors are only pre- literature and clinical databases to determine the clinical scribed to individuals without specific KRAS mutations in utility of variants identified through molecular profiling to codons 12 and 13. A comprehensive characterization of decide on the appropriate treatment option for their pa- untreated colorectal tumors estimated that 43% of non- tients. The same is true for clinical translational scientists hypermutated tumors had KRAS mutations, and these considering relevant therapeutic approaches to evaluate in mutations were generally oncogenic activating mutations, model systems or humans, using either single agents or which means that a large percentage of patients with co- combinations. In this setting, the number of possible lorectal cancer are left with few therapeutic options. Our molecular profiles that may be relevant and the number of framework and tool are seeking to remedy this issue. experimental agents create a combinatorial explosion of METHODS research possibilities among which prioritization is needed. Overview of Methods for Generating Several efforts are ongoing to capture, standardize, and Patient-Specific Networks share clinically relevant variants identified through mo- lecular diagnostic tests among several public, academic, The user inputs into CDGnet are the specific alterations 3-5 and private institutions, although challenges remain in found in a patient’s tumor and the patient’s cancer type. synthesizing evidence in a manner that is both systematic Part of the landing page is shown in Figure 2. These data 6,7 and timely. Our goal in this work is to expand the range are then integrated with biologic networks relevant to the of options for targeted therapies for patients with cancer cancer type (from the Kyoto Encyclopedia of Genes and who undergo molecular profiling by developing CDGnet Genomes [KEGG] database ), FDA-approved targeted (Cancer-Drug-Gene Network), a user-friendly, evidence- cancer therapies and indications (curated from DailyMed based approach that accounts for downstream effects therapy labels ), additional gene-drug connections in within pathways in cancer and is personalized for the in- the form of drug targets (from the DrugBank database ), dividual patient. Our tool, which uses the Shiny framework information on whether a gene is an oncogene (from KEGG). 8 9 with an R backend, is available online. We incorporate Users may consider different data sources by using the pathway information specifically by looking at downstream CDGnet code directly, for example, by considering the targets of oncogenes, which are genes that are constitu- predicted oncogenes from a recent comprehensive charac- 10 16 tively activated in cancer. This is illustrated in Figure 1.If terization of The Cancer Genome Atlas (TCGA) projects. an oncogene in a biologic pathway is activated, targeting Currently, the biologic networks we consider are the genes and proteins that are found upstream may no cancer-specific pathways in KEGG, and therefore, for now, longer be effective, leading to a focus on downstream we are also restricting the cancer types to those that have targets. This includes the scenario of EGFR inhibitors for KEGG pathways. We have developed 4 different therapy KRAS wild-type colorectal tumors. The EGFR protein categories that can be prioritized for patients, given their triggers a signaling cascade in cancer, which may be specific tumor alterations, ordered from “most evidence blocked by anti-EGFR drugs; however, this is only ef- that therapy works” to “least evidence that therapy works.” fective if KRAS, which is downstream of EGFR, is not (1) FDA-approved drugs for which the input genes/proteins mutated. Otherwise, certain KRAS mutations can lead to are biomarkers for their tumor type; (2) FDA-approved a lack of effectiveness of therapies that block EGFR. As drugs for which the input genes/proteins are biomarkers 72 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies Tumor with no oncogenic Tumor with oncogenic activation of protein B activation of protein B i. No therapy ii. Targeted therapy iii. No therapy iv. Targeted therapy Drug Drug A A A A against against protein A protein A B B C C C D D D E E E Abnormal Abnormal Abnormal cancer-related cancer-related cancer-related processes processes processes FIG 1. Simplified diagram showing the reasoning behind looking at downstream targets of activated oncogenes. A simple pathway is shown that consists of 5 proteins, A, B, C, D, and E, with A activating B, B activating C, and so on, with the final activation of E leading to various abnormal cancer-related processes. (i, ii) Scenario where a tumor has no oncogenic activation of protein B. (iii, iv) Scenario where protein B has gained an oncogenic mutation that renders it constitutively active. If there is no oncogenic activation of protein B, then targeting protein A, as in (ii), may be effective in stopping cancer growth. However, if there is oncogenic activation of protein B, this means that, in particular, it is not necessary for protein A to activate protein B, so that targeting protein A is not effective for turning off the pathway. in other tumor types; (3) drugs that have as targets the input We differentiate between targets and biomarkers because genes/proteins or as biomarkers/targets other genes/ in many cases, as a result of complicated biologic in- proteins that are downstream of the input oncogenes when teractions, the target of a therapy may be different from the considering the pathway corresponding to this tumor type; biomarker used to specify the indication, such as in the case and (4) drugs that have as biomarkers/targets other genes/ of EGFR inhibitors being administered for KRAS wild-type proteins that are downstream of input oncogenes when colorectal tumors or CDK4/6 inhibitors being administered for considering the pathways corresponding to other tumor ER-positive breast tumors. The general approach is pre- types. sented in Figure 3. The options used on the landing page to obtain the different therapy categories are shown in Figure 2. In categories 3 and 4, users have the option to consider only We also provide documentation for the tool, including a step- FDA-approved targeted cancer therapies, all FDA-approved by-step analysis for the built-in patient use case scenario. therapies, or all drugs in DrugBank; this allows clinical re- searchers to consider increasing numbers of therapies only Appendix Table A1 lists FDA-approved targeted cancer as needed, as opposed to being overwhelmed with a huge therapies and indications that were obtained by consider- number of therapies