Scientific Contributions
Table of Contents
Ongoing Projects
ARPA-H funded AI-driven Data Extraction
I currently serve as a co-investigator for the $1.9 million ARPA-H funded program for AI-driven data extraction from unstructured EHR text. We are refining a toolset called BRIM analytics, developed by the grant co-PI Daniel Fabbri, PhD, which we uses large language models to assist in extracting and organizing critical information from unstructured clinical notes and reports. With co-PI Christine Micheel, PhD, I offer scientific expertise in the clinical setting of cancer genomics and precision biomarkers, cancer precision therapies, and cancer clinical trials, in order to train the AI models, develop protocols for their extension to other settings, and design user-centered features that offer the broadest relevance to research and clinical applications. In the oncology setting streamlining data extraction will enable researchers to evaluate therapy efficacy more quickly and help oncologists quickly review patient histories. By making our tool broadly accessible, we hope to democratize precision cancer care delivery, including matching diverse patients to relevant clinical trials.
ARPA-H has a goal to fund high-risk, high-reward projects that have the potential to develop analytic infrastructure with broad-reaching applications across the public and private sectors, including those that improve emergency readiness. For instance, while our tool will be developed and trained in the setting of cancer-related data abstraction, this type of AI-driven EHR-based note extraction has applications in rapidly evolving infectious disease settings and could speed pandemic response.
You can find a brief description of our award from ARPA-H.
Electronic Medical Records & Genomics (eMERGE) Network
I currently serve as the EHR Integration site lead for Vanderbilt University Medical Center (VUMC) on the NHGRI-funded Electronic Medical Records & Genomics (eMERGE) Network. This role involves working directly with VUMC HealthIT to integrate study-generated discrete and non-discrete data, including family history and genomic data, directly into the EHR. The eMERGE study is utilizing MeTree™ for family history assessment, and I am overseeing the implementation of MeTree™ as a SMART on FHIR application at our site. I am also directing the integration of study-generated clinical decision support to the EHR.
Publications
Linder JE, Allworth A, Bland ST, Caraballo PJ, Chisholm RL, Clayton EW, Crosslin DR, Dikilitas O, DiVietro A, Esplin ED, Forman S, Freimuth RR, Gordon AS, Green R, Harden MV, Holm IA, Jarvik GP, Karlson EW, Labrecque S, Lennon NJ, Limdi NA, Mittendorf KF, et al. Returning integrated genomic risk and clinical recommendations: The eMERGE study. Genet Med. 2023 Jan 6;25(4):100006. doi: 10.1016/j.gim.2023.100006. Epub ahead of print. PMID: 36621880.
Song J, Forrest N, Gordon A, Kottyan L, Mittendorf KF, Wei WQ, Ramsey-Goldman R, Walunas T, Kho A. Utilization of electronic health record data to evaluate the association of urban environment on systemic lupus erythematosus symptoms. Rheumatology (Oxford). 2022 Nov 16:keac647. doi: 10.1093/rheumatology/keac647. Epub ahead of print. PMID: 36383166.
Family History and Cancer Risk Study (FOREST)
On the FOREST study, my role is to serve various needs in the fields of my expertise. I primarily work in the development of analyses plans, study outcomes measurement, and informatics spaces. In the informatics space of the project my work includes ensuring productive implementation of MeTree™ for family history assessment as a SMART on FHIR application that does not interfere with the MeTree™ implementation in eMERGE (see below). Further, I work on developing data models to track participant interactions.
