Automatic Parameter

Generating System for

Urinary Calculi Diagnosis
and Treatment Planning
based on 3D CT Images

Presented by
Ma Lai Yee Myint
Ph.D CEIT – 5 (13th Batch)
1
Introduction
• Urinary Calculi are small, hard deposits of mineral and acid
salts on the inner surfaces of the kidneys (drops along the
urinary track).

• Alternative names include:

• Nephrolithiasis
• Urolithiasis
• Ureterolithiasis

http://www.pilotfriend.com/aeromed/medical/images2/25.jpg
2
 Background
• Why doing research in urinary calculi Disease?
– Prevalence of urinary calculi disease and recurrence rate
Lifetime prevalence of incident kidney stones in the United

10.0%
8.8%
9.0%
8.0%

7.0%
6.0% 5.2%

5.0%

4.0% 3.2%

3.0%

2.0%
1.0%
States

0.0%

Fig: Prevalence of urinary calculi disease from 1976 to 2010 Fig: Recurrent rates of urinary calculi forming after treatment [1]
[1]
3
 Background
• GBD(GLOBAL BURDEN OF DISEASES) 2010
– Years of life lost (YLLs) quantify premature mortality in Myanmar [2][3]
Chronic Kidney Disease in period of 20 years (from 1990 to
2010)
1.80%
1.60%
1.40%
1.20%
1.00%
0.80%
0.60%
0.40%
0.20%
0.00%
1990 2010

% of total disease

4
 Related Researches Advantage Limitation
An artificial intelligence-based clinical Predict treatment procedure (stent Using patient history variable,
Previous Works
decision support system for large kidney
stone treatment [4].
place after stone removal, blood urinary stone parameters and
transfusion during the surgery or not) laboratory data

S.T.O.N.E. Nephrolithometry: Novel Surgical Predict only PCNL treatment success Using stone parameters: stone size
Classification System for Kidney Calculi [5] or the risk of perioperative (S), tract length (T), obstruction (O),
complications after PCNL number of involved calices (N), and
essence or stone density (E)

Measuring stone volume - Three-dimensional Classify: Small, Medium and Large No single formula for estimating
software reconstruction or an ellipsoid stone stone volume
algebra formula? [6]
Predicting shockwave lithotripsy outcome for Predict the probability of stone free Using patient history variable,
urolithiasis using clinical and stone computed outcome after ESWL treatment urinary stone parameters
tomography texture analysis Variables [7]

Kidney Stone Diagnosis and Management [8] Predict treatment type Using linear measurement of stone
1. Based on HU value (* inappropriate in irregular stone,
2. Based on size eg: staghorn calculi)
5
 Related Researches Advantage Limitation
Image Quality Improvement in Kidney Stone Detect the stones in CT, identify the size Using semi-automated program for

Previous Works
Detection on Computed Tomography Images [9] and location of stone based on pixel stone parameter prediction
count
The Usefulness of The Hounsfield Unit And Stone To evaluate the use of stone density Using software application, firstly
Heterogeneity Variation in Predicting The variation coefficient (SDVC) as an output mean and SD (standard
Shockwave Lithotripsy Outcome [10] indicator of stone heterogeneity for deviation) in manual-selected stone
predicting extracorporeal shock wave region, and then calculated SDVC as
lithotripsy (ESWL) outcome in urinary SD/mean *100.
calculi
3D Medical Image Visualization and Volume To visualize 3D stone object and to For 3D visualization, it was used an
Estimation of Pathology Zones [11] estimate stone volume algorithm using standard OpenGL
graphic library and extract the stone
region using histogram thresholds (semi-
automatic segmentation)
Noncontract Computed Tomography Parameters for Predict mean stone density, stone Using available imaging 3D software,
Predicting Shock Wave Lithotripsy Outcome in heterogeneity index (SHI) and variation stone parameters are measured to
Upper Urinary Tract Stone Cases [12] coefficient of stone density (VCSD) for provide ESWL outcomes prediction
predicting ESWL outcomes

Stone heterogeneity index as the standard To group High SHI and Low SHI for ESWL Define the stone heterogeneity index
deviation of Hounsfield units: A novel predictor for success and for stone fragility (SHI) as the standard deviation (SD) on
shockwave lithotripsy outcomes in ureter calculi CT using available imaging 3D software
[13]

