Name: 3d ultrasound imaging: fast and cost-effective morphometry of musculoskeletal tissue
Uploaded: 2017-11-27T12:30:00
Duration: 8 min 52 s
Description: Watch this Scientific Journal Video about 3D Ultrasound Imaging: Fast and Cost-effective Morphometry of Musculoskeletal Tissue at JoVE.com

The authors are very grateful to Adam Shortland and Nicola Fry who shared their algorithms for the 3-dimensional ultrasound in 2004, which were the inspiration for the development of the software used in this study.

The authors have nothing to disclose.

A valid and reliable 3DUS technique is presented that allows for the fast analysis of morphometric variables of skeletal muscles. Different 3DUS approaches for soft tissue imaging have been available for approximately a decade42,43, however the 3DUS approaches are still not used commonly. MRI is a &#39;gold standard&#39; for estimation of in vivo muscle volumes (e.g., references16,17,18,19,20). MRI validity has been tested and confirmed in studies comparing either phantoms or cadaveric organs of known volume to MRI-based volume estimates44,45. However, MRI availability for research is limited and scans are time consuming and costly. In addition, experimental subject postures are limited by the bore seize of the MRI scanners. Typical MR images generate insufficient contrast to perform measurements of variables of muscle geometry (fascicle lengths and angles). However, 3D muscle geometry can be assessed also using MRI by using additional techniques, e.g., DTI technique21. Similar to MRI, US imaging provides adequate distinction at interfaces between different types of tissues (i.e.&#160;visible within US images), providing a valid modality for soft tissues volume assessment1,30,44,46,47,48,49. In contrast to MRI, 3DUS images have sufficient contrast to perform analysis on both volume and muscle geometry from the same measurement.
In addition, the technique presented allows combining images of multiple sweeps into one array, for the study of larger muscles. This new 3DUS method provides a potential tool for clinical assessment of muscle morphology. This method can be used also for imaging soft-tissue structures other than muscle (e.g., tendons, internal organs, arteries).
Modifications to Improve Offline Processing Time:
Modifications of the 3DUS approach were mainly aimed at improving processing time and measuring larger muscles. The offline processing time of a 3DUS image depends on voxel array settings, sampling frequency, size of ROI, duration and speed of the sweep, number of sweeps, and the used workstation. Previously, a reconstruction time of &#8776; 2 h was necessary for reconstructing only one sweep yielding 750 US images (30 s at 25 Hz)15,25,30. With the present 3DUS method, the same sweep takes only 50 s reconstruction time (improving the &#39;offline&#39; processing time by 99%). This improvement can be explained by the enhanced filling algorithm that utilizes large vector operations to fill the voxels frame-by-frame, instead of pixel per pixel and increased random access memory (RAM) of workstations to construct larger voxel arrays. With the new 3DUS approach, a typical reconstruction representing a sweep length of 30 cm at a speed of 1 cm/s, with a target voxel size of 0.2 x 0.2 x 0.2 mm3 and a sampling frequency of 25 Hz, takes the following time to reconstruct:
a. Approximately 10 s to identify the synchronization pulse and select relevant US images. 
b. Approximately 120 s to determine the calibration transformation matrix (PrTIm). 
c. Approximately 10 s for the bin-filling stage. 
d. Approximately 30 s for executing the gap-filling steps.
In total, taking 170 s. Note, step b only needs to be performed once, assuming a rigid connection of the MoCap markers to the probe, leaving 50 s for the reconstruction of a single sweep. Combining two single sweep reconstructed voxel arrays takes approximately 10 s.
Limitations and Critical Steps:
There are several 3DUS imaging aspects that should be taken into account:
i. US image quality: Higher spatial resolution of 2D US images provide more pixels to be placed within the voxel array. This would allow the voxel dimensions to decrease, leading to higher voxel density. Several currently available ultrasound machines use spatial compounding to reduce the noisy granular texture, allowing for better artifact-free distinction of the interfaces of tissues. Another option to reduce speckle is edge enhancement. However, it should be noted that this approach is not desirable, since it deforms the image in an attempt to create distinct interfaces, thereby distorting the true anatomical position of the interfaces.
ii. MoCap accuracy: Pixels can only be accurately placed into a voxel, if the position sensor accurately quantifies the coordinates of the probe. With an increase in image resolution, MoCap accuracy becomes more important. The presented 3DUS setup works best with a voxel dimension of 0.2 x 0.2 x 0.2 mm3, using a MoCap system with an accuracy of 0.1 mm, providing ample accuracy to reconstruct the 3DUS voxel array.