from the start. We also note the differ- ing the targeted therapies listed by the National Cancer ence between categories 3 and 4; category 3 considers Institute and looking up the corresponding labels via the biologic pathway corresponding to the individual’s DailyMed. In particular, the indications and usage portion cancer type, whereas category 4 considers the pathways of the label was used to obtain the specificcancertype and corresponding to other cancer types. Given that pathways biomarker information, which is listed in the “Gene/Protein,” represent a simplification of a more complicated reality and “Data Type,” and “Alteration” columns; in the case of each tumor is unique, we found it necessary to allow for multiple biomarkers, these are listed in separate rows of possible connections between genes and proteins that may be the table. In cases where the biomarker indication is curated in cancer types different from that with which a patient unclear, the lists of FDA companion diagnostic tests were 19,20 presents, although in our experience, it is generally sufficient to also consulted. Note that although some targeted stop at category 3 therapies. therapies have specific biomarker indications, many do not. JCO Clinical Cancer Informatics 73 Kancherla et al FIG 2. Part of the landing page, which shows how users can select the cancer type and either input a tab-separated or comma-separated file or use the example data. The inset shows how under “Filter Recommended Therapies,” combinations of the first 2 checkboxes lead to the 4 different categories of therapy recommendations described in the text. Removing 1 or both of the last 2 checkboxes expands the range of therapies in categories 3 and 4 beyond US Food and Drug Administration (FDA)–approved drugs and FDA-approved targeted cancer drugs, respectively. For example, ibrutinib is a targeted therapy, administered for parsed and had identifiers converted using the KEGGREST, 22 23 a number of subtypes of leukemia/lymphoma, but not for KEGGgraph, and org.Hs.eg.db Bioconductor packages, a specific biomarker indication. If there is no biomarker respectively, and against the information input by the user, indication, this is noted as an asterisk in the table in the with the gene/protein names being normalized via the rDGIdb “Gene/Protein” column. The therapies are then cross- package, which is a wrapper for the Drug Gene Interaction referenced with DrugBank to obtain the targets for both 24,25 Database. the therapies with biomarker indications and those without To obtain the list of FDA-approved drugs, we used the data indications. The biomarkers and targets obtained in these files from the official Drugs@FDA resource. Drugs@FDA ways are checked against downstream targets from KEGG cancer-specific pathways, which were downloaded and contains several tab-separated value files that include Category 1: Patient-specific inputs Therapy recommendations FDA-approved drugs for these alterations in this Category 3: Molecular alterations cancer type (ie, the alterations represent Drugs that have as targets the input biomarkers in this genes/proteins or as biomarkers/targets other Cancer type cancer type) genes/proteins that are downstream of the input oncogenes when considering the pathway corresponding to this tumor type Category 2: FDA-approved FDA-approved drugs for Category 4: targeted therapies these alterations in other Drugs that have as biomarkers/targets other and biomarkers cancer types (ie, the genes/proteins that are downstream of input alterations represent oncogenes when considering the pathways Curation from biomarkers in other corresponding to other tumor types DailyMed cancer types) Pathways Gene-drug Oncogene for specific connections information cancer types KEGG DrugBank KEGG FIG 3. General approach for targeted therapy recommendations, including specific data sources. FDA, US Food and Drug Administration; KEGG, Kyoto Encyclopedia of Genes and Genomes. 74 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies information on the submission, review, and approval pro- column represents the pathway between the altered gene/ cess for various drugs. We use the products (list of all drugs) protein and the gene/protein that is a biomarker or target; and submission (review process for all drugs) files to filter for the alteration column represents the biomarker for an FDA- drugs that are approved or tentatively approved. The Drugs@ approved indication, if this exists, in which case the tumor FDA resource contains a list of all drugs approved since 1939, for which it is approved is also listed; the predicted effect some of which may have been discontinued. As a result, we column has the value “sensitive”, if the alteration column is use the marketingstatus file to remove any discontinued not empty, or “target”, if the drug targets the protein products from the list. The R scripts to parse and filter the according to the DrugBank data. Drugs@FDA data files are available in our GitHub repository. An architecture diagram for our system is shown in Shiny App and Visualization Figure 6. We use Shiny, an R package/framework for creating interactive and standalone Web applications di- For each of the 4 categories detailed, a sortable and rectly from R. Shiny applications can run on a Web page or searchable table of therapies is output with the FDA- be embedded in RMarkdown documents to build in- approved indications; for categories 3 and 4, network vi- teractive dashboards. They use the same technology that sualizations are also shown. The table also provides the powers Web applications (ie, HTML and JavaScript) and tumor type in which a particular therapy is approved. allow users to create intuitive and interactive user interfaces Figure 4 shows a Sankey flow diagram representation that and prototypes with an R computational backend. focuses on the flow of evidence between drug-gene and gene-gene connections, enabling an intuitive visualization To support interactive Sankey charts within Shiny appli- from the molecular profile to the inferred targets and cations, we developed a Shiny Web component for vi- recommended therapies. Figure 5 shows a portion of the sualizing Sankey flow diagrams, available to download sortable and searchable corresponding table. The path as an R package. Web components are custom HTML FIG 4. Sankey flow diagram focusing on the flow of evidence between drug-gene and gene connections for a putative patient with estrogen receptor–positive breast cancer and FGFR1 overexpression, showing category 3 recommendations, namely, targets downstream of FGFR1. Therapies can be clicked to