Publications
Mittendorf KF, Bland HT, Andujar J, Celaya-Cobbs N, Edwards C, Gerhart M, Hooker G, Hubert M, Jones SH, Marshall DR, Myers RA, Pratap S, Rosenbloom ST, Sadeghpour A, Wu RR, Orlando LA, Wiesner GL. Family history and cancer risk study (FOREST): A clinical trial assessing electronic patient-directed family history input for identifying patients at risk of hereditary cancer. Contemp Clin Trials. 2024 Oct 10:107714. doi: 10.1016/j.cct.2024.107714. Epub ahead of print. PMID: 39395532. DOI: 10.1016/j.cct.2024.107714
Bland HT, Gilmore MJ, Andujar J, Martin MA, Celaya-Cobbs N, Edwards C, Gerhart M, Hooker GW, Kraft SA, Marshall DR, Orlando LA, Paul NA, Pratap S, Rosenbloom ST, Wiesner GL, Mittendorf KF*. Conducting inclusive research in genetics for transgender, gender-diverse, and sex-diverse individuals: Case analyses and recommendations from a clinical genomics study. J Genet Couns. 2023 Sep 4. doi: 10.1002/jgc4.1785. Epub ahead of print. PMID: 37667436. PMCID: PMC10909936
Health and Healthcare of Transgender, gender-diverse, and sex-diverse persons
I currently work on several projects related to the health and healthcare of transgender, gender-diverse, and sex-diverse (TGSD) persons. I use the terminology TGSD to broadly encompass individuals 1) who are transgender, non-binary, or intersex, 2) whose gender identity does not align with their sex assigned at birth, 3) who have a variance of sex characteristic (e.g., who are intersex or otherwise have a sex characteristic that does not align with medical binary constructs of sex phenotypes), or 4) whose reproductive development and/or whose gender identity or expression is characterized by non-binary constructs of sex phenotype and or gender.
Areas of interest include:
Accurate and respectful data capture of constructs related to sex phenotypes and gender identity/expression for informing healthcare delivery and health disparities research
Genetic testing in TGSD people to inform cancer preventive services (e.g., hereditary cancer testing) and gender-affirming surgical decision making (e.g., extent of resection during mastectomy)
Inclusive precision oncology care in hormone-responsive cancers in individuals receiving gender-affirming hormones
The influence of gender affirming hormones on cardiac health and cardiac screening needs in TGSD people receiving hormonal care
Data privacy of LGBTQ+ and TGSD persons in an era of NIH-mandated data sharing and increasing legislative attacks on LGBTQ+ rights
Healthcare policy surrounding transgender healthcare
Publications
Clayton EW, Bland HT, Mittendorf KF. Protecting Privacy of Pregnant and LGBTQ+ Research Participants. JAMA. 2024 Apr 15. doi: 10.1001/jama.2024.4837. Epub ahead of print. PMID: 38619831.
Bland HT, Gilmore MJ, Andujar J, Martin MA, Celaya-Cobbs N, Edwards C, Gerhart M, Hooker GW, Kraft SA, Marshall DR, Orlando LA, Paul NA, Pratap S, Rosenbloom ST, Wiesner GL, Mittendorf KF*. Conducting inclusive research in genetics for transgender, gender-diverse, and sex-diverse individuals: Case analyses and recommendations from a clinical genomics study. J Genet Couns. 2023 Sep 4. doi: 10.1002/jgc4.1785. Epub ahead of print. PMID: 37667436. PMCID: PMC10909936
Rolf BA, Schneider JL, Amendola LM, Davis JV, Mittendorf KF, Schmidt MA, Jarvik GP, Wilfond BS, Goddard KAB, Ezzell Hunter J. Barriers to family history knowledge and family communication among LGBTQ+ individuals in the context of hereditary cancer risk assessment. Journal of Genetic Counseling. 2022 Feb;31(1):230-241. doi: 10.1002/jgc4.1476. Epub 2021 Jul 23. PubMed PMID: 34302314; PubMed Central PMCID: PMC8783924.
*senior author
CHESTcare: Cancer & Hereditary Risk Education & Support for Transgender and Nonbinary Individuals
Kim Zayhowski, MGC, CGC and I are co-developing a patient- and provider-facing educational tool on breast/chest cancer risk for transgender men and other gender-diverse persons where the currently heavily gendered nature of breast/chest cancer care is harmful. We are using a community-engaged approach where patient and providers act as collaborators in user-centered design. We aim to include hereditary cancer risk assessment and genetic testing as key components for qualified patients to inform pre-operative surgical decision making (e.g., extent of tissue resection in gender-affirming mastectomy) and post-operative surveillance. We have applied for NIH funding to support this project. Pilot work is funded to Kim under a National Society of Genetic Counselors Jane Engelberg Memorial Fund Award. More information on the JEMF pilot can be found in this press release about Kim's award.
