Mentorship Meetings

Anish S. Shah, MD, MS

University of Utah

University of Illinois Chicago

Emory University

04/24/24 | DD

Agenda

1. Active projects
   1. AFL/FH paper accepted to JAHA
   2. AFEQT/SDOH manuscript, tracking to Annals
   3. ECG/genetics project, UK BioBank proposal
   4. AF-CTRS, piloting
2. K23 plans and intergration
3. Mentorship timeline
   1. Current/future mentees
   2. Research mentership

Active projects

AFEQT/SDOH manuscript

• Draft to co-authors by end of April
• Submission to Annals end of May
• Abstract/manuscript for AHA Scientific Sessions + Early Career Award application

ECG/genetics

• UK BioBank proposal being written (Emelia and Seung Hoan to review to help with sponsorship)

Should this project be dropped?

AF-CTRS

Priority? Timeline?

K23

1. Identify scar patterns and their evolution in AF phenotypes
   • ECG-based phenotypes (UIC Multi-Ethnic AF Registry and MRI data in Utah registry)
   • Genetic variant influence on atrial myopathy patterns, incorporate the inverse problem for ECGi
   • Validation in larger cohorts (ARIC, MESA, UK BioBank)
2. Measure and quantify the electrical properties of different scar-based phenotypes EPS-modulation of dromotropic properties
   • EP-lab based
   • Obtain parameters that affect dromotropic properties (e.g. autonomic modulation)
3. Model the effect of scar on atrial arrhythmia prediction (AT, AF, AFL)
   • Computational aim with Ravi

Mentorship timeline

Shashank Sadhu

• James Scholar award on ECG/genetics
• Incorporation into ECG/genetic project

Darren Seaney

• T35 award on AF/diabetes
• Summer project
• Incorporation into ECG/genetic project

Future lab resources

• VCF pipeline for annotations
• Pipeline for CCTS/ECG dataset creations

• Consider continued mentorship with both during EP fellowship?
• 1:1 meetings q2wk with mentees?
• UIC account sponsorship (EPIC? CCTS? Research account?)
• Cluster for data storage + research?

04/23/24 | AJS

Agenda

03/13/24 | DD

Agenda

1. First-author products
   1. AFL/FH paper submission (R3) to JAHA
   2. AFEQT/SDOH paper (drafting), track to JACC EP
   3. HRV/CVD paper re-submission (R1) to JAMA Cardiology
   4. GEH/DM (CARRS) paper (drafting), deciding on submission location
   5. ECG/TTN presentation (3/27 to AFGen fellows)
   6. MUSE/LSPro annotation software (pending publication)
2. Co-author products
   1. Hemodynamic Reactivity (JAMA Open)
   2. HRV and MSIMI (JAHA)
   3. HRV and CHF (JACC HF)
   4. VUS/P/LP and AF/DCM (draft by Hill)
3. Research network
   1. AHA Epidemiology (March 2024) - meeting with Jared Magnani and Alvaro Alonso
4. K23 planning
   1. Review K23 aims with Emelia Benjamin and Ravi Ranjan

02/21/24 | AA

Agenda

1. General professional updates
   1. AFGen & working with EJB
   2. EP at Utah
2. AFEQT and SDOH work
   1. Table 1
   2. General findings
3. AFGen proposal
4. ARIC proposal
   1. Rate/rhythm by genotype

02/20/24 | AJS

Agenda

1. CARRS data
   1. Table 3 = <10% data loss in modeling
   2. Table 2a/2b = LSM is \(\propto\) factorial AOV
2. K23 update
   1. Specific aims
   2. Preliminary data
3. ECG/TTN update
   1. Variant annotation
   2. ECG/ML neural network learning

02/13/24 | AJS

Agenda

1. CARRS1 analysis
2. ECG/TTN variant analysis
3. K23 application

01/29/24 | AJS

Agenda

1. HRV/CVD paper submission
2. CARRS models
   1. Decision on CARRS2?
3. AFEQT/SDOH paper
4. ECG/TTN paper
   1. Variant annotation pipeline
   2. ECG/ML layer learning
5. Next steps
   1. GWAS/reactivity
   2. DSP/reactivity