obtain a panel with PubChem information. JCO Clinical Cancer Informatics 75 Kancherla et al FIG 5. Part of the sortable, searchable table for therapies in category 3 for a putative patient with estrogen receptor–positive breast cancer and FGFR1 overexpression, showing the subset of therapies that target MAP2K1. elements that are natively extensible and reusable and a pathway connection or information on a drug when can be integrated into any framework that supports HTML. a user selects/clicks on an edge or node. Selecting an The Sankey visualization uses a custom 3-column layout edge shows the downstream pathway information used to organize nodes in the graph: molecular profile and for inference. Selecting a recommended therapy dis- FDA-approved drugs, inferred targets, and recommen- plays the structure of the drug and linked publications ded therapies; it intuitively focuses the user on the flow from PubChem, using PubChem widgets. The Sankey of evidence from input parameters to recommended visualization is built on top of d3.js, a data visuali- therapies. The Sankey visualization also contains an zation library for JavaScript to build highly customizable information panel that displays evidence related to and interactive visualizations. User interface Server User input or loaded example Downloaded data from � Mo lecular alterations � C urated therapy labels � Ca ncer type � KEGG � D rugBank R/Shiny R backend computations R/Shiny Display of therapy Therapy recommendations recommendations Tables nfpmShinyComponent R package Sankey diagrams FIG 6. Architecture diagram for our system. KEGG, Kyoto Encyclopedia of Genes and Genomes. 76 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies RESULTS can be targeted by different cancer therapies. On the basis of the level of evidence, the clinical actionability of these We will now consider the scenario of a patient who has ER- pathways can be further tested in a laboratory or clinical trial positive breast cancer. ER-positive breast cancer, generally setting. Additionally, there is a growing field of research related treated with aromatase inhibitors or antiestrogens, employs to drug-target interactions and drug repositioning using an array of mechanisms that permit escape from these 37-40 network-based models, which may in the future be in- therapies. These include amplification or upregulation tegrated with our tool. of fibroblast growth factor receptor 1 (FGFR1), which is amplified in approximately 13% of ER-positive tumors from We aim to further enhance the data that drive the CDGnet 30-32 33 TCGA and leads to ligand-independent ER activation. tool by incorporating relevant information from additional FGFR activity has also recently been shown to confer re- precision oncology efforts, tools, and resources. Users who sistance to CDK4/6 inhibitors in ER-positive breast download or connect to these resources may currently use cancer. Pan-FGFR antagonists have been combined them in the context of our approach by modifying our with endocrine therapies in prior clinical studies (eg, code. Expert-curated precision oncology databases CTKI258A2210), but the efficacy of this combination has include Clinical Interpretations of Variants in Cancer 5,41 42,43 44,45 been minimal, even in patients preselected for alterations in (CIViC), Cancer Genome Interpreter, OncoKB, 46,47 the FGFR pathway. A potential underlying explanation for Database of Evidence for Precision Oncology (DEPO), 48,49 this lack of benefit is that FGFR alterations impinge upon and Precision Medicine Knowledge Base (PMKB), and downstream signaling networks shared by many other more general resources include ClinVar. These additional receptor tyrosine kinases. Figure 4 shows CDGnet rec- sources may further strengthen the clinical annotations ommendations for a patient with breast cancer with over- and evidence related to germ line and somatic alterations in expression of both ESR1 (gene encoding ER) and FGFR1, our database and provide options between curated drug when considering only FDA-approved targeted therapies. label information and DrugBank targets. CIViC is an open- Therapy recommendations include PIK3CA, MAPK, and access, open-source, community-driven Web resource that RAF inhibitors, which may have utility in this context, along allows clinical interpretations of mutations related to can- with the standard targeted therapies prescribed for ER- cer. Cancer Genome Interpreter is an online tool that positive breast cancer. Figure 5 shows the subset of the connects genes and drugs along with their effects and corresponding table that consists of FDA-approved publication sources, not in a network format, but in a tab- MAP2K1 inhibitors, which are approved for either ABL1 ular format. OncoKB is another online precision oncology fusions in chronic myeloid leukemia, or specific BRAF knowledge base that contains information about the effects mutations in melanoma, non–small-cell lung cancer, and and treatment implications of specific cancer gene alter- anaplastic thyroid cancer. ations. DEPO contains druggable variant information such as drug therapies, evidence levels (FDA approved, clinical DISCUSSION trials, case reports, and preclinical), and cancer types for We developed the CDGnet tool using an approach that intended treatments. PMKB provides information about considers biologic pathways and connections among clinical cancer variants and interpretations. We are also genes, proteins, and drugs to prioritize targeted therapies using a simplified model for incorporating pathway in- for patients with cancer. Our approach integrates many formation via the consideration of targets that are down- disparate sources of knowledge and provides results in an stream of oncogenes; there are scenarios we do not capture easily accessible and usable format. With our tool, users are where upstream targeting can also be useful, for example, able to quickly obtain information on the FDA-approved 51-53 in the case of a feedback loop. We will incorporate more therapies (category 1) and potential off-label therapies complex information in future iterations of our tool. Our tool (category 2) associated with a patient’s molecular profile. partly relies on manual curation of information for FDA- Our definitions of categories 1 and 2 in CDGnet are in approved targeted therapies and thus has challenges alignment with the tier I and II evidence level classifications similar to those of other tools in this space, including the recommended by the Association for Molecular Pathology, time- and labor-intensive nature of this process. However, American College of Medical