Computer aided diagnosis

1
Computer aided diagnosis application in digital
mammography and acute abdominal pain
By
Dr . Usama Gameel Bassuoni Executive manger of
HMIS
Director of medical training Ahmed Maher Teaching
Hospital
2
Medical decision process 3 elements
Information gathering Data analysis
Decision making Human life is too important
to be left to a computer Patients
receive the best treatment on the best
evidence available
3
We must accept our limitations Computers
allow us to access up to date research
worldwide research is there to assist us not
undermine our ability Shared information
advances our knowledge . Encourages us to expand
our practices . Scrutiny of all research to
validate validity . No one person can have all
expert knowledge it is important to
collaborate with all professional colleagues
Use good karma always
4
is a soft ware is an
increasingly role as decision making tools
are being applied in several health care
areas e.g digital mammography and AAP
CAD
CAD
CAD
5

CAD in digital mammography
Cancer Breast ? ? 1-15 40 of CA Breast
Benign In Norway 2100 new cases,800
deaths/year Early detection ? mortality
Screening program based on mammography ?
mortality ? no of countries started mass
screening programs ?increased no mammography
interpretation.
6
Abnormalities embedded , camouflaged by
varying densities of breast tissue 10 30
of ca breast missed by radiologist CAD
design to detect , classification clustered
micro calcification nodules
7
Image segmentation detection , extraction of
clustered micro calcified nodules from back
ground breast tissue Extracted micro
calcification nodules categorized as benign or
malignant ( image classification )
8
Clustered micro calcification early signs (
potential cancerous changes) Micro
calcification small ca deposit accumulate is
breast tissue as a bright spot Cluster defined
to be at least 3 micro calcifications with in 1-2
cm region of mammogram Individual micro
calcifications range 0.1 1.0 mm ( in
mammogram)
9
Main goal of ACD system provide a second
opinion 2nd opinion radiologist use result
of pc analysis of mammogram in making a diagnosis
Digital mammogram ? feature extraction ?
classification ? analysis ? detected tumor.
10
The main goal of CAD is to develop feature
extraction schemes
To generate segment the mammogram into 2
classes clusters of micro calcification
normal tissues



To generate features used to
discrimination benign and malignant clusters of
micro calcifications
11
CAD support in AAP
Over last 20 Years studies in U.K , Europe in
AAP shows 2-3.000.000 / year with AAP
Cost Euro 1.5 million/ year AAP not a
diagnosis in the strict sense but rather a
symptom
12
The most common frequent decision which surgeon
called upon is not merely related to precise
diagnosis but rather to operate or observe?
General rules of investigation Know what you
are doing? dont waste time?, know what you want
?  
13
Main goal early diagnosis and correct
management Initial diagnosis accuracy 45
Perforated app rate 30 (delayed diagnosis ) ,
-ve app rate 30 In Europe last year 200.000
app removed too late , 200.000 normal app removed

14
Health care data classified into 4 categories
. Patient centric data ( direct related to
patient ) . Aggregates data based on performance
and utilization ( resources management data) .
Transforming based data for planning , clinical
and management decision support . Comparative
data for health services researches.
15

• Computer aided decision support system CADSS
• is AAP began in 1970 by professor Tim de-Dambal
• Decision making improved by fall in ve laprotomy
  and
• rate of perforated appendicectomies

16

•  Initial diagnostic accuracy rate rose from
• 45.6 - 63.5
• -ve laprotomy rate fall by almost half
• perforation rate amongst patient with
  appendicitis
• fall 23.7 - 11.5
• Mortality rate fall by 22.0
• Direct cost saving of over 5 m Euro

17

• CAD useful system for improving diagnosis and
  better clinical practice
•   450 physician is 64 hospitals is 19 countries
• ? residual diagnosis error rate by 40
• ? unnecessary operation by 25
• ? in perforation rate of appendicitis case by
  half

18
Changes is like dragon fight it you will loose
.,ignore it., it will eat you., success only
favors those who challenge it. Patients needs
change constantly so therefore we must respond
accordingly to new advancement in technology
19
GOD the phrase of BILL GATES grow or die
20
Thanks a lot