01/24/24 | DD

Agenda

1. AFL/FH revisions
   1. Reviewer response by Friday
   2. Draft of paper with revisions by next Monday
2. AFEQT/SDOH
   1. JBVA data
   2. Worsening AFEQT ~ Black Race + Divorced + High Deprivation (NDI) + AAD3
3. Genetics and ANS dz (arrhythmia surrogates)

01/23/24 | AJS

Agenda

1. HRV/CVD paper
   1. Submission to JAMA Cardiology
   2. Co-authors revisions needed?
2. CARRS project = GEH/DM
   1. ECG data ~ n ~= 5000 overall
   2. CARRS1 ~ n ~= 1500 with ECG data (out of n ~= 16,000)
   3. Results preview under R/results.html (pending)
3. ECG/TTN project
4. Next projects
   1. MS/AF
   2. Harmonic/autoregressive predictors of stress reactivity
   3. GWAS for HRV

01/19/24 | SHC

Agenda

1. Project overview
2. Computational methods
3. Proposal
4. Statistical concerns

Project overview

Can we predict the effect of TTN variants on atrial myopathy (and subsequent atrial fibrillation) using machine learning algorithms trained on ECG data?

• Cohort of “early” (age < 65 onset) of AF with mix of WES and high-density candidate gene sequencing
• n=298 individuals overall and n=36 individuals with LP/P variants in TTN (using SIFT/PolyPhen scores)
• n=7953 unique ECGs that were thought to be in sinus rhythm from 2010 to 2023

Computational methods

• Digital signal processing to extract ECG data (P wave, R wave, T wave) while in sinus
• Deep learning approach for feature extraction and classification
• Novelty is in part using 12-lead approach and recurrent data (e.g. multiple ECG over time)
• End-to-end classification layer to combine pre-processing, filtering, clean-up

Proposal

• Assumes multi-hit hypothesis on effect of TTN variants on atrial myopathy (e.g. effect on strain, fibrosis, etc)
• Assumes fibrosis/increase duration of AF episodes (unlikely to lead to initiation but contributes to maintenance)
• Identification of TTN-VUS in AF patients that may play a role in atrial remodeling
• Using VEP to evaluate 1) location of variants in bands of sarcomere, 2) focus on missense proteins (?) in addition to truncating and/or LOF variants
• Assess impact of variants on ECG

Statistical concerns

• Power is limited based on number of individuals with TTN variants
• Multiple machine learning approaches needed that are “explainable” in terms of effect on ECG signal
• Validation of dataset
• Correct “ground truth” of pathology of TTN variants is difficult to confirm

01/10/24 | DD

Agenda

1. Mentorship updates:
   1. EP fellowship - need to reach out to Ranjan
   2. AFGen Fellowship - RIP meeting in March, Benjamin = K23 remote mentor, Choi = assistance with statistical genetics
   3. Sean/Shashank - ML/ECG pipeline, James Scholar and T35 application
2. ECG/TTN:
   1. Training models on TTN LP/P variants
   2. Categorizing variants based on location
3. Phenotyping pAF:
   1. Assessing differential trajectories of AF based on quantified ECG/TTE
   2. TTN variants and trajectory of impact on AF ECG/TTE trajectory (compared to controls)
4. Stress reactivity & genetics - considering GWAS approach

01/09/24 | AJS

1. Questions on HRV/CVD mortality paper
2. CARRS data analysis pipeline
   1. CARRS 1 dataset
   2. CARRS 2 dataset needs XML data re-analysis
3. Future projects - stress reactivity and genetics?
   1. GWAS approach
   2. PRS approach

01/02/24 | AJS

Agenda

1. Update on HRV/CVD JAMA Cardiology revisions
2. CARRS project updates
   1. Define timeline
   2. Assign roles
3. Other projects
   1. AFEQT/SDOH
   2. ECG/TTN