6
 Bone

Stone
Stent

Problem Formulation Calcification

Failure in some kidney stone cases

 Spreading of kidney stone in the urinary system (kidney, ureter and bladder )
 Consisting of stent, calcification, bone fragment together with stone (appear on CT_ bright-
white appearance similarity, variety of size, shape, intensity and location)
 Consisting of one or more types of stone in each patient (appear on CT_ bright-white
appearance similarity, variety of size, shape, intensity and location)
 Consisting many slices of a stone (≈ 500-600 slices)
 Using drawing tool to select the stone area
 Variety of size, shape, intensity and location to apply Standard formula for CT parameter
measurements
R. Ray, R. Mahapatra, P. Mondal, and D. Pal, “Long-term complications of JJ stent and its management: A 5 years review,” Urol. Ann., vol. 7, no. 1, p. 41, 2015.
7 pp. 268–
H. J. Lee et al., “Differentiation of urinary stone and vascular calcifications on non-contrast CT images: An initial experience using computer aided diagnosis,” J. Digit. Imaging, vol. 23, no. 3,
276, 2010.
 Problem Statement

Failure in some kidney stone cases

• Difficulty to distinguish the stone among other high-intensity objects (stent,

Segmentation
calcification, bone fragment)

Slice Selection • Difficulty to select the largest slice among all included slices in a stone

Parameter • Difficulty to measure the stone parameters for diagnose the stone problem and
Measuring to make medical treatment decisions without the help of expert person 8
Objectives

To obtain all kidney stones in the urinary system simultaneously

without any other unwanted regions (to be clear output object)
To provide the essential information (for stone parameter), 3D
visualization (for stone shape) and 2D internal structure
visualization (for stone composition) for diagnosis and treatment
planning without any other expert persons
To develop automatic parameter generating system for urinary
calculi diagnosis and treatment planning
9
 Start

System Design
Input Imag e

3D segmentation

Soft_organ removing

Bed-map removing and Bony

skeleton removing

False Pos itive reduction

Segmented obj ect (ROI)

No
Is there stone?

Yes

Feature Extraction

CT parameters
and 2D and 3D
visualization for
Diagnosis and
Treatment type
decis ion

End
10
Contribution

11

11
Contribution

Stone size

12

Stone density
(only HU value)

Surgical Management of Urinary Lithasis, by MohammadBaghdadi,

Urology resident KKNGH, https://slideplayer.com/slide/1570041/
12
 Contribution

Stone size

Stone size

Stone density
(only HU value) 13

Stone density
(only HU value)

Stone
heterogeneity
index
J. Y. Lee, J. H. Kim, D. H. Kang, D. Y. Chung, D. H. Lee, and H. Do, “Stone heterogeneity index as the
Surgical Management of Urinary Lithasis, by MohammadBaghdadi,
standard deviation of Hounsfield units : A novel predictor for shock- wave lithotripsy outcomes in
Urology resident KKNGH, https://slideplayer.com/slide/1570041/
13
ureter calculi,” Sci. Rep., vol. 6, no. 23988, pp. 1–7, 2016, doi: 10.1038/srep23988.
 Contribution
Previous work New contribution
• Stone size, density (only HU value) & other hardware • Seven effective CT parameters (Area, 3D
devices (urine, blood, patient’s history, etc…) volume, Diameter, Density, Average HU,
• Stone parameter using CT SHI, VCSD) {not only for diagnosis but
also for treatment planning]
 Largest cross-sectional slice of the largest stone where • Automate stone region selection (without
more than one stone was treated. Stone volume is the sum any noise, without using any drawing
of individual stone volumes in cases with more than one tools)
stone • Auto 2D stone parameters measurements
(auto largest slice selection)
 Stone area on this selected slice (Manually or • Auto 3D stone parameters measurements
• Completely depend on radiologist (may be variability-reduce
Automatically)
to get Time consuming & variability
accurate CT information) (Not only one slice)
• 3D visualization (to show the shape and external structure of • Automated computer aided system
the stone together with other high intensity objects such • 3D visualization and 2D internal structure
as bones) visualization (to display shape, number,
the stone composition), clear output
without any disturbances