Genetics and Genomics, the KEGG and DrugBank components only need to be ASCO, and College of American Pathologists. However, downloaded and reprocessed through our existing code CDGnet categories 3 and 4 are unique to our evidence- when updates are desired. based network approach and enable users to evaluate additional targeted therapy options based on an individual’s Consortia such as the Clinical Genome Resource Somatic 3,54 tumor profile. It is important to note that the targeted therapy Cancer Working Group and the Global Alliance for recommendations in categories 3 and 4 have lower evidence Genomics and Health Variant Interpretation for Cancer levels and may or may not have proven clinical significance Consortium have ongoing efforts to standardize and in ongoing clinical trials. However, by examining the down- harmonize the expert-curated data in these different stream targets of candidate biomarkers, clinical researchers knowledge bases, with the goal of enhancing the in- can derive key insights into potential biologic pathways that teroperability among these databases. We will align the JCO Clinical Cancer Informatics 77 Kancherla et al future development of CDGnet with the guidelines and a personalized tool may eventually expand the range of consensus frameworks developed by these consortia. options of targeted therapies for patients with cancer in the CDGnet can also serve as an informative tool for oncolo- clinical setting, a key goal of precision oncology. gists, molecular pathologists, and genomic scientists who We currently consider clinical or translational researchers routinely participate in molecular tumor board discussions. to be the primary target user group for our tool. For in- 56,57 Tools similar to CDGnet include PreMedKB and the stance, if they are interested in a particular combination of Drug Gene Interaction Network. PreMedKB is an in- molecular alterations for a specific cancer type and gen- tegrated precision medicine knowledgebase for interpret- erally find the recommendations to be for drugs prescribed ing relationships among diseases, genes, variants, and in a different cancer type, they may decide to pursue formal drugs. The Drug Gene Interaction Network is a commercial studies of drug repurposing, which is made easier by tool offered by Seqome (Tramore, Ireland) that builds drug- knowing whether they are considering an FDA-approved gene interaction networks to predict clinical response from targeted drug, FDA-approved drug, or non–FDA-approved multiomics data sets. The advantage of CDGnet over these drug. Our eventual goal is to allow for the use of this tool by tools is that our approach allows users to input specific clinicians, especially for the care of patients with advanced- alterations found in a patient’s tumor and cancer type and stage disease for whom the immediate FDA-approved outputs therapy options ordered based on priority. Such therapy choices have been exhausted. AFFILIATIONS AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF University of Maryland, College Park, MD INTEREST Georgetown University, Washington, DC The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless Preprint version available on bioRxiv. otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the CORRESPONDING AUTHOR subject matter of this manuscript. For more information about ASCO’s Simina M. Boca, PhD, Georgetown University Medical Center, 2115 conflict of interest policy, please refer to www.asco.org/rwc or ascopubs. Wisconsin Ave, Suite 110, Washington, DC 20007; e-mail: smb310@ org/cci/author-center. georgetown.edu. Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments). EQUAL CONTRIBUTION J.K. and S.R. contributed equally to this work. Shruti Rao Research Funding: Symphony (Inst) SUPPORT Robert A. Beckman Supported by National Institutes of Health grant No. R21CA220398 and Leadership: Onco-Mind supplements No. R21CA220398-02 (S.M.B.), U01HG007437 (S.M.), Stock and Other Ownership Interests: Johnson & Johnson U24CA237719 (S.M.), and P30CA051008 (S.M.B., via pilot award); Consulting or Advisory Role: AstraZeneca, Zymeworks, Vertex, Department of Defense Breast Cancer Research Program award No. EMDSerono, CStone. W81XWH-17-1-0615 (R.B.R.); National Institute of General Medical Patents, Royalties, Other Intellectual Property: Two patents for dynamic Sciences grant No. R01GM114267 (H.C.B.); and National Institute of precision medicine, a novel approach to precision medicine, have been Mental Health grant No. R24MH114815 (H.C.B.). granted in Japan and Taiwan and transferred to Onco-Mind; these patents are pending in the United States and European Union. Subha Madhavan AUTHOR CONTRIBUTIONS Leadership: Perthera Conception and design: Robert A. Beckman, Subha Madhavan, Hector Stock and Other Ownership Interests: Perthera Corrada Bravo, Simina M. Boca Consulting or Advisory Role: Perthera Financial support: Rebecca B. Riggins, Subha Madhavan, Hector Corrado Research Funding: Teewinot Life Sciences (Inst) Bravo, Simina M. Boca Administrative support: Simina M. Boca Hector Corrada Bravo Collection and assembly of data: Jayaram Kancherla, Shruti Rao, Simina Consulting or Advisory Role: Genentech/Roche M. Boca Travel, Accommodations, Expenses: Genentech/Roche Data analysis and interpretation: Krithika Bhuvaneshwar, Rebecca B. Simina M. Boca Riggins, Robert A. Beckman, Subha Madhavan, Simina M. Boca Research Funding: Symphogen (Inst) Manuscript writing: All authors Final approval of manuscript: All authors Accountable for all aspects of the work: All authors No other potential conflicts of interest were reported. 78 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies REFERENCES 1. Amado RG, Wolf M, Peeters M, et al: Wild-type KRAS is required for panitumumab efficacy in patients with metastatic colorectal cancer. J Clin Oncol 26:1626-1634, 2008 2. Waks AG, Winer EP: Breast cancer treatment: A review. JAMA 321:288-300, 2019 3. Ritter DI, Roychowdhury S, Roy A, et al: Somatic cancer variant curation and harmonization through consensus minimum variant level data. Genome Med 8:117, 2016 4. Dolman L, Page A, Babb L, et al: ClinGen advancing genomic data-sharing standards as a GA4GH driver project. Hum Mutat 39:1686-1689, 2018 5. Griffith M, Spies NC, Krysiak K, et al: CIViC is a community knowledgebase for expert crowdsourcing the clinical interpretation of variants in cancer. Nat Genet 49:170-174, 2017 6. Boca SM, Panagiotou OA, Rao S, et al: Future of evidence synthesis in precision oncology: Between systematic reviews and biocuration. JCO Precis Oncol 10.1200/PO.17.00175 7. Rao S, Beckman RA, Riazi S, et al: Quantification