12/13/23 | DD

Agenda

1. AFGen fellowship updates
2. K23 mentorship committee expansion
3. Decision on abstract submissions
   1. AFL/FH should be submitted
   2. ECG/TTN should be held until manuscript is complete
4. AFEQT manuscript strategy
   1. Co-authors for expanded feedback
   2. Increasing impact of paper, higher level journal
5. Timelines

12/12/23 | AJS

Agenda

1. Review of big-picture items
2. Current project update on ECG/TTN variants
   1. https://asshah4.github.io/research/t32/rip.html#/december-10-2023
3. CARRS project
   1. Repository
   2. Paper plan

12/05/23 | AJS

Agenda

• Review publication & project priorities
• If time is available, walk through CARRS repository
• Review of ECG/TTN variant project

11/21/23 | DD

Agenda

• High level updates
  • EP fellowship ranklist
  • Lab meeting structure
  • AFL/FH paper: reviewer comments, discussion thoughts
• AFEQT/SDOH
  • Missing data issues
  • Defining outcomes
• ECG/TTN variants
  • Tentative plan and timeline

AFEQT/SDOH

• n = 51 patients had missing encounter dates but had provided baseline AFEQT scores
• n = 388 unique patients from UIC currently with AFEQT at baseline and follow-up (not including JBVA data), however missing or inappropriate baseline dates (e.g. before AF registry started), approximately ~150 additional patients

Questions:

• How should we handle missingness?
• Should we call to obtain “baseline” data for recent patients? (approximately 150 additional patients from 2020 to now are available)

ECG/TTN

1. Working on ECG findings with TTN variants currently
2. Abstract focus is on trainability of ECG-based model to help differentiate ECG in pathogenic variants of TTN versus controls in patients with atrial myopathy
3. Expect to have initial analyses complete in first week of December
4. HRS abstract deadline is December 15th

11/13/23 | AJS

Agenda

• High level updates
  • EP fellowship
  • Lab meetings / collaborations (Louis Li?)
• HRV/CVD paper
• ECG/TTN variants
• Next projects?
  • DM/GEH papaer
  • Twins/cosinor and \(\psi\)
  • Stress-decrease and PRS
  • Stress-decrease and atrial arrhythmias

HRV/CVD

1. What elements of the results should we make sure to include? As MI, competing risk, etc are taken out, our points made in discussion about cause-specific nature are limited?
2. By taking out competing risks paragraph, the results section now is < 1 page (too few results was previously a concern)
3. Major findings in discussion should be both high risk and low risk groups? Dichotomous and categorical groups?

11/07/23 | MM

Potential Locations

Program	Comments
Case Western	Arruda, Waldo (no NIH-EP)
Houston Methodist	Valderrabano, out-of-match
MCW	supportive faculty, no NIH-EP
Northwestern	Arora leaving, Passman
NYU	Park, recent R01
UC San Diego	No NIH-EP, Narayan left
UChicago	Arora, Yeshwant
UIC	Darbar, McCauley
UMich	Ghanbari
UPMC	Saba
UTHouston	offered 3-year protected
U Utah	Ranjan, Steinberg
UWisconsin	Eckhardt

11/07/23 | AJS

Agenda

1. Opinions on fellowship and future career
2. Updates after meeting with RL
3. Revisions to HRV/CVD mortality paper
   1. Context of MW paper
   2. Context of JO paper
4. Consideration of adding AA on to the AFL/FH manuscript
5. Next steps

11/01/23 | DD

Agenda

1. Dr. Darbar’s current priorities & thoughts for Anish
2. Current limitations/challenges
   • September/October general cardiology boards
   • Interview burden
3. Active projects:
   • AFL/FH revisions
   • AFEQT data collection + phenotyping AF
   • ECG/TTN data cleaning
4. EP fellowship interview updates
   • Key locations
   • Ranking advice
5. Next steps

AFL/FH Remaining Questions

• Can a comparison group with isolated AF be analyzed as part of a sensitivity analysis?
• Why do we think AA population had higher risk of AFL?
• Why is there an inverse relationship between AFL and BB/CCB usage?
• Why did we combine White with Latinx?