14
 Proposed Methods
3D Parameter Measuring and 3D
Visualization

False Positive • Binarization • Iso-surface generation

• Thresholding Reduction techniques using Marching Cubes
• Logical AND
and Otsu’s operation Algorithm for 3D reconstruction
method • Geometric
• Morphological • Geometric
Features (Area,
operation feature-
Size, Volume,
based
Density, stone
3D Segmentation thresholding • Fuzzy Logic for Largest slice
heterogeneity
(Noise Removal) index and selection
variation • MATLAB function to
coefficient of represent HU values of each
stone density ) pixel
2D Parameter Measuring and
Feature Extraction Internal Structure
Visualization15
Proposed Methods

3D Parameter
Measuring
and 3D
Visualization
3D
Segmentation False Positive Feature
(Noise Reduction Extraction
Removal)
2D Parameter
Measuring
and Internal
Structure
Visualization

16
 Thresholding and Otsu’s Method
• Thresholding:
– Simplest methods for image segmentation
– Assign _ every pixel above a defined intensity (threshold value) as one type of region (foreground),
whereas every pixel below that intensity as another type of region (background).
– Two types of thresholding :
• Global Thresholding: only one threshold value for entire image
• Local Thresholding: different value for different regions
• Otsu’s Method
– Used to find the threshold value for segmentation
– Based on selecting the lowest point between
two classes (peaks).

17
 Morphological Operations
• Morphological operation:
– A broad set of shape-based image processing operations.
– The value of each pixel in the output image is based on a comparison of the corresponding
pixel in the input image with its neighbors. By choosing the size and shape (Circular disks,
crosses, diamonds, lines and squares etc.) of the neighborhood, a morphological operation
can be constructed that it is sensitive to specific shapes in the input image.
• Four basic morphological operations
– Morphological Dilation : to fill holes and close gaps in lines
– Morphological Erosion : for shrinking an object in the image and removing isolated pixels
– Morphological Opening : application of erosion followed by dilation (used to remove small
objects)
– Morphological Closing : application of dilation followed by erosion (used to close small holes
in an objects) 18
 Morphological Operations
• Morphological operation:
– A broad set of shape-based image processing operations.
– The value of each pixel in the output image is based on a comparison of the corresponding
pixel in the input image with its neighbors. By choosing the size and shape (Circular disks,
crosses, diamonds, lines and squares etc.) of the neighborhood, a morphological operation
can be constructed that it is sensitive to specific shapes in the input image.
• Four basic morphological operations
– Morphological Dilation : to fill holes and close gaps in lines
– Morphological Erosion : for shrinking an object in the image and removing isolated pixels
– Morphological Opening : application of erosion followed by dilation (used to remove small
objects)
– Morphological Closing : application of dilation followed by erosion (used to close small holes
in an objects) 19
3D Segmentation (Noise Removal)

• Low_intensity region removing

– Intensity-based thresholding method
• Unneeded high_intensity region removing
– Size-based thresholding method

20
Intensity-based thresholding method

 In step 1, Two constant threshold values, and (200

HU and 2800 HU values )
 Step 2, foreground and background are segmented
into two groups of pixels and , using and :
(1)
 Step 3, finally, the new image is outputted with,
after eliminating group.
21
Size-based thresholding method

 Step 1, Calculate the size (longest diameter )of each object

through the bounding box calculation.
 Step 2, Two constant threshold values, and (3mm and 50
mm).
 Step 3, Foreground and background are segmented into two
groups of pixels and without using, and
(2)
 Step 4, Finally, the new image is outputted with , after
elimination group.
22
Proposed Methods

3D Parameter
Measuring
and 3D
Visualization
3D
Segmentation False Positive Feature
(Noise Reduction Extraction
Removal)
2D Parameter
Measuring
and Internal
Structure
Visualization

23
False-positive reducing

• Volume and average HU values -based thresholding

method

* One of new contribution of the research

24
Volume and average HU values -based thresholding method

 Step 1, Calculate the required threshold values ( and ) using following equation, and
regarded with 0.03 and 35 as consideration of the relationship between volume and
HU of kidney stone.


– is the average HU for each object, n is the number of voxels in an object. µ is the HU value of each voxel in an object.