and expert evaluation of evidence for chemopredictive biomarkers to personalize cancer treatment. Oncotarget 8:37923-37934, 2017 8. Chang W, Cheng J, Allaire JJ, et al: Shiny: Web application framework for R. https://rdrr.io/cran/shiny/ 9. CDGnet: Therapy recommendations using biological networks. http://epiviz.cbcb.umd.edu/shiny/CDGnet/ 10. Vogelstein B, Kinzler KW: Cancer genes and the pathways they control. Nat Med 10:789-799, 2004 11. Cancer Genome Atlas Network: Comprehensive molecular characterization of human colon and rectal cancer. Nature 487:330-337, 2012 12. Kanehisa M, Furumichi M, Tanabe M, et al: KEGG: New perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res 45:D353-D361, 2017 13. US National Library of Medicine: DailyMed. https://dailymed.nlm.nih.gov/dailymed/ 14. Wishart DS, Feunang YD, Guo AC, et al: DrugBank 5.0: A major update to the DrugBank database for 2018. Nucleic Acids Res 46:D1074-D1082, 2018 15. CDGnet: Repository for code related to the CDGnet tool for prioritizing targeted cancer therapies. https://github.com/SiminaB/CDGnet/ 16. Bailey MH, Tokheim C, Porta-Pardo E, et al: Comprehensive characterization of cancer driver genes and mutations. Cell 173: 371-385.e18, 2018 [Erratum: Cell 174:1034-1035, 2018] 17. CDGnet: CDGnet documentation. https://github.com/SiminaB/CDGnet/blob/master/CDGnet_documentation.pdf 18. National Cancer Institute: Targeted cancer therapies. https://www.cancer.gov/about-cancer/treatment/types/targeted-therapies/targeted-therapies-fact-sheet 19. US Food and Drug Administration: List of cleared or approved companion diagnostic devices (in vitro and imaging tools). https://www.fda.gov/medicaldevices/ productsandmedicalprocedures/invitrodiagnostics/ucm301431.htm 20. US Food and Drug Administration: Nucleic acid based tests. https://www.fda.gov/medical-devices/vitro-diagnostics/nucleic-acid-based-tests 21. Tenenbaum D: KEGGREST: Client-side REST access to KEGG. https://rdrr.io/bioc/KEGGREST/ 22. Zhang JD, Wiemann S: KEGGgraph: A graph approach to KEGG PATHWAY in R and Bioconductor. Bioinformatics 25:1470-1471, 2009 23. Carlson M: Genome wide annotation for human. http://bioconductor.org/packages/release/data/annotation/html/org.Hs.eg.db.html 24. Thurnherr T, Singer F, Stekhoven DJ, et al: Genomic variant annotation workflow for clinical applications. F1000 Res 5:1963, 2016 25. Wagner AH, Coffman AC, Ainscough BJ, et al: DGIdb 2.0: Mining clinically relevant drug-gene interactions. Nucleic Acids Res 44:D1036-D1044, 2016 26. US Food and Drug Administration: Drugs@FDA data files. http://www.fda.gov/Drugs/InformationOnDrugs/ucm079750.htm 27. Shiny Component: API for interactive network visualization of gene-drug and gene-gene interactions. https://github.com/jkanche/nfpmShinyComponent 28. Kim S, Chen J, Cheng T, et al: PubChem 2019 update: Improved access to chemical data. Nucleic Acids Res 47:D1102-D1109, 2019 29. Bostock M, Ogievetsky V, Heer J: D : Data-driven documents. IEEE Trans Vis Comput Graph 17:2301-2309, 2011 30. Gao J, Aksoy BA, Dogrusoz U, et al: Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 6:pl1, 2013 31. Cerami E, Gao J, Dogrusoz U, et al: The cBio cancer genomics portal: An open platform for exploring multidimensional cancer genomics data. Cancer Discov 2:401-404, 2012 32. Ciriello G, Gatza ML, Beck AH, et al: Comprehensive molecular portraits of invasive lobular breast cancer. Cell 163:506-519, 2015 33. Formisano L, Stauffer KM, Young CD, et al: Association of FGFR1 with ERαLmaintains ligand-independent ER transcription and mediates resistance to estrogen deprivation in ER+ breast cancer. Clin Cancer Res 23:6138-6150, 2017 [Erratum: Clin Cancer Res 25:1433, 2019] 34. Formisano L, Lu Y, Servetto A, et al: Aberrant FGFR signaling mediates resistance to CDK4/6 inhibitors in ER+ breast cancer. Nat Commun 10:1373, 2019 35. Perez-Garcia J, Muñoz-Couselo E, Soberino J, et al: Targeting FGFR pathway in breast cancer. Breast 37:126-133, 2018 36. Li MM, Datto M, Duncavage EJ, et al: Standards and guidelines for the interpretation and reporting of sequence variants in cancer: A joint consensus recommendation of the Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists. J Mol Diagn 19:4-23, 37. Zhang W, Chien J, Yong J, et al: Network-based machine learning and graph theory algorithms for precision oncology. NPJ Precis Oncol 1:25, 2017 38. Jiang XL, Martinez-Ledesma E, Morcos F: Revealing protein networks and gene-drug connectivity in cancer from direct information. Sci Rep 7:3739, 2017 39. Zhang Y, Tao C: Network analysis of cancer-focused association network reveals distinct network association patterns. Cancer Inform 13:45-51, 2014 (suppl 3) [Erratum: Cancer Inform 13:89, 2014 (suppl 3)] 40. Cheng F, Liu C, Jiang J, et al: Prediction of drug-target interactions and drug repositioning via network-based inference. PLOS Comput Biol 8:e1002503, 2012 41. CIViC: Clinical Interpretations of Variants in Cancer. https://civicdb.org/ 42. Tamborero D, Rubio-Perez C, Deu-Pons J, et al: Cancer Genome Interpreter annotates the biological and clinical relevance of tumor alterations. Genome Med 10:25, 2018 43. Cancer Genome Interpreter: Cancer Biomarkers database. https://www.cancergenomeinterpreter.org/biomarkers 44. Chakravarty D, Gao J, Phillips SM, et al: OncoKB: A precision oncology knowledge base. JCO Precis Oncol 10.1200/PO.17.00011 45. OncoKB: Precision oncology knowledge base. https://oncokb.org/ 46. Sun SQ, Mashl RJ, Sengupta S, et al: Database of evidence for precision oncology portal. Bioinformatics 34:4315-4317, 2018 47. DEPO: Database of Evidence for Precision Oncology. http://depo-dinglab.ddns.net/ 48. Huang L, Fernandes H, Zia H, et al: The cancer precision medicine knowledge base for structured clinical-grade mutations and interpretations. J Am Med Inform Assoc 24:513-519, 2017 49. PMKB: Precision Medicine Knowledge Base. https://pmkb.weill.cornell.edu/ JCO Clinical Cancer Informatics 79 Kancherla et al 50. Landrum MJ, Lee JM, Benson M, et al: ClinVar: Public archive of interpretations of clinically relevant variants. Nucleic Acids Res 44:D862-D868, 2016 51. Prahallad A, Sun C, Huang S, et al: Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR. Nature 483:100-103, 52. Solit DB, Janne PA: Translational medicine: Primed for resistance. Nature 483:44-45, 2012 53. Chandarlapaty S, Sawai A, Scaltriti M, et al: AKT inhibition relieves feedback suppression of receptor tyrosine kinase expression and activity. Cancer Cell 19:58-71, 2011 54. Madhavan S, Ritter D, Micheel C, et al: ClinGen Cancer Somatic Working Group: Standardizing and democratizing access to cancer molecular diagnostic