AFEQT/SDOH/Phenotying AF

Quantified approaches to measure AF, progression, treatment:

• AFEQT on multiple visits consider including this in EPIC in all AF consult/clinic notes
• SDOH will include…
  • Sociodemographics, including education/income for some participants
  • Neighborhood deprivation indices
  • Payment method / insurance types
• Progression composed of repeat ECG in sinus or AF (including heart rate), medication intensification (now have dosages as well at each visit from CCTS), and intervention (DCCV, PVI)

My next 3-4 weeks will be focused on cleaning and processing this data from CCTS

ECG/TTN

• Shashank and Darren have joined on this project, will teach them DSP and ML approaches (taking them to the EP lab)
• Need to train an algorithm to identify chambers (P/QRS/T waves) with improved accuracy over current mechanisms
• Have to generate training and test data sets

Next week will be dedicated to ML/signal processing efforts

PRS/AF

• Contacted NW about the SNP data
  • There are 165 SNPs that they utilized
  • Effect sizes for each SNP were also given (from their cohort)
• Actively completing power calculations now
  • Have to decide how to use our VCF data from WES and from GWAS data
  • Can we leverage both the dense array data and WES to match the 165 SNPs?
• Will meet with Chen to discuss further

EP Fellowships

• Case Western
• Houston Methodist (Valderrábano)
• Northwestern
• NYU
• UCSD
• U. Chicago
• U. Michigan
• UPMC (Pittsburgh)
• U. Wisconsin

09/08/23 | DD

EP Interviews (notable figures)

• U. Wisconsin (Eckhardt)
• UPMC
• U. Chicago
• NW (Arora, Passman)
• U. Arizona (Tung)
• Boston (Emilia Benjamin)
• Case Western
• UC San Diego
• University of Utah (Freedman)
• U. Mich

EP Potential Programs

• Vanderbilt (Stevenson, PD = Arvindh Kanagasundram)
• UCLA (Shivkumar, PD = Noel Boyle)
• UCSF (Marcus, PD = Joshua Moss)
• Mayo (PD = Sam Asirvatham)
• JHU (PD = Ronald Berger)

Research Priorities

1. Genetic variant annotation
2. ECG & machine learning approaches
3. EP lab (research & clinical)
4. AF clinical phenotyping + SDOH

08/16/23 | DD

AFL/FH

• Decision on journal submission
  • Circulation EP or Journal of American Heart Association
  • American Heart Journal
  • International Journal of Cardiology
  • British Medical Journal Heart
• Revisions to discussions and analyses

(a)

(b)

Figure 1: Examples of pedigrees in high prevalence families.

Phenotype AF

• Progress: on schedule, with support from committee, with goal is to complete ECG-based features by end of August
• Machine learning: Generating diagnoses/beat-segmentation through PhysioNet and applying to our local data
• Phenotype data: Starting to collect clinical variables to assess status and burden
• eMERGE: pending power calculation with Chen, no additional proposal needed with eMERGE, but will need to put together manuscript concept sheet

Can we phenotype a patient with AF based on ECG alone?

Additional projects & training

• PhysioNet: AF labeling/detection for ML algorithms
• SDOH: Involving Alvaro & Amit from epidemiology perspective, evaluating multiple aspects for AF progression/intensification. Current areas of focus include neighborhood-based resources, insurance, and health literacy
• Language/AF: Expanding work from grant-student on this
• Computational genetics: Finished coursework on this in July, learning how to filter/annotate VCF
• EP fellowship: 4-5 interviews at mid-range programs, pending “reach” programs (e.g. Vanderbilt, Northwestern, JHU)

08/03/23 | DD

Outline

1. EP fellowship and timeline
2. Prioritizing projects
3. SDOH and arrhythmia outcomes
4. Phenotyping AF

EP fellowship

Which locations should I prioritize for in terms of ideal training? When should I ask for personal recommendations?

