KEMBAR78
Artificial Intelligence and Diagnostics | PPT
Artificial Intelligence in Diagnostics
Dr. Sunita Maheshwari
Chief Dreamer and Loop closer
Teleradiology Solutions and Telerad Tech
The Need
Increased Imaging for Chronic Diseases and
Emergencies
Increasing Utilization of Imaging
Diagnostics
High speed, High resolution – More data,
More images for analysis
By whom?
Global Specialists Shortage
Radiologists
•US – 1: 10,000
•Singapore – 1: 20,000
•Japan – 1: 35,000
•India - 1: 100,000
•Bangladesh – 1: 1,000,000
•Africa – Acute shortage
Solutions?
Teleradiology
A Technology cum Service Innovation
Creating a Global Workforce in Radiology
t
How can we make a Radiologist more
efficient?
Auto detect a lesion, pull up relevant priors and clinical history
Measure the lesion, compare to previous scans
Have a templated report ready
AVERAGE: ONE RAD CAN
READ 6-10 RVU’S/HR
IF HE CAN READ 20
RVU’S/HR HALF THE
SHORTAGE ISSUE OF
THE WORLD IS
RESOLVED
AI companies and algorithms
• Osteoporosis – an algorithm that receives CT
Chest/Abdomen studies and delivers an alert if the data
suggests low bone mineral density (osteopenia or
osteoporosis)
• Emphysema – an algorithm that receives CT Chest data
and delivers an alert if the data suggest emphysematous
lungs.
• Compression Fractures – an algorithm that receives CT
Chest/Abdomen data and delivers an alert if the data
suggests a vertebral compression fracture.
• Malignant tissue- an algorithm that received 2D
Mammography and alerts if suspicious malignant tissue if
occurs.
• Algorithm that Detects Opacities on Xray Scans
• Xrays: Algorithms that auto detect fractures
Confused clinician?
• There is ‘AI’ on my workflow but will it pick up all the lesions or
just alert on a few diagnoses?
• Different AI companies have different algorithms for different
diseases? How do I access all of them on my one system?
• If AI is not 100%, if accuracy is 80%-should I trust it?
AI today
feels like the 12
taxes!
Need to have
All algorithms on
one platform
Radiology
Auto-detection of Disease
• Brain: Bleeds, Strokes, Tumors
• Bones: Fractures, Tumors
• Heart: Coronary disease
• Lungs: Pneumonia, Cancer
• AI onto PACS: IBM/Merge, Philips, Agfa,
Siemens, Zebra onto RadSpa
AI is not just for radiology…..From
Radiology to Pathology
Digital Slides: Auto Detect Malaria,
Anemia, Leukemia and more
Neurology and AI: Strokes and more
• Auto reporting of EEG’s/EMG’s
• Report generation for review
Cardiologists and AI, Heart attacks and more
• Pedi cardiologist perspective : Can AI help me not miss a hole
in the heart on echo?
• Auto reading of ECG’s, algorithm based MI detection-
already in clinical use, saving lives; auto
measurement of ejection fraction
Arterys recently received clearance from the US Food and Drug Administration (FDA)
to market its Arterys Cardio DL application, which uses AI to provide automated,
editable ventricle segmentations based on conventional cardiac MRI images. Jun 2017
Opthalmology
• Auto detection of diabetic retinopathy
Because people can’t get
screened, half suffer
vision loss before they’re
detected.
There are more than 400
million people worldwide
with diabetes, including
70 million in India.
Opthalmology
MINE -Microsoft Intelligent Network for Eyecare -is a
global consortium of eyecare providers, research and
academic institutions that will apply artificial intelligence to
help in eyecare delivery.
Parting note...
• Google's Peng dismissed suggestions that while this technology might be
good for patients, it could mean fewer jobs for doctors. She said Google’s
algorithms would instead do screening work not being done today due to
skills shortages while freeing physicians for more important tasks.
“There’s not enough expertise to go, we need to have our specialists
working on treating people who are sick.”
Let's not tell a Doctor that AI will replace
him…
It can help him avoid a miss, 1 less malpractice case
It can help him get more efficient and do more work
in the same period of time
It can reduce physician stress (eg have priors ready,
clinical history made available, comb through
thousands of data sets) yet at the same time make
his patient care superior
Tell him it will take him from human doc to...
We are not just Telemedicine providers or Scientists or
Techies…
We are enablers of a better future of Patients and Doctors,
Globally.
info@telradsol.com
www.telradsol.com

Artificial Intelligence and Diagnostics

  • 1.