data to drive translational research. Pac Symp Biocomput 23:247-258, 2018 55. Wagner AH, Walsh B, Mayfield G, et al: A harmonized meta-knowledgebase of clinical interpretations of cancer genomic variants. https://www.biorxiv.org/ content/10.1101/366856v1 56. PreMedKB: Precision Medicine Knowledgebase. http://www.fudan-pgx.org/premedkb/index.html#/home 57. Yu Y, Wang Y, Xia Z, et al: PreMedKB: An integrated precision medicine knowledgebase for interpreting relationships between diseases, genes, variants and drugs. Nucleic Acids Res 47:D1090-D1101, 2019 58. Seqome: Drug Gene Interaction Network. http://seqome.com/drug-gene-network/ nn n 80 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies APPENDIX TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source ER Abemaciclib (Verzenio) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Abemaciclib (Verzenio) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Abiraterone Prostate cancer Target NCI/FDA acetate (Zytiga) * Acalabrutinib (Calquence) Mantle cell lymphoma Target NCI/FDA HER2 Ado-trastuzumab emtansine (Kadcyla) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression EGFR Afatinib dimaleate (Gilotrif) NSCLC Sensitive Mutation Exon 19 deletion NCI/FDA EGFR Afatinib dimaleate (Gilotrif) NSCLC Sensitive Mutation L858R NCI/FDA * Afatinib dimaleate (Gilotrif) NSCLC Target NCI/FDA ALK Alectinib (Alecensa) NSCLC Sensitive Mutation Rearrangement NCI/FDA * Alemtuzumab (Campath) B-cell chronic Target NCI/FDA lymphocytic leukemia * Alitretinoin (Panretin) Kaposi sarcoma Target NCI/FDA ER Anastrozole (Arimidex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Anastrozole (Arimidex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Apalutamide (Erleada) Prostate cancer Target NCI/FDA * Atezolizumab (Tecentriq) NSCLC Target NCI/FDA PD-L1 Atezolizumab (Tecentriq) Urothelial carcinoma Sensitive Gene or protein Overexpression NCI/FDA expression * Atezolizumab (Tecentriq) Urothelial carcinoma Target NCI/FDA * Avelumab (Bavencio) Urothelial carcinoma Target NCI/FDA * Avelumab (Bavencio) Merkel cell carcinoma Target NCI/FDA * Axicabtagene ciloleucel (Yescarta) Large B-cell lymphoma Target NCI/FDA * Axitinib (Inlyta) Kidney cancer Target NCI/FDA * Belinostat (Beleodaq) Peripheral T-cell Target NCI/FDA lymphoma * Bevacizumab (Avastin) Glioblastoma Target NCI/FDA * Bevacizumab (Avastin) Cervical cancer Target NCI/FDA * bevacizumab (Avastin) Colorectal cancer Target NCI/FDA * Bevacizumab (Avastin) Fallopian tube cancer Target NCI/FDA * Bevacizumab (Avastin) Renal cell carcinoma Target NCI/FDA * Bevacizumab (Avastin) NSCLC Target NCI/FDA * Bevacizumab (Avastin) Ovarian cancer Target NCI/FDA * Bevacizumab (Avastin) Primary peritoneal Target NCI/FDA cancer * Bexarotene (Targretin) Cutaneous T-cell Target NCI/FDA lymphoma * Blinatumomab (Blincyto) B-cell precursor acute Target NCI/FDA lymphoblastic leukemia * Bortezomib (Velcade) Mantle cell lymphoma Target NCI/FDA * Bortezomib (Velcade) Multiple myeloma Target NCI/FDA (Continued on following page) JCO Clinical Cancer Informatics 81 Kancherla et al TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source ABL/BCR Bosutinib (Bosulif) Chronic myeloid Sensitive Mutation Fusion NCI/FDA leukemia * Brentuximab vedotin (Adcetris) Classic Hodgkin Target NCI/FDA lymphoma * Brentuximab vedotin (Adcetris) Systemic anaplastic Target NCI/FDA large-cell lymphoma * Brentuximab vedotin (Adcetris) Primary cutaneous Target NCI/FDA anaplastic large-cell lymphoma CD30 Brentuximab vedotin (Adcetris) Mycosis fungoides Sensitive Gene or protein Overexpression NCI/FDA expression ALK Brigatinib (Alunbrig) NSCLC Sensitive Mutation Rearrangement NCI/FDA * Cabazitaxel (Jevtana) Prostate cancer Target NCI/FDA * Cabozantinib (Cabometyx) Renal cell carcinoma Target NCI/FDA * Cabozantinib (Cometriq) Medullary thyroid cancer Target NCI/FDA * Carfilzomib (Kyprolis) Multiple myeloma Target NCI/FDA ALK Ceritinib (LDK378/Zykadia) NSCLC Sensitive Mutation Rearrangement NCI/FDA KRAS Cetuximab (Erbitux) Colorectal cancer Sensitive Mutation WT NCI/FDA * Cetuximab (Erbitux) Head and neck Target NCI/FDA squamous cell carcinoma BRAF Cobimetinib (Cotellic) Melanoma Sensitive Mutation V600E NCI/FDA BRAF Cobimetinib (Cotellic) Melanoma Sensitive Mutation V600K NCI/FDA * Copanlisib hydrochloride (Aliqopa) Follicular Target NCI/FDA lymphoma ALK Crizotinib NSCLC Sensitive Mutation Rearrangement NCI/FDA (Xalkori) ROS1 Crizotinib NSCLC Sensitive Mutation Rearrangement NCI/FDA (Xalkori) BRAF Dabrafenib (Tafinlar) NSCLC Sensitive Mutation V600E NCI/FDA BRAF Dabrafenib (Tafinlar) Melanoma Sensitive Mutation V600E NCI/FDA BRAF Dabrafenib (Tafinlar) Melanoma Sensitive Mutation V600K NCI/FDA BRAF Dabrafenib (Tafinlar) Anaplastic thyroid cancer Sensitive Mutation V600E NCI/FDA * Daratumumab (Darzalex) Multiple myeloma Target NCI/FDA ABL/BCR Dasatinib Chronic myeloid Sensitive Mutation Fusion NCI/FDA (Sprycel) leukemia ABL/BCR Dasatinib Acute lymphoblastic Sensitive Mutation Fusion NCI/FDA (Sprycel) leukemia * Denileukin Cutaneous T-cell Target NCI/FDA diftitox (Ontak) lymphoma * Denosumab Giant-cell tumor of the Target NCI/FDA (Xgeva) bone * Dinutuximab (Unituxin) Neuroblastoma Target NCI/FDA * Durvalumab (Imfinzi) Urothelial carcinoma Target NCI/FDA * Durvalumab (Imfinzi) NSCLC Target NCI/FDA * Elotuzumab (Empliciti) Multiple myeloma Target NCI/FDA IDH2 Enasidenib AML Sensitive Mutation R140Q NCI/FDA mesylate (Idhifa) IDH2 Enasidenib AML Sensitive Mutation R140L NCI/FDA mesylate (Idhifa) (Continued on following page) 82 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source IDH2 Enasidenib AML Sensitive Mutation R140G NCI/FDA mesylate (Idhifa) IDH2 Enasidenib mesylate (Idhifa) AML Sensitive Mutation R140W NCI/FDA IDH2 Enasidenib mesylate (Idhifa) AML Sensitive Mutation R172K NCI/FDA IDH2 Enasidenib AML Sensitive Mutation R172M NCI/FDA mesylate (Idhifa) IDH2 Enasidenib AML Sensitive Mutation R172G NCI/FDA mesylate (Idhifa) IDH2 Enasidenib AML Sensitive Mutation R172S NCI/FDA mesylate (Idhifa) IDH2 Enasidenib AML Sensitive Mutation R172W NCI/FDA mesylate (Idhifa) * Enzalutamide (Xtandi) Prostate cancer Target NCI/FDA EGFR Erlotinib NSCLC Sensitive Mutation Exon 19 deletion NCI/FDA (Tarceva) EGFR Erlotinib NSCLC Sensitive Mutation L858R NCI/FDA (Tarceva) * Erlotinib Pancreatic Target NCI/FDA (Tarceva) cancer ER Everolimus Breast cancer Sensitive Gene or protein Overexpression NCI/FDA (Afinitor) expression PR Everolimus Breast cancer Sensitive Gene or protein Overexpression NCI/FDA (Afinitor) expression * Everolimus Renal cell carcinoma Target NCI/FDA (Afinitor) * Everolimus GI neuroendocrine tumor Target