• Northwestern
• Vanderbilt
• UPenn
• Johns Hopkins
• Stanford
• UCLA

Prioritizing projects

1. AFL/FH revisions: plan for being done by the weekend
2. Phenotyping AF: complex project, currently working on ECG analysis with TTN genotypes
3. Mentees: currently have requests from x2 medical students, x2 residents, and am currently working with x1 student on SDOH/language barriers in AF
4. Committee: Currently everyone is reviewing my phenotyping proposal - most excitement is for ECG identification of genetic variants
5. eMERGE: need help with original and revised proposal - okay to ask VUMC staff about supporting/example documents?

SDOH/Arrhythmia

• Urban Health Disparities grant to evaluate language barriers as mediator of AF care intensification (e.g. medications, cardioversion, referrals)
• Plans are to expand for ALL arrhythmias and EP referrals
• Mediation analysis for neighborhood
• Roughly n=3500 English speakers (control) and n=500 Spanish speakers (exposure)

Phenotyping AF

• Exploration of identifying ECG features with TTN mutations. Key mutations to include… TTN, FLNC, MYBPC3, SCN5A, LMNA
• Of these WES (n=305), there are (n=56) with TTN variants (different mutations however)
• Over 1k ECGs associated with this cohort
• Approach: using ECG-based tensors, will trial multiple neural networks and see efficacy, but will require larger validation cohort
• Requires cluster computing → likely more cost-effective than purchasing our own hardware

07/19/23 | AJS, DD, AA

Outline

1. Long-term timeline
2. Research foci
3. Manuscript/project proposals
4. Parallel clinical training
5. 2023-2024 timeline

gantt
    title Timeline
    dateFormat MM-YYYY
    section Clinical
        General CV fellowship       :07-2020, 06-2024
        EP fellowship                       :07-2024, 06-2026
        Junior faculty                  :07-2026, 06-2028
    section Academics
        TL1 fellowship                  :07-2019, 06-2020
        T32/F32 fellowship          :07-2022, 06-2024
        K23 application                 :03-2026, 09-2026

K23 application in ~2.5 years, with following variables:

1. EP fellowship + faculty position institution
2. Department and divisional support
3. Changes / discoveries in research field/space

Areas of interest

• Stress: arrhythmic component in abnormal stress reactions (autonomic-mediated)
• pAF: trajectory, phenotypes and subphenotypes
• ECG: beat segmentation, morphology, machine learning approaches
• EPS: novel approaches to electrogram classification, 3D mapping analysis (conduction velocity, scar mapping)
• Genetic variants: GWAS for burden/progression, ECG-markers, cardiovagal receptors

Research questions

Do certain phenotypes (EMR, ECG, TTE, VUS) that have different rate of progression of AF?

Are there genetic variants in cardiovagal receptors that increase the risk of autonomic dysfunction to stress?

Are markers of abnormal stress reactivity predictive of future arrhythmic events?

Specific proposals

1. ECG/PAF = ECG classification of progression/trajectory of pAF (terminal P wave forces in NSR, coarseness of F waves in AF) builds on DSP and EPI skills
2. GWAS/PAF = GWAS with ECG markers of pAF trajectory in AF Registry, validate in ARIC (?) to develop skills in computational genetics
3. GWAS/ANS = Exploratory association of genetic variants with “maladaptive ANS response” from Emory MIMS/MIPS dataset as fundamental for ANS-mediated arrhythmias
4. COSINOR = Circadian rhythm (multiphase) of HRV/ECG markers with neuropsychological stress (e.g. depression/anxiety) builds on previous software publications
5. ECG/MSNA = Correlation between SNA and ECG parameters (PRD, GEH, P-wave morphology) strengthens DSP and clinical relevance
6. REVIEW = Review of trigger-mediated versus scar-mediated pAF

Specific Aims

The following aims will be supported by preliminary and published data generated from this research year, combining DSP, epidemiology, and genetic approaches. Expect some fluctuation as the research evolves.