    Artificial Intelligence inDiagnostics Dr. Sunita Maheshwari Chief Dreamer and Loop closer Teleradiology Solutions and Telerad Tech
  • 2.
    The Need Increased Imagingfor Chronic Diseases and Emergencies
  • 3.
  • 4.
    Diagnostics High speed, Highresolution – More data, More images for analysis By whom?
  • 5.
    Global Specialists Shortage Radiologists •US– 1: 10,000 •Singapore – 1: 20,000 •Japan – 1: 35,000 •India - 1: 100,000 •Bangladesh – 1: 1,000,000 •Africa – Acute shortage
  • 6.
  • 7.
  • 8.
    Creating a GlobalWorkforce in Radiology
  • 9.
    t How can wemake a Radiologist more efficient? Auto detect a lesion, pull up relevant priors and clinical history Measure the lesion, compare to previous scans Have a templated report ready AVERAGE: ONE RAD CAN READ 6-10 RVU’S/HR IF HE CAN READ 20 RVU’S/HR HALF THE SHORTAGE ISSUE OF THE WORLD IS RESOLVED
  • 10.
    AI companies andalgorithms • Osteoporosis – an algorithm that receives CT Chest/Abdomen studies and delivers an alert if the data suggests low bone mineral density (osteopenia or osteoporosis) • Emphysema – an algorithm that receives CT Chest data and delivers an alert if the data suggest emphysematous lungs. • Compression Fractures – an algorithm that receives CT Chest/Abdomen data and delivers an alert if the data suggests a vertebral compression fracture. • Malignant tissue- an algorithm that received 2D Mammography and alerts if suspicious malignant tissue if occurs. • Algorithm that Detects Opacities on Xray Scans • Xrays: Algorithms that auto detect fractures
  • 11.
    Confused clinician? • Thereis ‘AI’ on my workflow but will it pick up all the lesions or just alert on a few diagnoses? • Different AI companies have different algorithms for different diseases? How do I access all of them on my one system? • If AI is not 100%, if accuracy is 80%-should I trust it? AI today feels like the 12 taxes! Need to have All algorithms on one platform
  • 12.
    Radiology Auto-detection of Disease •Brain: Bleeds, Strokes, Tumors • Bones: Fractures, Tumors • Heart: Coronary disease • Lungs: Pneumonia, Cancer • AI onto PACS: IBM/Merge, Philips, Agfa, Siemens, Zebra onto RadSpa
  • 13.
    AI is notjust for radiology…..From Radiology to Pathology Digital Slides: Auto Detect Malaria, Anemia, Leukemia and more
  • 14.
    Neurology and AI:Strokes and more • Auto reporting of EEG’s/EMG’s • Report generation for review
  • 15.
    Cardiologists and AI,Heart attacks and more • Pedi cardiologist perspective : Can AI help me not miss a hole in the heart on echo? • Auto reading of ECG’s, algorithm based MI detection- already in clinical use, saving lives; auto measurement of ejection fraction Arterys recently received clearance from the US Food and Drug Administration (FDA) to market its Arterys Cardio DL application, which uses AI to provide automated, editable ventricle segmentations based on conventional cardiac MRI images. Jun 2017
  • 16.
    Opthalmology • Auto detectionof diabetic retinopathy Because people can’t get screened, half suffer vision loss before they’re detected. There are more than 400 million people worldwide with diabetes, including 70 million in India.
  • 17.
    Opthalmology MINE -Microsoft IntelligentNetwork for Eyecare -is a global consortium of eyecare providers, research and academic institutions that will apply artificial intelligence to help in eyecare delivery.
  • 18.
    Parting note... • Google'sPeng dismissed suggestions that while this technology might be good for patients, it could mean fewer jobs for doctors. She said Google’s algorithms would instead do screening work not being done today due to skills shortages while freeing physicians for more important tasks. “There’s not enough expertise to go, we need to have our specialists working on treating people who are sick.”
  • 19.
    Let's not tella Doctor that AI will replace him… It can help him avoid a miss, 1 less malpractice case It can help him get more efficient and do more work in the same period of time It can reduce physician stress (eg have priors ready, clinical history made available, comb through thousands of data sets) yet at the same time make his patient care superior Tell him it will take him from human doc to...
  • 20.
    We are notjust Telemedicine providers or Scientists or Techies… We are enablers of a better future of Patients and Doctors, Globally. info@telradsol.com www.telradsol.com