NCI/FDA (Afinitor) * Everolimus Pulmonary Target NCI/FDA (Afinitor) neuroendocrine tumor * Everolimus Pancreatic Target NCI/FDA (Afinitor) neuroendocrine tumor * Everolimus Tuberous sclerosis Target NCI/FDA (Afinitor) complex–associated renal angiomyolipoma * Everolimus Tuberous sclerosis Target NCI/FDA (Afinitor) complex–associated subependymal giant cell astrocytoma ER Exemestane (Aromasin) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression ER Fulvestrant (Faslodex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Fulvestrant (Faslodex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression EGFR Gefitinib (Iressa) NSCLC Sensitive Mutation Exon 19 deletion NCI/FDA EGFR Gefitinib (Iressa) NSCLC Sensitive Mutation L858R NCI/FDA CD33 Gemtuzumab ozogamicin (Mylotarg) AML Sensitive Gene or protein Overexpression NCI/FDA expression * Ibritumomab tiuxetan (Zevalin) Follicular B-cell Target NCI/FDA non-Hodgkin lymphoma * Ibritumomab tiuxetan (Zevalin) Follicular non-Hodgkin Target NCI/FDA lymphoma (Continued on following page) JCO Clinical Cancer Informatics 83 Kancherla et al TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source * Ibrutinib Mantle cell lymphoma Target NCI/FDA (Imbruvica) * Ibrutinib Marginal zone lymphoma Target NCI/FDA (Imbruvica) * Ibrutinib Chronic lymphocytic Target NCI/FDA (Imbruvica) leukemia * Ibrutinib Waldenstrom Target NCI/FDA (Imbruvica) macroglobulinemia * Idelalisib (Zydelig) Chronic lymphocytic Target NCI/FDA leukemia * Idelalisib (Zydelig) Follicular B-cell Target NCI/FDA non-Hodgkin lymphoma * Idelalisib (Zydelig) Small lymphocytic Target NCI/FDA lymphoma * Imatinib mesylate (Gleevec) Dermatofibrosarcoma Target NCI/FDA protuberans c-Kit Imatinib mesylate (Gleevec) GI stromal tumor Sensitive Gene or protein Overexpression NCI/FDA expression ABL/BCR Imatinib mesylate (Gleevec) Chronic myeloid Sensitive Mutation Fusion NCI/FDA leukemia ABL/BCR Imatinib mesylate (Gleevec) Acute lymphoblastic Sensitive Mutation Fusion NCI/FDA leukemia PDGFRB Imatinib mesylate (Gleevec) Myelodysplastic/ Sensitive Mutation Rearrangement NCI/FDA myeloproliferative disorders c-Kit Imatinib mesylate (Gleevec) Systemic mastocytosis Resistant Mutation D816V NCI/FDA c-Kit Imatinib mesylate (Gleevec) Systemic mastocytosis Sensitive Mutation Unknown status NCI/FDA * Inotuzumab ozogamicin (Besponsa) B-cell precursor acute Target NCI/FDA lymphoblastic leukemia * Ipilimumab (Yervoy) Renal cell carcinoma Target NCI/FDA * Ipilimumab (Yervoy) Melanoma Target NCI/FDA MSI-H Ipilimumab (Yervoy) Colorectal cancer Sensitive MSI-H Positive NCI/FDA dMMR Ipilimumab (Yervoy) Colorectal cancer Sensitive dMMR Positive NCI/FDA * Ixazomib citrate (Ninlaro) Multiple myeloma Target NCI/FDA * Lanreotide acetate (Somatuline Depot) Gastroenteropancreatic Target NCI/FDA neuroendocrine tumor HER2 Lapatinib (Tykerb) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Lenvatinib mesylate (Lenvima) Renal cell carcinoma Target NCI/FDA * Lenvatinib mesylate (Lenvima) Differentiated thyroid Target NCI/FDA cancer ER Letrozole (Femara) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Letrozole (Femara) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Lutetium Lu-dotatate (Lutathera) Gastroenteropancreatic Target NCI/FDA neuroendocrine tumor FLT3 Midostaurin (Rydapt) AML Sensitive Mutation Internal tandem NCI/FDA duplication FLT3 Midostaurin (Rydapt) AML Sensitive Mutation D835X NCI/FDA (Continued on following page) 84 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source FLT3 Midostaurin (Rydapt) AML Sensitive Mutation I836X NCI/FDA * Midostaurin (Rydapt) Mast cell leukemia Target NCI/FDA * Midostaurin (Rydapt) Systemic mastocytosis Target NCI/FDA * Necitumumab (Portrazza) NSCLC Target NCI/FDA HER2 Neratinib maleate (Nerlynx) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression ABL/BCR Nilotinib (Tasigna) Chronic myeloid Sensitive Mutation Fusion NCI/FDA leukemia * Niraparib tosylate monohydrate (Zejula) Fallopian tube cancer Target NCI/FDA * Niraparib tosylate monohydrate (Zejula) Ovarian cancer Target NCI/FDA * Niraparib tosylate monohydrate (Zejula) Primary peritoneal Target NCI/FDA cancer MSI-H Nivolumab (Opdivo) Colorectal cancer Sensitive MSI-H Positive NCI/FDA dMMR Nivolumab (Opdivo) Colorectal cancer Sensitive dMMR Positive NCI/FDA * Nivolumab (Opdivo) Head and neck Target NCI/FDA squamous cell carcinoma * Nivolumab (Opdivo) Renal cell carcinoma Target NCI/FDA * Nivolumab (Opdivo) Hepatocellular Target NCI/FDA carcinoma * Nivolumab (Opdivo) NSCLC Target NCI/FDA * Nivolumab (Opdivo) Classic Hodgkin Target NCI/FDA lymphoma * Nivolumab (Opdivo) Melanoma Target NCI/FDA * Nivolumab (Opdivo) Urothelial carcinoma Target NCI/FDA * Obinutuzumab (Gazyva) Chronic lymphocytic Target NCI/FDA leukemia * Obinutuzumab (Gazyva) Follicular lymphoma Target NCI/FDA * Ofatumumab (Arzerra) Chronic lymphocytic Target NCI/FDA leukemia BRCA1 Olaparib (Lynparza) Breast cancer Sensitive Mutation Deleterious NCI/FDA BRCA2 Olaparib (Lynparza) Breast cancer Sensitive g.mutation Deleterious NCI/FDA * Olaparib (Lynparza) Fallopian tube cancer Target NCI/FDA BRCA1 Olaparib (Lynparza) Ovarian cancer Sensitive g.mutation Deleterious NCI/FDA BRCA2 Olaparib (Lynparza) Ovarian cancer Sensitive g.mutation Deleterious NCI/FDA * Olaparib (Lynparza) Ovarian cancer Target NCI/FDA * Olaparib (Lynparza) Primary peritoneal Target NCI/FDA cancer * Olaratumab (Lartruvo) Soft tissue sarcoma Target NCI/FDA EGFR Osimertinib (Tagrisso) NSCLC Sensitive Mutation Exon 19 deletion NCI/FDA EGFR Osimertinib (Tagrisso) NSCLC Sensitive Mutation L858R NCI/FDA EGFR Osimertinib (Tagrisso) NSCLC Sensitive Mutation T790M NCI/FDA ER Palbociclib (Ibrance) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Palbociclib (Ibrance) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression KRAS Panitumumab (Vectibix) Colorectal cancer Sensitive Mutation WT NCI/FDA (Continued on following page) JCO Clinical Cancer Informatics 85 Kancherla et al TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source NRAS Panitumumab (Vectibix) Colorectal cancer Sensitive Mutation WT NCI/FDA * Panobinostat (Farydak) Multiple myeloma Target NCI/FDA * Pazopanib (Votrient) Renal cell carcinoma Target NCI/FDA * Pazopanib (Votrient) Soft tissue sarcoma Target NCI/FDA PD-L1 Pembrolizumab (Keytruda) Urothelial carcinoma Sensitive Gene or protein Overexpression NCI/FDA expression PD-L1 Pembrolizumab (Keytruda) Cervical cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Pembrolizumab (Keytruda) Head and neck Target NCI/FDA squamous cell carcinoma * Pembrolizumab (Keytruda) NSCLC Target NCI/FDA PD-L1 Pembrolizumab (Keytruda) NSCLC