 Step 2, Segment the image into two groups of pixels and without using, and
•
 Step 3, Develop the new image with , after elimination group.
25
Proposed Methods

3D Parameter
Measuring
and 3D
Visualization
3D
Segmentation False Positive Feature
(Noise Reduction Extraction
Removal)
2D Parameter
Measuring
and Internal
Structure
Visualization

26
Geometric Features
Localization 100 500 120

• Binarization (1 for foreground and 0 for background) 900 800 200

Logical AND operation I 700 600 180

• ⨂0=0, 0⨂1=0, 1⨂0=0, 1⨂1=1 0 1 0

1 1 0

Area
M 1 1 0

• Area=Total number of voxel in the selected region

0 500 0

Volume
I۪ M=R 900 800 0

• Volume = area of object × voxel resolution (voxel spacing) 700 600 0

Geometric Features
Average HU
𝑛

∑ µ𝑖
𝑖 − 1
𝐻𝑈 𝑎𝑣𝑔 =
𝑛

Diameter

• Calculate x(length) and y(width) of stone using “Bounding Box” method

• Calculate diameter “D” by choosing the larger one from x and y

HU Density
Stone
𝐻𝑈 Heterogeneity Index (as SD)
𝑎𝑣𝑔
𝐻𝑈𝐷 =
𝐷

variation coefficient of stone density 𝑆𝐷 =

𝑆𝐷
√ ∑ ( 𝑥𝑖 −
𝑁 −
𝑉𝐶𝑆𝐷 =
Proposed Methods

3D Parameter
Measuring
and 3D
Visualization
3D
Segmentation False Positive Feature
(Noise Reduction Extraction
Removal)
2D Parameter
Measuring
and Internal
Structure
Visualization

29
Largest Slice Selection

• Fuzzy Logic (If-then rule)

 Step 1: Set the slice, Si=(S1, S2, S3, …, Si)
 Step 2: Set the area of each slice, ai=(a1, a2, a3, …, ai)
 Step 3: Larea=0, Lslice=1;
 For Si=1:i
 If (Larea<Si(ai))
 Lslice=i
 End
 Larea=ai
 End

* One of new contribution of the research 30

2D Parameter Measurements

• Apply all methods of 100

100
500
500
120
120

feature extraction
900 800 200
900 800 200

II 700
700
600
600
180
180

process on the ROI of

a selected slice
0 1 0
0 1 0

11 11 00

M 11 11 00

00 500
500 00

I۪ M=R 900
900 800
800 00

700
700 600
600 00
31
2D Internal Structure Visualization
•Step 1. Set HU= (x1, x2, x3, …, xi), xi is represented the HU values of each pixel in the specified
region.
•Step 2. For i=1to i,
if ( xi > 3000) out =8;
else if ( xi > 2500 && HU <= 3000 ) out =7;
else if ( xi > 2000 && HU <= 2500 ) out =6;
else if ( xi > 1500 && HU <= 2000 ) out =5;
else if ( xi > 1000 && HU <= 1500 ) out =4;
else if ( xi > 500 && HU <=1000 ) out =3;
else if ( xi > 200 && HU <= 500 ) out =2;
else out=1; * One of new contribution of the research
End
End
•In here, 1,2,3,..,8 is described as the eight different colors to represent each pixel intensity range.
•Step 3. Display with each value of pixels with 1,2,3,…,8. 32
Proposed Methods

3D Parameter
Measuring
and 3D
Visualization
3D
Segmentation False Positive Feature
(Noise Reduction Extraction
Removal)
2D Parameter
Measuring
and Internal
Structure
Visualization

