greyscale measurements in tendons/tendinopathy

[Simon Locke-2]

Good Afternoon all,

 

I am a clinician and am part of a group that combines basic science and
clinical research into tendons. I say this because I do not have programming
skills but have access to someone.

 

The basic science research we are involved in involves response of tendon to
exercise (load). The clinical component of this research is now exploring
first-order greyscale statistics from longitudinal B mode ultrasound in
various tendons.

 

Our  images are stored as RGB images which are then transferred in "Image J"
to 8-bit greyscale. Over various sites of a tendon, we take multiple linear
samples (ROIs) of the ultrasound image.

 

In the Achilles tendon for example, we wish to define the tendon edge and
exclude the paratenon. In other tendons, we need to exclude synovial sheath.
Generally, these have increase echogenicity in our experience.

 

So, it seems to me that I should use the "edge" control to define the tendon
edge, seek the coordinates (X, Y) at the edge, note them and then return to
the original image and use the linear tool to define the ROI according to
the chosen dimensions (XY coordinates). The linear ROI will provide
first-order statistics or a list that can be exported into a statistics
program.

 

After having followed the various discussion threads over the last few
months, it seems to me that this is a rather long winded way of proceeding.
There is probably a macro that can be used to expedite this.

 

Or indeed, is this the correct way to get dimensions of a tendon in Image J
and the appropriate first-order statistics in a linear ROI?

 

Does anybody have any comment?

 

I do apologise for having limited computer skills. I've also been reading a
lot of the discussion threads with great deal of interest.  Some of them I
have actually stored in my own help file.

 

Thank you for your thoughts in anticipation.

 

Regards

 

Dr Simon Locke

MBBS BMed Sci FACSP FFSEM (UK) FRACGP

Sport and Exercise Physician

Queensland Centre for Sport+Movement

48 Annerley rd., Woolloongabba

 

Ph. 07 3393 2001

Fax [phone]

 

Queensland Centre for Sport+ Movement is now on Facebook

http://www.facebook.com

 

 


--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Anderson, Charles (DNR)]

Attaching an image (or a link to one) with one or more transects marked would help us to generate informed suggestions, especially on ways to segment tendon from surrounding sheath and other tissues.  After that, are you wanting to pick transects automatically or manually?  What statistics do you really want for each transect? Or do you really want locations and grey values saved or exported for analysis?

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Simon Locke
Sent: Wednesday, January 16, 2013 10:41 PM
To: [hidden email]
Subject: greyscale measurements in tendons/tendinopathy

Good Afternoon all,

I am a clinician and am part of a group that combines basic science and clinical research into tendons. I say this because I do not have programming skills but have access to someone.

The basic science research we are involved in involves response of tendon to exercise (load). The clinical component of this research is now exploring first-order greyscale statistics from longitudinal B mode ultrasound in various tendons.

Our  images are stored as RGB images which are then transferred in "Image J"
to 8-bit greyscale. Over various sites of a tendon, we take multiple linear samples (ROIs) of the ultrasound image.

In the Achilles tendon for example, we wish to define the tendon edge and exclude the paratenon. In other tendons, we need to exclude synovial sheath.
Generally, these have increase echogenicity in our experience.

So, it seems to me that I should use the "edge" control to define the tendon edge, seek the coordinates (X, Y) at the edge, note them and then return to the original image and use the linear tool to define the ROI according to the chosen dimensions (XY coordinates). The linear ROI will provide first-order statistics or a list that can be exported into a statistics program.

After having followed the various discussion threads over the last few months, it seems to me that this is a rather long winded way of proceeding.
There is probably a macro that can be used to expedite this.

Or indeed, is this the correct way to get dimensions of a tendon in Image J and the appropriate first-order statistics in a linear ROI?

Does anybody have any comment?

I do apologise for having limited computer skills. I've also been reading a lot of the discussion threads with great deal of interest.  Some of them I have actually stored in my own help file.

Thank you for your thoughts in anticipation.

Regards

Dr Simon Locke
MBBS BMed Sci FACSP FFSEM (UK) FRACGP
Sport and Exercise Physician
Queensland Centre for Sport+Movement
48 Annerley rd., Woolloongabba
Ph. 07 3393 2001
Fax [phone]

Queensland Centre for Sport+ Movement is now on Facebook
http://www.facebook.com
--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Simon Locke-2]

Good morning Charles,

Thank you for replying.