Sensitive Gene or protein Overexpression NCI/FDA expression * Pembrolizumab (Keytruda) Classic Hodgkin Target NCI/FDA lymphoma * Pembrolizumab (Keytruda) Primary mediastinal large Target NCI/FDA B-cell lymphoma MSI-H Pembrolizumab (Keytruda) Colorectal cancer Sensitive MSI-H Positive NCI/FDA dMMR pembrolizumab (Keytruda) Colorectal cancer Sensitive dMMR Positive NCI/FDA MSI-H Pembrolizumab (Keytruda) Solid tumors Sensitive MSI-H Positive NCI/FDA dMMR Pembrolizumab (Keytruda) Solid tumors Sensitive dMMR Positive NCI/FDA * Pembrolizumab (Keytruda) Melanoma Target NCI/FDA PD-L1 Pembrolizumab (Keytruda) Gastric cancer Sensitive Gene or protein Overexpression NCI/FDA expression HER2 Pertuzumab (Perjeta) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Ponatinib hydrochloride (Iclusig) Chronic myeloid Target NCI/FDA leukemia ABL/BCR Ponatinib hydrochloride (Iclusig) Chronic myeloid Sensitive Mutation T315I NCI/FDA leukemia ABL/BCR Ponatinib hydrochloride (Iclusig) AML Sensitive Mutation Fusion NCI/FDA ABL/BCR Ponatinib hydrochloride (Iclusig) AML Sensitive Mutation T315I NCI/FDA * Pralatrexate (Folotyn) Peripheral T-cell Target NCI/FDA lymphoma * Radium 223 dichloride (Xofigo) Prostate cancer Target NCI/FDA * Ramucirumab (Cyramza) Colorectal cancer Target NCI/FDA * Ramucirumab (Cyramza) Gastric cancer Target NCI/FDA * Ramucirumab (Cyramza) NSCLC Target NCI/FDA * Regorafenib (Stivarga) Colorectal cancer Target NCI/FDA * Regorafenib (Stivarga) GI stromal tumor Target NCI/FDA * Regorafenib (Stivarga) Hepatocellular Target NCI/FDA carcinoma ER Ribociclib (Kisqali) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Ribociclib (Kisqali) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression CD20 Rituximab (Rituxan) Chronic lymphocytic Sensitive Gene or protein Overexpression NCI/FDA leukemia expression (Continued on following page) 86 © 2020 by American Society of Clinical Oncology Evidence-Based Network Approach to Recommend Targeted Therapies TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source CD20 Rituximab (Rituxan) Non-Hodgkin lymphoma Sensitive Gene or protein Overexpression NCI/FDA expression * Rituximab and hyaluronidase human Chronic lymphocytic Target NCI/FDA (Rituxan Hycela) leukemia * Rituximab and hyaluronidase human DLBCL Target NCI/FDA (Rituxan Hycela) * Rituximab and hyaluronidase human Follicular lymphoma Target NCI/FDA (Rituxan Hycela) * Romidepsin (Istodax) Cutaneous T-cell Target NCI/FDA lymphoma * Romidepsin (Istodax) Peripheral T-cell Target NCI/FDA lymphoma * Rucaparib camsylate (Rubraca) Fallopian tube cancer Target NCI/FDA BRCA1 Rucaparib camsylate (Rubraca) Fallopian tube cancer Sensitive Mutation Deleterious NCI/FDA BRCA2 Rucaparib camsylate (Rubraca) Fallopian tube cancer Sensitive Mutation Deleterious NCI/FDA * Rucaparib camsylate (Rubraca) Ovarian cancer Target NCI/FDA BRCA1 Rucaparib camsylate (Rubraca) Ovarian cancer Sensitive Mutation Deleterious NCI/FDA BRCA2 Rucaparib camsylate (Rubraca) Ovarian cancer Sensitive Mutation Deleterious NCI/FDA * Rucaparib camsylate (Rubraca) Primary peritoneal Target NCI/FDA cancer BRCA1 Rucaparib camsylate (Rubraca) Primary peritoneal Sensitive Mutation Deleterious NCI/FDA cancer BRCA2 Rucaparib camsylate (Rubraca) Primary peritoneal Sensitive Mutation Deleterious NCI/FDA cancer * Ruxolitinib phosphate (Jakafi) Myelofibrosis Target NCI/FDA * Ruxolitinib phosphate (Jakafi) Polycythemia vera Target NCI/FDA * Siltuximab (Sylvant) Multicentric Castleman Target NCI/FDA disease * Sonidegib (Odomzo) Basal cell carcinoma Target NCI/FDA * Sorafenib (Nexavar) Renal cell carcinoma Target NCI/FDA * Sorafenib (Nexavar) Hepatocellular Target NCI/FDA carcinoma * Sorafenib (Nexavar) Thyroid carcinoma Target NCI/FDA c-Kit Sunitinib (Sutent) GI stromal tumor Sensitive Gene or protein Overexpression NCI/FDA expression * Sunitinib (Sutent) Renal cell carcinoma Target NCI/FDA * Sunitinib (Sutent) Pancreatic Target NCI/FDA neuroendocrine tumor ER Tamoxifen (Nolvadex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression PR Tamoxifen (Nolvadex) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression * Temsirolimus (Torisel) Renal cell carcinoma Target NCI/FDA * Tisagenlecleucel (Kymriah) B-cell acute Target NCI/FDA lymphoblastic leukemia * Tisagenlecleucel (Kymriah) DLBCL Target NCI/FDA ER Toremifene (Fareston) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression BRAF Trametinib (Mekinist) NSCLC Sensitive Mutation V600E NCI/FDA BRAF Trametinib (Mekinist) Melanoma Sensitive Mutation V600E NCI/FDA (Continued on following page) JCO Clinical Cancer Informatics 87 Kancherla et al TABLE A1. Curated FDA-Targeted Cancer Therapies Using Label Information (Continued) Data Type (gene Association/Predicted expression, DNA copy Alteration (over- Effect (resistant or number, gene orunderexpressed, sensitive to treatment mutation, protein mutation type based on Gene/Protein Drug Disease in humans) expression) data type) Source BRAF Trametinib (Mekinist) Melanoma Sensitive Mutation V600K NCI/FDA BRAF Trametinib (Mekinist) Anaplastic thyroid cancer Sensitive Mutation V600E NCI/FDA HER2 Trastuzumab (Herceptin) Breast cancer Sensitive Gene or protein Overexpression NCI/FDA expression HER2 Trastuzumab (Herceptin) Gastric cancer Sensitive Gene or protein Overexpression NCI/FDA expression RARA Tretinoin (Vesanoid) Acute promyelocytic Sensitive Mutation t(15;19) NCI/FDA leukemia * Vandetanib (Caprelsa) Medullary thyroid cancer Target NCI/FDA BRAF Vemurafenib (Zelboraf) Erdheim-Chester disease Sensitive Mutation V600X NCI/FDA BRAF Vemurafenib (Zelboraf) Melanoma Sensitive Mutation V600E NCI/FDA * Venetoclax (Venclexta) Chronic lymphocytic Target NCI/FDA leukemia * Venetoclax (Venclexta) Small lymphocytic Target NCI/FDA lymphoma * Vismodegib (Erivedge) Basal cell carcinoma Target NCI/FDA * Vorinostat (Zolinza) Cutaneous T-cell Target NCI/FDA lymphoma * Ziv-aflibercept (Zaltrap) Colorectal cancer Target NCI/FDA NOTE. Last updated on July 12, 2018. Asterisks denote no biomarker indication. Abbreviations: AML, acute myeloid leukemia; DLBCL, diffuse large B-cell lymphoma; dMMR, deficient mismatch repair; EGFR, epidermal growth factor receptor; ER, estrogen receptor; FDA, US Food and Drug Administration; HER2, human epidermal growth factor receptor 2; MSI-H, microsatellite instability high; NCI, National Cancer Institute; NSCLC, non–small-cell lung cancer; PD-L1, programmed death 1 ligand; PR, progesterone receptor; WT, wild type. 88 © 2020 by American Society of Clinical Oncology

Journal

JCO Clinical Cancer InformaticsWolters Kluwer Health

Published: Jan 28, 2020

There are no references for this article.