33
3D Parameter Measurements 100
100
500
100

900
500
120
500

800
120
120

200
100 500900 120800 200

I
100 500900 120800
700
200
600 180

I
100 500 120

• Apply all methods of

900 800700 200600 180

I 900 800700 200600 180

I 900 800
700
200
600 180

feature extraction I
0 1 0
700 600 180

I
0 1 0
700 600 180
0 1 0
1 1 0

z (number of slices), compare x, y and z

process on the ROI of
0 1 1 0 1 0

M0 1 1 0 1
1
0
1 0

M0 1
1
0
1 1 0 1 0

all slices of a stone M

100 500 120

1 1 1 0 1 0
900 800 200

M 1 1 0
100 500 120

I 700 600 180 1 1 0

M
100 500 120
900 800 200 0 500 0
1 1 0

M
900 800 200

I 0 500 0

I۪M=R
M=R
700 600
0 180
1 0
100 500 120
1 1 0
I 700 600 180
0 500 0

I۪
100 500
900 120
800 200
0
1

1
1

0
0
900 800 0
M

I۪ M=R
I 900 800
700 200
600
0 180
1
1
0
1

1
1

0
0
0 500900 0 800 0
I 700 600 180
0 500900 0 800 0

I۪M=R
M=R
1 1 0

M 0 1 0
1 01 0
500 0 700 600 0
M 0 500 0
I۪ M=R
I۪
1 1 0

0 1 0
1 0 900 800 0 900 800700 0 600 0

I۪ M=R
0 500 0

M 700 600 0 900 800700 0 600 0

I۪ M=R
1 1 00
1 0
500 0
900 800 0

I۪ M=R
M 1 1

0
0
900

500
800
700
0
0
600 0
900 800
700
0
600 0
700 600 0
700 600 0
I۪ M=R 0 500
900 0
800 0

700 600 0
I۪ M=R 900 800
700 0
600 0

700 600 0
34
Proposed Methods
• Iso-surface generation techniques using Marching Cubes
Algorithm for 3D reconstruction
 Isosurface generation is a technique that visualizes the 3D volume
datasets using intermediate polygonal based representation
 Marching cubes, Dividing Cubes and Dual Contouring
• Marching Cubes Algorithm
 Employ a divide and conquer approach
 Operates on two basic steps: detection of active cells and provoking
the triangles inside each active cell
35
Marching Cubes Algorithm
• Input Data
 Input (voxelization representing a scalar field v=f(x,y,z))
 isovalue (threshold value) to classify samples (inside/outside the surface)
• Function
 Generate : Geometry (vertices ) and Topology (triangles connecting these
vertices)
 1 for vertices
• Output Surface
 Isosurface joining all points with the
choosen isovalue.
36
 Input (CT) Image

System Design of Experimental Results Image Preprocessing and

Segmentation

Start Segmented Object (stone)

Input Image Largest Slice Selection

Measuring CT values

3D Auto Measurement
Measuring CT values
Hypodense and Isodense
(Average HU, Area, (Average HU, Area, Volume,

2D Auto Measurement
Region Removing
Diameter, Density)

Feature Extraction
Diameter, Density)

Unneeded Hyperdense
Region Removing Stone Heterogeneity Index Stone Heterogeneity Index

False Positive Reducing

Variation Coefficient of Stone Variation Coefficient of Stone
Density Density

Segmented Object
(ROI)

Internal Structure of the Stone 3D Visualization

End 37
 Experimental Results
(Pre-Processing)

 Working on 551 CT slices of a patient, our experiments

demonstrated that the proposed algorithm is removed unwanted
surrounding objects step by step.

2D Slide mode of 3D Volume mode of Result of after Low Intensity

original DICOM original DICOM image region removing
image 38
Experimental Results
(Pre-Processing)

Result of after Unneeded high intensity Result of after false positive removing
region removing
39
Results Analysis 1
(Pre-Processing)

• After loading 500-600 slices of each patient, it is constructed the original image which consists of
over 500000 labels.
• In the first process, the algorithm can remove all regions having hypodense and isodense structure
from the abdomen CT image and reduce the unwanted objects up to 4300 labels on average from
over 500000 labels.
• In the second process, hyperdense region removing, the algorithm removes the bony skeleton, bed
mat, and some bones but the average of the remaining labels is about 112.
• The final output is clear in visualization and remains only an average of 21 labels in an image.
• So, it can remove most of unwanted objects.

Hypodense and isodense region Hyperdense region removing False-positive reducing

removing
4300 labels 112 labels 21 labels

Table I. Average Result of All 50 Patients in Each Process 40

Results Analysis 2
(Pre-Processing)


 where, TP and FN are used for kidney stones which are correctly detected and not detected in system,

TP FN Sensitivity
66 3 95.7%

Could carry out the unneeded regions removing with a significant stone detecting.

