Click here to view the agenda of the Second Machine Learning Vulnerable Patient Symposium

symposiumlogos2016

The First Machine Learning Vulnerable Patient Symposium

A Focus Group Meeting on Developing an Artificial Intelligence-based Forecast System for Prediction of Heart Attacks within 12 Months

A Satellite Event in Conjunction with
2016 Annual Scientific Sessions of American Heart Association

american_heart_association_logo

Le Meridien Hotel, New Orleans, LA
November 14, 7-10 PM

This event is open to public.

Agenda

Welcome

 


Morteza Naghavi, M.D.
Program Director, Founder of SHAPE and Executive Chairman of the SHAPE Task Force
Download Slides
 


 

 

 

Opening Remarks

 


Valentin Fuster, M.D., Ph.D.
Professor of Medicine and Physician-in-Chief, Mount Sinai Hospital and Icahn School of Medicine
Download Slides
 


 

 

 

Moderators:

aha2016-maron-saha2016-narula-s
David Maron, M.D.
Clinical Professor and Director, Preventive Cardiology, Stanford University
Jagat Narula M.D., Ph.D.
Chief of Cardiology, Mount Sinai West & St. Luke’s Hospitals Associate,
Dean, Arnhold Institute for Global Health at Mount Sinai Icahn School of
Medicine

Featured Speaker:

 


Ioannis Kakadiaris, Ph.D.
Professor of Computer Science and Biomedical Engineering,
Director of Machine Learning Laboratory
University of Houston
Download Slides
 


Topic: What is Machine Learning and How Can It Shape the Future of Healthcare?

Invited Online Presentations:
Two Examples of Machine Learning Studies in CVD Risk Assessment (10 minutes each)

(1) CVD risk prediction using support vector machine based on Australian Blue Mountains Eye study database

kumararjunan

 

Dinesh Kumar, Ph.D.                                                                                          Sridhar Arjunan, Ph.D.
Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, Melbourne, Australia
 

(2) Machine learning improves image analysis, diagnosis and prognosis in cardiac imaging

 

Download Slides
 


 


Piotr Slomka, Ph.D.
Chief Scientist, Artificial Intelligence in Medicine Program, Department of Imaging Cedars-Sinai Medical Center, Professor, UCLA School of Medicine, Los Angeles, CA
Download Slides
 


Moderated Discussions on the Vulnerable Patient Project
Machine Learning for Prediction of  Near-Term CHD Events
All investigators will be asked to give a very brief introduction of their study and how it can fit in

“Machine Learning Vulnerable Patient Project”.

Project Summary

Background: Imagine instead of the existing daily weather forecasts and hurricane alerts we were told the probability of a storm within the next 10 years! This is how heart attacks are predicted today. We teach our physicians to calculate the 10-year probability of a heart attack and sudden cardiac death based on their patients’ risk factors. Long term predictions do not trigger immediate preventive actions. Although some people develop warning symptoms, half of men and two-thirds of women who die suddenly of coronary heart disease (CHD) have no previous symptoms. Imagine if we could alert people months, weeks, or even days before a heart attack and trigger immediate preventive actions.

The Idea: Use machine learning to create new algorithms to detect who will experience a CHD event  within a year (The Vulnerable Patient). Algorithms will be based on banked biospecimen and information collected days up to 12 months prior to the event. We will utilize existing cohorts such as MESA, Heinz Nixdorf Recall Study, Framingham Heart Study, BioImage Study and the Dallas Heart Study.  External validation to test for discrimination and calibration will be conducted using other longitudinal observational studies that provide adjudicated cardiovascular event information such as the MiHeart, JHS, DANRISK and ROBINSCA. Additionally, we will use machine learning to characterize individuals who, despite high conventional risk, have lived over 80 years with no CHD events (The Invulnerable ). We expect to discover new targets for drug and possibly vaccine development. We will make the algorithms available as an open source tool to collect additional data over time and increase its predictive value.
 
Organizers:
SHAPE as the originating and organizing center for the entire project, recruiting new studies and biobanks, conducting workshops with researchers from each study, fundraising, creating an open source platform community for future enhancement and collaborations.
Stanford as the coordinating center for collecting data and samples, and basic science labs.
Mount Sinai as the data review and publication center.
Machine Learning Lab to be decided, either Google, Apple, IBM, Facebook, Amazon or wherever we find a strong industry partner or sponsor.

cohort-studies

Introductory Presentation to Final Topic:

Challenges in Predicting Acute Coronary Events

picture1

Armin A. Zadeh, M.D. Ph.D., M.P.H.