Pathogenic Variant Prediction for Hereditary Cancer: PREMM5 and PREMMplus
PREMM5 is a risk assessment algorithm that predicts the probability of a pathogenic variant in any of the five Lynch syndrome-related genes. On the PREMM project, I led the adaptation of the PREMM5™ provider-facing application to a patient-facing risk assessment web application in an extensive community-engaged iterative adaptation. I coordinated two multi-site workgroups and an onsite software development team that collaborated with patient stakeholders from Denver Health in an iterative design process to ensure the application was accessible to patients from medically underserved populations. I also led the development of a data model that supported structured capture of participant interactions with the tool and trained the multi-site recruitment team in its use. In my mentored supplement to CHARM, I led an interdisciplinary qualitative and quantitative evaluation of these patient-facing web tools and trained in qualitative interview and analysis techniques. I also received formal didactic training from Oregon Health and Sciences University (OHSU). This version of PREMM5 has since been adapted and incorporated into the EHR at Dana Farber Cancer Institute, where it is administered to GI oncology patients. Further, it has informed the language of the patient-facing adaptation of PREMMplus, an updated algorithm that assesses pathogenic variant risk in 19 genes associated with 18 cancers and neoplasms. I have been working on this adaptation of PREMMplus, and am co-investigator on the PREMMplus R01 (5R01CA132829). I am also be co-investigator on a recently awarded R01 (R01CA292900), whose goal is to adapt implement and test the patient-facing PREMMplus in a diverse population of individuals, including adapting the PREMMplus into a SMART on FHIR application for EHR integration in a vendor-agnostic fashion.
Publications
Mittendorf KF, Lewis HS, Duenas DM, Eubanks DJ, Gilmore MJ, Goddard KAB, Joseph G, Kauffman TL, Kraft SA, Lindberg NM, Reyes AA, Shuster E, Syngal S, Ukaegbu C, Zepp JM, Wilfond BS, Porter KM. Literacy-adapted, electronic family history assessment for genetics referral in primary care: patient user insights from qualitative interviews. Hereditary Cancer in Clinical Practice. 2022 Jun 10;20(1):22. doi: 10.1186/s13053-022-00231-3. PMID: 35689290; PMCID: PMC9188215.
Mittendorf KF*, Ukaegbu C*, Gilmore MJ, Lindberg NM, Kauffman TL, Eubanks DJ, Shuster E, Allen J, McMullen C, Feigelson HS, Anderson KP, Leo MC, Hunter JE, Sasaki SO, Zepp JM, Syngal S, Wilfond BS, Goddard KAB. Adaptation and early implementation of the PREdiction model for gene mutations (PREMM 5™) for lynch syndrome risk assessment in a diverse population. Familial Cancer. 2021 Mar 23. Online ahead of print. doi: 10.1007/s10689-021-00243-3. PubMed PMID: 33754278; PubMed Central PMCID: PMC8458476.
Feigelson HS, Mittendorf KF, Okuyama S, Porter KM, Bulkley J, Shuster E, Anderson KP, Gilmore MJ, Zepp JZ, Kauffman TK, Lindberg NM, Muessig KR, Bellcross C, Ukaegbu C, Syngal S, Leo MC, Wilfond BS, On Behalf of the Cancer Health Assessments Reaching Many (CHARM) Study. Feasibility of an electronic patient-facing cancer family history tool in medically underserved populations. Genet Med Open. Vol 2, 101860. Epub 2024 Jun 25. DOI: https://doi.org/10.1016/j.gimo.2024.101860
Oncology Knowledge Rapid Alerts (OKRA)
I currently serve as key personnel on the NCI-funded OKRA project (1R21CA274545), where we are developing a novel tool to develop and test an EHR-integrated precision oncology clinical decision support tool. This tool leverages biomarker-driven assertions housed in the internationally renowned My Cancer Genome (MCG) knowledge base (see "Past Projects" below) to match patients to appropriate therapeutic assertions based on their diagnosis and NGS genetic testing results. We are initially testing this tool on MCG's assertion knowledgebase for non-small cell lung cancer (NSCLC). We aim to close gaps in NGS-driven prescribing of first line therapies, as over half of community oncologists do not appropriately use NGS to drive treatment decisions. OKRA aims to close these gaps by bringing precision oncology out of academic medical centers and into the community, through an EHR-integrable solution that can be readily ported into other EHR systems.