1. Identify ECG-based phenotypes of pAF
2. Determine cardiovagal biomarker patterns of sympathetically-mediated pAF during EPS
3. Evaluate role of genetic variants in pAF progression

Parallel clinical training

1. Studying for general cardiology boards
2. 50% conference attendance (supplemented by research meetings)
3. 1/2 clinic per week
4. 1/2 day of EP lab per week
5. 2-3 weeks of clinical service in spring prior to fellowship

Timetable

gantt
    dateFormat MM-YYYY
    axisFormat %b-%Y
    section Clinical
        Echo boards                                 :milestone, 07-2023, 0d
        General boards                          :milestone, 10-2023, 0d
        EP fellowship match                 :milestone, 11-2023, 0d 
        EP fellowship starts                :milestone, 07-2024, 0d
    section Atrial Fibrillation
        AFGEN application                       :milestone, 07-2023, 0d
        AFL/FH manuscript                       :milestone, 08-2023, 0d
        Chart review for AF                 :08-2023, 11-2023
        NLP for AF burden                       :09-2023, 01-2024
        Clinical trajectories               :10-2023, 04-2024
        Phenotype AF manuscript         :milestone, 05-2024, 0d
    section Stress Reactivity
        Stress/AF preliminary           :07-2023, 09-2023 
        Stress/AF drafting                  :09-2023, 11-2023
        Stress/AF manuscript                :milestone, 01-2024, 0d
        GWAS/ANS proposal                       :08-2023, 09-2023
        GWAS/ANS data                           :09-2023, 01-2024
        GWAS/ANS drafting                       :12-2024, 03-2023
        GWAS/ANS manuscript                 :milestone, 03-2024, 0d
    section Computational
        Digitization of ECG                 :06-2023, 08-2023
        ECG parameters generation       :07-2023, 10-2023
        EPS scar-burden                         :09-2023, 11-2023
        ECG/AF manuscript                       :milestone, 11-2023, 0d
        

Milestone feedback

• Establish milestones that are appropriate for next 12 months of dedicated research
• Frequency and purpose of mentorship meetings
• Role of clinical training in this research year (e.g. EP lab, consults)
• Importance of research skill development
• Mininum number of publications

06/20/23 | AJS

HRV/CVD revisions

• Better define baseline ANS dysfunction and stress-induced ANS inflexibility
• Consider moving HF HRV from supplement to main tables as they are similar in significance
• Methods confusion about sensitivity analyses

06/16/23 | AB

Premises

1. Atrial fibrillation is a macro-reentry rhythm predominantly initiated by triggered electrical activity
2. Triggered activity can come from maladaptive autonomic tone ± focal scar ± tissue heterogeneity
3. Each triggered activation may not lead to a sustained or measurable episode of AF
4. Duration of AF is mediated by amount of fibrosis or scar available to sustain arrhythmia

Hypothesis: Sympathetic activity drives triggered atrial activity that leads to and sustains episodes of AF in those with structurally normal atria

Available Data

• ~145k patients with CVD diagnoses (overly sensitive)
• ~500k digital ECGs from 2010 to 2023
• Clinical notes from providers + support staff
• Procedure reports for cardiac studies (catheterization, ablation, device interrogation, echocardiography, cardioversion)

Goals: Identify patients at onset of paroxysmal AF, and identify time point when burden of AF increases or transitions to persistent AF

Persistent AF is poorly defined as >7 days of AF duration

Approach

1. Identify diagnosis of atrial fibrillation: ICD codes, ECG diagnoses
2. Confirm diagnoses: EP notes, interventions, hospitalizations, procedures
3. Quantify burden: documentation in notes, interventions, duration of episodes, change on repeat ECGs
4. Cross-sectional parameters:
   1. Clinical = weight, comorbidites
   2. Electrical = F-wave voltage, P-wave morphology, scar-burden on EPS
   3. Echocardiography = LA size/volume, LV size, LV function, diastology
   4. Treatment = PVI, ECV, antiarrhythmic drugs, BB
5. Evaluate trajectory: characteristics that associate with rapidity of diagnosis change