Director, Cardiac Computed Tomography, Associate Professor of Medicine, Johns Hopkins University Division of Cardiology, The Johns Hopkins Hospital

 

Download Slides
 


Final Topic: A Billion Dollar Question to All Participants

Moderators:

shape-meeting-2016-thumb-moderators-jpg

Imagine the new machine learning Vulnerable Patient detection algorithm (Heart Attack Forecaster) is created and validated. If studies confirm that the new artificial intelligence based Heart Attack Forecaster is able to detect the Vulnerable Patient with 50% or more certainty, (in other words, 1 out of 2 patients classified as Vulnerable Patient goes to have a CVD event in the following 12 months), then the big questions we need to answer are:

A)    What preventive actions would you take if your asymptomatic patient tested positive as a Vulnerable Patient?

B)    What preventive actions would you take if the patient was you? (This question is meant to circumvent regulatory and financial limitations that may apply to your patients but may not hold you back).

Dr. Morteza Naghavi presents the vulnerable patient video that shows an individual’s risk of a coronary event is defined by three components of plaque, blood, and myocardial vulnerability. Dr. Jagat Narula completely agrees with the presentation and conclusions made by Dr Naghavi.
 
 
Dr. Morteza Naghavi presents a rerun of Dr. Eugene Braunwald’s presentation on coronary risk management at SHAPE’s Vulnerable Patient symposium in 2004. Dr. Braunwald points to pre-emptive interventions such as bypass surgery and stents if an individual’s 1yr risk exceeds 25%. and Dr. Jagat Narula discusses the points.
 
 
Dr. Daniel Berman answers the billion dollar question “what would you do when you find the vulnerable patient?”.
 
 
Dr. Ahmed Tawakol discusses the billion dollar question (what would you do when you find the vulnerable patient?) and the challenge that Dr. Fuster outlined namely educating the patient to adhere to heart healthy life style.
Dr. Zahi Fayad discusses his views on the billion dollar question (what would you do when you find the vulnerable patient?) and that new governmental policies at the societal level needs to be put in place.
Dr. Morteza Naghavi poses a question to co-chair Dr. David Maron regarding patients on statins getting complacent and gaining weight, and possibly mislead about their risk.
Dr. Amit Khera discusses the billion dollar question (what would you do when you find the vulnerable patient?) to and that once an asymptomatic CVD patient is told you are going to die in the next 12 months if you don’t do anything about lowering your risk, it will likely trigger actions. Dr. David Maron adds his input.
Dr. Morteza Naghavi discusses some of the answers given to the billion dollar question (what would you do when you find the vulnerable patient?) by cardiologist colleagues who were asked what would do you if you were the vulnerable patient?
Dr. Michael Blaha discusses his views that current primary prevention policies have pushed the incidence of coronary events down. Dr. Morteza Naghavi discusses the decline is not rapid enough and with the slope eradication of heart attacks may take 100 years. Dr. Jagat Narula and Dr. Ahmed Tawakol weigh in specially regarding patient compliance to treatment. Dr. Morteza Naghavi draws the analogy between the future approach to vulnerable patient vs. current approach to cancer treatment once cancer or a tumor is detected. He further drew analogy between PSA and metastatic prostate cancer vs. the future vulnerable patient test and MI or STEMI as equivalent to metastatic prostate cancer.
Dr. Harry de Koning discusses the vulnerable patient detection test and the billion dollar question. Dr. Wolfgang Koenig further discusses the topics and emphasizes on the existing strategies. Dr. Morteza Naghavi reiterates the need for a multipronged approach to prevention of coronary events. Dr. Jagat Narula
Dr. Erling Falk weighs in on risk assessment of asymptomatic vulnerable patient and Dr. Morteza Naghavi discusses that machine learning and new high tech advances can help us detect the signature of vulnerability in asymptomatic individuals and we have a duty to make this happen
 Dr. Marcia Bittencourt discusses his views that there is still room to make progress with exiting primordial and primary prevention. He argues we need to do more for socio behavioral interventions. Dr. Morteza Naghavi counters that the socio behavioral intervention, and dieting and exercise that American Heart Association has taught over the past few decades is unlikely to make a drastic change towards the eradication of heart attacks, and that new strategies are needed both for detection of asymptomatic vulnerable patients and treatment of such cases. Dr. Michael Blaha argues that some important physiological information are missing in such vulnerable cases. Dr. Morteza Naghavi further discusses that vulnerable patients may have periodical signals that may last few months and referred to Dr. Attilio Maseri’s presentation at SHAPE’s 2004 Vulnerable Patient Symposium. Dr. Erling Falk has the last words and questioned if with this approach we would have to exam asymptomatic individuals every 6 months. The focus group brainstorming ends shortly after.

 

 

aha-2016-shape-machine-learning-vulnerable-patient-meeting-dr-naghavi-openingaha-2016-shape-fuster-2aha-2016-shape-naghavi-fuster

Write a comment:

Your email address will not be published.