My role involves working directly with software developers to build a tool that pairs My Cancer Genome assertions with HL7-formatted genetic testing results that are compatible with structured genomics modules in EHRs. This involves defining the algorithmic requirements to match MCG's variant mapping algorithms and assertions algorithms (e.g., when two or more variants are part of an assertion) to HL7-formatted genomics data.
Our first manuscript describing the development of OKRA infrastructure is in preparation. See past projects for more information about MCG.
GE-VUMC Digital Precision Oncology Partnership
I currently serve as scientific personnel a GE funded academic-industry partnership. In this partnership, our goal is to refine precision oncology care delivery. Early in this partnership, our work has primarily focused on developing models for prediction of toxicity and effectiveness of cancer immunotherapy. Immunotherapies show great promise, but up to 50% of patients experience an autoimmune toxicity. In an ideal scenario, oncologists would be able to balance the likelihood of response in a specific patient with their personal risk of toxicity. To do this, we have developed AI-driven4
My roles encompass the development of back-end data models, expert human verification of data curation necessary for model inputs, providing key scientific insight into the modeling components (e.g., suggesting time windows), among others.
A patent application associated with this work is pending. Read a press release about the first publication below from GE Healthcare.
Selected publications and presentations:
Lippenszky L, Mittendorf KF, Kiss Z, LeNoue-Newton ML, Napan-Molina P, Rahman P, Ye C, Laczi B, Csernai E, Jain NM, Holt ME, Maxwell CN, Ball M, Ma Y, Mitchell MB, Johnson DB, Smith DS, Park BH, Micheel CM, Fabbri D, Wolber J, Osterman TJ. Prediction of Effectiveness and Toxicities of Immune Checkpoint Inhibitors Using Real-World Patient Data. JCO Clinical Cancer Informatics. 2024 Feb;8:e2300207. doi: 10.1200/CCI.23.00207. PMID: 38427922; PMCID: PMC10919473.
Rahman P, Ye C, Mittendorf KF, Lenoue-Newton M, Micheel C, Wolber J, Osterman T, Fabbri D. Accelerated curation of checkpoint inhibitor-induced colitis cases from electronic health records. JAMIA Open. 2023 Apr 1;6(1):ooad017. doi: 10.1093/jamiaopen/ooad017. PMID: 37012912; PMCID: PMC10066800.
Past Projects
Cancer Health Assessments Reaching Many (CHARM)
From 2017-2021, I worked on the Cancer Health Assessments Reaching Many (CHARM) study, a CSER Consortium study funded by the NHGRI with co-funding by NCI. CHARM investigates strategies to improve access to genetic counseling and testing for a diverse group of adults with increased risk of hereditary cancer syndromes. For CHARM, I led the adaptation of two validated, guideline-recommended risk assessment algorithms (PREMM5™ and B-RST™ 3.0) to patient-facing web applications for implementation in the low-literacy, low-resource study population. I led two multi-site workgroups (one for each tool) and an onsite software development team, as well as collaborated with patient stakeholders in an iterative design process. I also led the software development team in their implementation of a web-based informed consent for study-provided genomic testing for those participants who screened at risk. Additionally, I worked with the development team to design and implement a data model supporting data capture of patient interactions with the web-based risk assessment and consent, and provided analytic support for downstream analyses leveraging this data model. I later received a career development supplement to CHARM to evaluate the effectiveness of our adaptations at improving access to risk assessment in our underserved study population.