41
 Outcomes
(Pre-Processing)

• The proposed method, unwanted region removing, gives good support in kidney stone
detection, and correctly remove the unneeded region and provide the output image in a
clear form.
• Because of using thresholding methods based on the nature of kidney stones on CT
imaging, the proposed scheme is simple and easy to understand for kidney stone
classification.
• The result is good and satisfactory.
• According to the analysis, the proposed model could detect the kidney stones with 95.7%
in sensitivity.
• Therefore, the performance of the proposed scheme is effective and efficient with good
sensitivity. 42
Experimental Results
(Manual Parameter Measurements)

Area
Diameter
Average HU
SD (stone heterogeneity index, SHI)
Variation Coefficient of Stone Density (VCSD)
43
Density (from average HU, Average HU/diameter)
Experimental Results
(2D and 3D Parameter Measurements)

Manual Measurement 2D Measurement 3D measurement

Area 472 737 7172
Volume 4388.85
Average HU 1077.75 1140.52 977.62
Diameter (cm) 2.35 2.345 2.346
Density 466.94 470.81 416.71
SHI 357.57 357.73 370.04
VCSD 0.33 0.32 0.37
44
Experimental Results
(2D and 3D Parameter Measurements)

45
Experimental Results
(2D and 3D Parameter Measurements)

46
 Results Analysis
(2D and 3D Parameter Measurements)

Table III. Summarization the Results of Each Analysis of the Proposed Study 47
Outcomes
(2D and 3D Parameter Measurements)

• Can choose the largest slice among all input slices for 2D Measurement.
• Can give the CT information (parameters) accurately and automatically for
both 2D and 3D measurement.
• Can measure in any circular shape and irregular shape because of pixel-based
measurement (manual with ellipse or circle tools, shape-based measurement).
• Can give CT information for all stones in a patient simultaneously.
• More robust, more accurate, less time-consuming
• Can reduce the variability which can give better information for diagnosis
• Can provide the clear output with 3D visualization that could rotate in any
direction
• Can display the internal structure of the stone with pixel intensity revealing 48
Expected Benefits of research

To be a good assist for urologist in making medical decision

Speed up the diagnosis process

(By giving CT information with 7 parameters)

Decide the precise treatment decisions

(By giving 2D internal structure and 3D visualization)

49
Limitations
• More robust 3D parameter (slightly difference compare with 2D
measuring)
• Impossible to get 3D ground truth data in Myanmar (only 2D ground truth
data)
• Produce only the CT parameter measuring because of ground truth
dataset limitation.
• Require 2.2 GHz for CPU and 2 GB of GPU to speed up the processing time
and produce 3D images quickly.