Statistics:
Our preliminary information from various tendons indicate the greyscale is
normally distributed. As a result, comparison of means, match pairs analysis
will be appropriate. So, we would generate mean (approximately 90 units),
standard deviation (average to date is approximately 20 units) etc. It
appears that exporting the quantitative values for analysis provides more
freedom in choosing the appropriate analysis (statistical advice) especially
when comparing normal to abnormal.  Abnormal tendons tendon not to be
normally distributed and may require transformation. Also, the distribution
of the greyscale within the tendon really depends on the type of injury. For
example, there may be a linear tear within a tibialis posterior tendon. As
you can see I've attached a longitudinal image of an Achilles tendon. Image
J already gives us first-order statistics but not the means to analyse
similarities or differences and different sites. This is from my preliminary
assessment over the last couple of months and I would be very pleased to
learn otherwise.

Methods:
At this stage, our plan is to collect linear ROIs manually. There would be a
small number of tendons (20) and we do not believe that a large number of
samples will help our cause. I would be very pleased to hear your opinion.

Image:
I have attached a longitudinal image of an Achilles tendon. This is one of
the tendons we are exploring. Just happens to be big and easy. The yellow
line is measuring the AP dimension and is taken directly from the machine.
We would use the linear tool to select the ROI in such a fashion. We would
use ROI manager to store multiple sites and then export them. As you can
see, removing the paratenon on in this instance or the synovial sheath is
important if possible.


Thank you for your reply.
Regards

Dr Simon Locke
MBBS BMed Sci FACSP FFSEM (UK) FRACGP
Sport and Exercise Physician
Queensland Centre for Sport+Movement
48 Annerley rd., Woolloongabba

Ph. 07 3393 2001
Fax [phone]

Queensland Centre for Sport+ Movement is now on Facebook
http://www.facebook.com


-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of
Anderson, Charles (DNR)
Sent: Friday, 18 January 2013 1:30 AM
To: [hidden email]
Subject: Re: greyscale measurements in tendons/tendinopathy

Attaching an image (or a link to one) with one or more transects marked
would help us to generate informed suggestions, especially on ways to
segment tendon from surrounding sheath and other tissues.  After that, are
you wanting to pick transects automatically or manually?  What statistics do
you really want for each transect? Or do you really want locations and grey
values saved or exported for analysis?

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Simon
Locke
Sent: Wednesday, January 16, 2013 10:41 PM
To: [hidden email]
Subject: greyscale measurements in tendons/tendinopathy

Good Afternoon all,

I am a clinician and am part of a group that combines basic science and
clinical research into tendons. I say this because I do not have programming
skills but have access to someone.

The basic science research we are involved in involves response of tendon to
exercise (load). The clinical component of this research is now exploring
first-order greyscale statistics from longitudinal B mode ultrasound in
various tendons.

Our  images are stored as RGB images which are then transferred in "Image J"
to 8-bit greyscale. Over various sites of a tendon, we take multiple linear
samples (ROIs) of the ultrasound image.

In the Achilles tendon for example, we wish to define the tendon edge and
exclude the paratenon. In other tendons, we need to exclude synovial sheath.
Generally, these have increase echogenicity in our experience.

So, it seems to me that I should use the "edge" control to define the tendon
edge, seek the coordinates (X, Y) at the edge, note them and then return to
the original image and use the linear tool to define the ROI according to
the chosen dimensions (XY coordinates). The linear ROI will provide
first-order statistics or a list that can be exported into a statistics
program.

After having followed the various discussion threads over the last few
months, it seems to me that this is a rather long winded way of proceeding.
There is probably a macro that can be used to expedite this.

Or indeed, is this the correct way to get dimensions of a tendon in Image J
and the appropriate first-order statistics in a linear ROI?

Does anybody have any comment?

I do apologise for having limited computer skills. I've also been reading a
lot of the discussion threads with great deal of interest.  Some of them I
have actually stored in my own help file.

Thank you for your thoughts in anticipation.

Regards

Dr Simon Locke
MBBS BMed Sci FACSP FFSEM (UK) FRACGP
Sport and Exercise Physician
Queensland Centre for Sport+Movement
48 Annerley rd., Woolloongabba
Ph. 07 3393 2001
Fax [phone]

Queensland Centre for Sport+ Movement is now on Facebook
http://www.facebook.com
--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html