Selected publications and presentations:
Mittendorf KF*, Kauffman TL*, Amendola LM, Anderson KP, Biesecker BB, Dorschner MO, Duenas DM, Eubanks DJ, Feigelson HS, Gilmore MJ, Hunter JE, Joseph G, Kraft SA, Lee SSJ, Leo MC, Liles EG, Lindberg NM, Muessig KR, Okuyama S, Porter KM, Riddle LS, Rolf BA, Rope AF, Zepp JM, Jarvik GP, Wilfond BS, Goddard KAB. Cancer Health Assessments Reaching Many (CHARM): A clinical trial assessing a multimodal cancer genetics services delivery program and its impact on diverse populations. Contemporary Clinical Trials. 2021 Jul;106:106432. doi: 10.1016/j.cct.2021.106432. Epub 2021 May 11. PubMed PMID: 33984519; PubMed Central PMCID: PMC8336568. Corrigendum in doi: 10.1016/j.cct.2022.106682
Mittendorf KF, Lewis HS, Duenas DM, Eubanks DJ, Gilmore MJ, Goddard KAB, Joseph G, Kauffman TL, Kraft SA, Lindberg NM, Reyes AA, Shuster E, Syngal S, Ukaegbu C, Zepp JM, Wilfond BS, Porter KM. Literacy-adapted, electronic family history assessment for genetics referral in primary care: patient user insights from qualitative interviews. Hereditary Cancer in Clinical Practice. 2022 Jun 10;20(1):22. doi: 10.1186/s13053-022-00231-3. PMID: 35689290; PMCID: PMC9188215.
Mittendorf KF*, Ukaegbu C*, Gilmore MJ, Lindberg NM, Kauffman TL, Eubanks DJ, Shuster E, Allen J, McMullen C, Feigelson HS, Anderson KP, Leo MC, Hunter JE, Sasaki SO, Zepp JM, Syngal S, Wilfond BS, Goddard KAB. Adaptation and early implementation of the PREdiction model for gene mutations (PREMM 5™) for lynch syndrome risk assessment in a diverse population. Familial Cancer. 2021 Mar 23. Online ahead of print. doi: 10.1007/s10689-021-00243-3. PubMed PMID: 33754278; PubMed Central PMCID: PMC8458476.
Mittendorf KF, Knerr S, Kauffman TL, Lindberg NM, Anderson KP, Feigelson HS, Gilmore MJ, Hunter JE, Joseph G, Kraft SA, Zepp JM, Syngal S, Wilfond BS, Goddard KAB. Systemic Barriers to Risk-Reducing Interventions for Hereditary Cancer Syndromes: Implications for Health Care Inequities. JCO Precision Oncology. 2021;5. doi: 10.1200/PO.21.00233. eCollection 2021. Review. PubMed PMID: 34778694; PubMed Central PMCID: PMC8585306.
*first author or co-first author
Clinical Genome Resource (ClinGen)
From 2017-2021, I was a member of the Clinical Genome Resource (ClinGen) Actionability Working Group. In this role, I assisted with adaptation of an actionability protocol from the adult setting to the pediatric setting. These protocols outline a standardized process to identify and synthesize evidence regarding the clinical actionability of genes and disorders associated with secondary findings during genetic testing. I applied both protocols to curate reports that were scored by experts and disseminated to the public on clinicalgenome.org. I also applied my experience with structured genomic data models and ontologies to provide user specifications for the actionability curation interface. The results of this work are used by the ACMG Secondary Findings WG and the Centers for Disease Control to provide professional recommendations about the return of secondary findings to patients undergoing genome-wide sequencing in both the adult and pediatric settings.
Selected publications and presentations:
Paquin RS, Mittendorf KF, Lewis MA, Hunter JE, Lee K, Berg JS, Williams MS, Goddard KAB. Expert and lay perspectives on burden, risk, tolerability and acceptability of clinical interventions for genetic disorders. Genetics in Medicine 2019 Apr 26. PMCID: PMC6815237
Webber EM, Hunter JE, Biesecker LG, Buchanan AH, Clarke EV, Currey E, Dagan‐Rosenfeld O, Lee K, Lindor NM, Martin CL, Milosavljevic A, Mittendorf KF, Muessig KR, O'Daniel JM, Patel RY, Ramos EM, Rego S, Slavotinek AM, Sobriera NM, Weaver MA, Williams MS, Evans JP, Goddard KAB, on behalf of the ClinGen Resource. Evidence-based assessments of clinical actionability in the context of secondary findings: Updates from ClinGen's Actionability Working Group. Human Mutation 2018;39:1677–1685. PMCID: PMC6211797
Mittendorf KF, Paquin RS, Lewis MA, Zulkiewicz BA, Lee K, Nyongesa DB, Leo MC, Berg JS, Williams MS, Goddard, KAB. Clinician Versus General Population Perceptions of the Nature of Clinical Interventions for Delaying or Preventing Outcomes Related to Inherited Conditions. (Poster Presentation). ACMG; April 10-13, 2018; Bethesda, MA.