50
 Further Extension

• It is envisaged that more CT kidney stone datasets with more ground truth data
categories will be collected in the future to assist treatment type classification
in addition to CT information.
• If more ground truth data and CT data sets can be collected, more accurate
treatment decisions can be made for urologists.
• By using CT images, not only stones in the urinary system, but also stones in
other organs in the human body, such as gallstones, can be carried out based
on this research system, and 3D automated system models can be continued.
51
 Conclusion
• Seven types of parameters as CT information that can speed up the diagnosis process were
presented.
• The stone region was automatically selected without any noise.
• The largest slice was also automatically selected using automatic 2D stone parameter
measurements.
• The proposed study was able to show 3D visualization and 2D internal structure visualization
using automated computer system to display the stone composition.
• By using the proposed system, CT can be used to accurately and automatically measure any
shape, reduce variability, and provide clear output with 3D visualization that can rotate in any
direction.
• Believe that the research work can provide to assist in diagnosing the stone problem and
making medical treatment decisions without the help of an expert person.
52
 References
[1] M. S. C. Morgan and M. S. Pearle, “Medical management of renal stones,” Bmj, vol. 2005, p. i52, 2016.
[2] YLLs, “WHO-Health metrics_GBD profile: Myanmar,” vol. 2010, no. Gbd, pp. 1–4, 1990.
[3] “SUPPORT LIFE EXPECTANCY RESEARCH!,” WORLD HEALTH ORGANIZATION 2017, 2017. [Online]. Available:
http://www.worldlifeexpectancy.com/myanmar-kidney-disease.
[4] Shabaniyan, T., Parsaei, H., Aminsharifi, A. et al. An artificial intelligence-based clinical decision support system for
large kidney stone treatment. Australas Phys Eng Sci Med 42, 771–779 (2019).
https://doi.org/10.1007/s13246-019-00780-3
[5] Zhamshid Okhunov, M.D., Arvin K. George, M.D., Arthur D. Smith, M.D., and Zeph Okeke, M.D., Arthur Smith Institute
for Urology, Hofstra North Shore-Long Island Jewish School of Medicine, New Hyde Park, New York
http://dx.doi.org/10.1016/j.urology.2012.10.086
[6] W. Finch, R. Johnston, N. Shaida, A. Winterbottom, and O. Wiseman, “Measuring stone volume - Three-dimensional
software reconstruction or an ellipsoid algebra formula?,” BJU Int., vol. 113, no. 4, pp. 610–614, 2014.
[7] HelenW. Cui, Mafalda D. Silva, AndrewW. Mills, BernardV. North and BenjaminW.Turney, “Predicting shockwave
lithotripsy outcome for urolithiasis using clinical and stone computed tomography texture analysis variables,”
Natureresearch, 11 October 2019.
53
 References (cont’d)
[8] Dushyant V Sahani, MD, “Kidney Stone Diagnosis and Management.,” Department of Radiology, Massachusetts General
Hospital., vol. 27, no. 2, pp. 162–7, 2013.
[9] Ebrahimi, S., & Mariano, V.Y. (2015). Image Quality Improvement in Kidney Stone Detection on Computed Tomography
Images. Journal of Image and Graphics, 3.
[10] Oktay C, Çoraplı M, Tutuş A. The usefulness of the Hounsfield unit and stone heterogeneity variation in predicting the
shockwave lithotripsy outcome. Diagn Interv Radiol. 2022 May;28(3):187-192. doi: 10.5152/dir.2022.20945. PMID:
35748199; PMCID: PMC9634939.
[11] Krechetova, K., Glaz, A., Platkajis, A. (2008). 3D Medical Image Visualization and Volume Estimation of Pathology Zones.
In: Katashev, A., Dekhtyar, Y., Spigulis, J. (eds) 14th Nordic-Baltic Conference on Biomedical Engineering and Medical
Physics. IFMBE Proceedings, vol 20. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-69367-3_142
[12] Yamashita, S., Kohjimoto, Y., Iwahashi, Y., Iguchi, T., Nishizawa, S., Kikkawa, K., & Hara, I. (2018). Noncontrast Computed
Tomography Parameters for Predicting Shock Wave Lithotripsy Outcome in Upper Urinary Tract Stone Cases. BioMed
Research International, 2018. https://doi.org/10.1155/2018/9253952
[13] J. Y. Lee, J. H. Kim, D. H. Kang, D. Y. Chung, D. H. Lee, and H. Do, “Stone heterogeneity index as the standard deviation of
Hounsfield units : A novel predictor for shock- wave lithotripsy outcomes in ureter calculi,” Sci. Rep., vol. 6, no. 23988, pp.
1–7, 2016, doi: 10.1038/srep23988. 54
 Achievements

1. Lai Yee Myint, Su Su Maung, and Khine Thin Zar, "Removal of Unwanted Object in 3D CT Kidney Stone
Images and 3D Visualization," 2020 24th International Computer Science and Engineering Conference
(ICSEC), 2020, pp. 1-5, doi: 10.1109/ICSEC51790.2020.9375155.

2. Lai Yee Myint and Khine Thin Zar, “Automated Size Measurements of Kidney Stone on 3D CT Image,”
Mandalay Technological University Journal of Science, Engineering and Technology, Volume 8, December
2021, pp. 348-352.

3. Lai Yee Myint and Khine Thin Zar, “Modify HU Density Estimation for Urinary Calculi on 3D CT Image,”
Engineering and Technology, Yangon Technological University, December 2022.

4. Lai Yee Myint and Khine Thin Zar, “Automated Measurement of Stone Heterogeneity Index and
Variation Coefficient of Stone Density on CT Image,” Journal of Engineering Technology and Applied
Physics, Vol 5 No 1 (2023), pp. 54-62, https://doi.org/10.33093/jetap.2023.5.1
55
THANK YOU VERY MUCH FOR
YOUR ATTENTION !!

56