NLP approach

• Filter clinical documentation to provider-only information
• NLP approach is to use bidirectional transformer approach (BERT)
• Add fine-tuning layer with named entities for AF (symptoms/burden/recurrence)
• Compare with AF Registry as training data, and EMR as test data

06/13/23 | AJS

Wrap up of HRV/CVD paper

• Vaccarino feedback
• Quyyumi feedback
• Lampert feedback

Zombie projects

• Circadian profile of depression, PTSD, anxiety
• Relationship of DYX with CAD burden (F32)
• CARRS ECG parameters and diabetes
• ECG-based age prediction and telomere length

Aims-focused project

1. ↑ sympathetic activity + ↓ cardiovagal activity is malignant/maladaptive
2. Etiology of vagolysis-prone individuals is likely both polygenic + acquired
3. Phenotypes may show: abnormal variants in HRV-associated and cardiovagal-associated genes, abnormal ECG-markers or patterns (e.g. GEH), clinical covariates

Question: Are there common/rare variants associated with these patients with higher risk (or is this a truly acquired disease)?

06/09/23 | DD

End-of-year meeting

1. Research training and productivity
2. Career planning (EP fellowship)
3. Upcoming milestones and research plans
4. Formative feedback

Research training

Completed training:

• Signal processing coursework through MIT OpenCourseWare in Aug 2022 to Dec 2022
• NLP coursework through Stanford Online in Feb 2023 to May 2023

In process:

• Computational Genomics series through Harvard Bioinformatics May 2023 to present

Research products

First Author:

• AF/AFL Manuscript = JAMA Cardiology JACC EP
• HRV/CVD Manuscript = Circulation EHJ
• LSPro/WFDB Abstract = HRx 2023
• HRV/CVD Abstract = ACC 2023
• LSPro/WFDB Software = CRAN ¹

Middle Author:

• Abstracts: ACC 2023 x 3, HRS 2023 x 1, AHA 2023 x 2
• Manuscripts:
  • HRV & Incident Heart Failure WIP
  • AF Recurrence & Race WIP

Article Reviewer:

• IJC x 1, AHJ x 1, Heart x 1

1. Peer-reviewed software repository

Career planning

Applying to EP fellowship:

1. LOR: (1) Darbar/Amit (combined letter), (2) Yeshwant (clinical), (3) Avitall (clinical), (4) Auseon (program director)
2. Locations: NW, UIC, UofC, Loyola, UPenn, Vermont (Spektor), UCLA (Shivkumar), VCU (Ellenbogen), JHU

Fellowship location will likely determine long-term K23 application

Paper proposal

Project to evaluate ECG-correlates of genetic variants with paroxysmal AF directly, similar to Verweij et al. (2020) approach. Basis of AFGEN fellowship application.

Exposure: common and rare genetic variants

Outcome:

1. ECG parameters during sinus (300 individuals x 4-10 ECGs x avg 10 beats = 20k to 35k ECG volume)
2. dominant frequency and coarseness of F waves during AF

Hypothesis: Genetic variants in genes associated with paroxysmal AF, including novel variants, will have significant association with P wave abnormalities

Genetic variants =

• cardiomyopathy genes (TTN, NPPA, TBX5, MYL4, TTN, LMNA)
• arrhythmia genes (KCNQ1, SCN5A, SCN10A, EXT1)
• cardiovagal genes (NPY2R, GALR1)

Completed milestones

• Routine/weekly 1:1 mentorship meetings with Darbar/Amit
• Clinical rotation obligations (Echo, consults, etc)
• Twice weekly clinic
• Educational coursework in DSP and NLP
• K23 specific aims draft (1-2 years ahead of schedule)
• Identified core research committee (Darbar, Shah, Alvaro) and advisors (Boyd, Lampert)
• Data collection of ECG, CV-procedures, EP study data for patients in AF registry (superset)
• Introduction to genetic analysis software (PLINK, BioConductor, Samtools, BCFTools)
• First-author manuscript submission x2