My Cancer Genome
Under Drs. Mia Levy, Travis Osterman, and Christine Micheel, I received training in bioinformatics applications in somatic cancer genomics and contributed to enhancing data models supporting structured curation of biomarker-driven assertions that power the public-facing resource My Cancer Genome (MCG), a website that is viewed 8,000 times a week by individuals in 211 countries and territories around the world. Our data model was used in partnership with GenomeOncology’s Workbench services, where it powered the generation of over 40,000 interpretative reports for 31 academic medical centers and commercial labs licensing these services. This content was made available through APIs for use in electronic health records as well as in generation of molecular pathology interpretive reports. I also worked with an expert interdisciplinary team to develop a clinical molecular oncology consult service to increase access to personalized medicine for patients seeing Vanderbilt-affiliated community oncologists. Finally, I contributed to two Moonshot to Cure Cancer working group documents. I have now returned to this team to collaborate on the development of informatics- and knowledge-driven clinical decision support at VUMC.
Selected publications and presentations:
Holt ME, Mittendorf KF, LeNoue-Newton M, Jain NM, Anderson I, Lovly CM, Osterman T, Micheel CM, Levy MA. My Cancer Genome: coevolution of precision oncology and a molecular oncology knowledgebase. JCO Clinical Cancer Informatics. 2021. 5:995-1004. DOI: 10.1200/CCI.21.00084
Neha J, Mittendorf KF, Holt M, Lenoue-Newton M, Maurer I, Miller C, Stachowiak M, Botyrius M, Cole J, Micheel C, Levy M. The My Cancer genome clinical trial data model and trial curation workflow. J Am Med Inform Assoc. 2020. 27:1057-1066. PMCID: PMC7647323
Levy M, Micheel C, Jain N, Mittendorf K. Assessment of actionability of cancer genomic testing panels based on a structured clinical trial knowledge base. Journal of Clinical Oncology 35, no. 15_suppl (May 20 2017) 6533-6533. Published online May 30, 2017. DOI: 10.1200/JCO.2017.35.15_suppl.6533
Levy M, Osterman T, Jain Neha, Mittendorf K, Micheel, C. Utility of adding clinical data to a molecular results portal for improving clinical trial prescreening efficiency. Journal of Clinical Oncology 35, no. 15_suppl. [Epub 2017 May 30] DOI: 10.1200/JCO.2017.35.15_suppl.e18182
Vaccine Safety Datalink
I have participated in the CDC-funded Vaccine Safety Datalink (VSD) project at the Research Associate III and co-investigator levels at the Kaiser Permanent Northwest site. I was actively involved in workgroups and proposal development that addressed safety outcomes associated with adolescent vaccination. In this role, I have contributed to the development of two VSD proposals and designed the data model for evaluation of outcomes for one of these funded proposals. I also contributed to work on the association of primary ovarian insufficiency/infertility with adolescent vaccinations, published in Pediatrics.
Publications
Naleway AL, Mittendorf KF, Irving SA, Henninger ML, Crane B, Smith N, Daley MF, Gee J. Primary ovarian insufficiency and adolescent vaccination. Pediatrics. 2018 Sep;142(3). pii: e20180943. doi: 10.1542/peds.2018-0943. [Epub 2018 Aug 21]. PMCID: PMC6719304
Groom HC, Brooks NB, Weintraub ES, Slaughter MT, Mittendorf KF, Naleway AL. Incidence of Adolescent Syncope and Related Injuries Following Vaccination and Routine Venipuncture. Journal of Adolescent Health. 2024 Apr;74(4):696-702. doi: 10.1016/j.jadohealth.2023.11.005. Epub 2023 Dec 9. PMID: 38069938; PMCID: PMC10960660. Featured in issue editorial.