Upcoming milestones

• Committee meeting = 06/2023 (confirm milestones with Darbar/Amit/Auseon)
• Echocardiography boards = July 2023
• AFGEN application = July 2023
• Attend 3 conferences this year (ACC, HRS, and R::conf), with associated abstract submissions = 2023 to 2024
• Complete ECG analyses by August 2023, genetic analyses by September 2023, and submit manuscript on ECG/genetic in AF by October 2023
• General cardiology boards = October 2023
• Match into (top-tier) EP fellowship = December 2023

04/28/23 | DD

• Preliminary data for AIMS
  • NLP/phenotyping needs coursework and mentorship with Boyd
  • ML/DSP needs significant engineering/programming time
• AF phenotyping to identify those with heavier percentage of “triggers” versus those with higher burden of “substrate”
• EGM assessment of AF, evaluating scar-burden, conduction velocity, with underlying theory of “wavelet re-entry” as a boundary problem

• AFL/FH manuscript to JAMA Cardiology
• HRV/CVD manuscript to co-authors
• HRV/CVD manuscript to European Heart Journal
• AF/CA outcomes statistical review
• Analysis of CCTS data pull
• MUSE ECG data in XML format
• IRB for EP lab research

04/11/23 | AJS

• Submitting paper to JAMA cardiology on AFL/FH
• Completed revised concordance and AUC on HRV/CVD paper
• WFDB for EGM, initialized I/O
• NLP/ontology planning started

• Analyses for non-first author paper on AF/ablation outcomes
• HRV/CVD paper thoughts on concordance and NRI
• Specific aims review
• Final committee decisions: Darbar, Shah, Alvaro, Boyd, Lampert?
• Ontology of “stress”?

03/31/23 | DD

HRV/CVD: Finalizing paper draft, to submit EHJ

CAR: Adjudicated ECGs, designed data collection, pending analyses

AFL/FH:

• Paper draft, second round
• Identify all authors
• Finalize JAMA cardiology submission

Phenotyping AF:

• eMERGE PRS and PhenotypeKB
• AWS HPC access
• CCTS DRA for data pull
• Ablation database

K23:

• Draft of aims
• Feedback on aims
• Prelim data on AF phenotypes
• Prelim data in the EP lab

Mentorship:

• Identified majority of mentorship committee = Darbar, Shah, Alonso, Boyd (advisor), Lampert
• Meeting with McCauley (3/3)
• Meeting with Boyd (3/16)
• Genetics mentor, local or at another institution? → will need more genetics training

Training:

• DSP training → coursework
• NLP training → coursework
• Computational training in genetics

Career:

• Delineate timeline and strategy for EP fellowship
• Identify potential institutions and letters of support/introduction

03/03/23 | DD

HRV/CVD:

• Non-linearity evaluated, analyses being repeated
• Identify appropriate level journal (EHJ)
• Presenting to Amit MS-EP group

AFL/FH:

• Paper draft has been drafted
• Genetic analyses
• Identify all authors
• Send in JAMA Cardiology format

CAR:

• Adjudicated ECG and designed data collection pattern
• Role and next steps
• REDCap inconsistencies noted → responsibility?

K23:

• Draft of aims
• Feedback on aims
• Which areas require prelim data?

Mentorship:

• Identified majority of mentorship committee = Darbar, Shah, Alonso, McCauley, Boyd, Lampert
• Meeting with McCauley (3/3)
• Meeting with Boyd (3/16)
• Genetics mentor, local or at another institution?

Training:

• Computational training in signal processing
• Computational training in genetics

Career:

• Delineate timeline and strategy for EP fellowship

References

::: {#refs}

Verweij, Niek, Jan-Walter Benjamins, Michael P. Morley, Yordi J. Van De Vegte, Alexander Teumer, Teresa Trenkwalder, Wibke Reinhard, Thomas P. Cappola, and Pim Van Der Harst. 2020. “The Genetic Makeup of the Electrocardiogram.” Cell Systems 11 (3): 229–238.e5. https://doi.org/10.1016/j.cels.2020.08.005.