iii. Sample frequency: The lowest temporal resolution of either the US images or MoCap data stream determines the sample frequency. This affects the sweep time or the voxel array settings. For instance, doubling the sample frequency from 25 to 50 Hz allows a sweep to be performed in half the time. Alternatively, not changing the sweep speed, provides more images to fill the voxel array, leaving fewer gaps to be filled and thereby potentially increasing the voxel array resolution. However, increasing the voxel array resolution, without increasing the sampling frequency, requires a slower scan, which will increase the potential of movement artifacts.
iv. Image reconstruction time: Fast reconstructions require a powerful workstation with sufficient available RAM. In addition, reconstruction time varies largely based on the voxel array volume and complexity of the gap-filling process.
v. Experimental protocol: Standardization of the experimental protocol, as exemplified in the present study for the VL and GM, is essential for comparison of morphological measurements (e.g., fascicle length, fascicle angle, muscle belly length, tendon length, aponeurosis length) between subjects and monitoring within subjects in longitudinal studies. However, note that the morphology assessed at rest may alter during muscle activation. For example, for the VL experiment, the knee extensor morphology during maximal contraction may demonstrate a high pennation angle and shorter fascicles in 60&#176; knee flexion, in comparison with morphology at rest50. In certain conditions (e.g., spasticity), electromyography (EMG) may be used to verify resting muscle activity levels during examination.
vi. Probe pressure and tissue deformation: If ample ultrasound gel is applied on the ROI, the amount of pressure to remain for full contact between probe and the skin is limited. As guidance, we advise that scanning a ROI should feel like hovering over the skin, and pressure should only be applied to keep in contact with the gel and thereby the skin. However, slight tissue deformation may be inevitable, even with a generous amount of ultrasound gel. Probe size and a curved ROI affect the required amount of pressure or gel used. Larger probe size and a more curved ROI require more pressure and/or more gel, than smaller probes with a similar curved ROI. Another possible solution is to discard the reverberation (i.e.&#160;non-skin-contact) region of the US images. In addition, tissue deformation is most likely to occur in the first tissue layers, such as skin and subcutaneous adipose tissue layers. Note that subjects with little to no subcutaneous adipose tissue are therefore more prone to adverse effects of pressure. In addition, the tissue deformation occurs most likely at the center of the probe, which is typically not the region of overlap with other sweeps.
vii. Imaging and anatomical knowledge: Another important consideration in using any imaging modality is that the knowledge of both the anatomy and the imaging modality is essential to obtain meaningful interpretation. Anatomical variation between subjects and image artifacts needs to be recognized and taken into account in the identification process of anatomical structures. Even with healthy and/or well developed muscles, clear identification can be difficult because it requires anatomical knowledge to differentiate between different components of one muscle or between muscle groups51. However, in atrophied muscle (i.e.&#160;elderly, in the case of pathology, or a cadaver), the clear identification is even more complicated because of a smaller size and decreased image contrast, and therefore less distinct tissue interfaces (Figure 4). We believe that without prior anatomical knowledge, we would have been limited in making correct judgements in designing this 3DUS approach and in performing the 3DUS measurements. For example, for GM experiments, different footplate angles do not necessarily cause expected changes in muscle tendon complex lengths, due to deformation within the foot7. Also detailed anatomical information on curvature of the distal aponeurosis was essential for an adequate selection of the mid-longitudinal plane in all subjects38.
<img alt="Figure 4" class="xfigimg" src="/files/ftp_upload/55943/55943fig4.jpg" /> 
Figure 4: Variation and quality of reconstructed anatomical cross-sectional 3DUS images of the quadriceps muscle halfway along the thigh. (A) Example of a male human cadaver shows an image of an atrophied state at death (death age: 81 years). Identification of the boundaries of individual heads of the quadriceps muscle is difficult. (B) Example of a sedentary male (aged 30 years). (C) Example of a male athlete rower (aged 30 years). The white squares for scale represent 1 cm x 1 cm. <a href="//cloudflare.jove.com/files/ftp_upload/55943/55943fig4large.jpg" style="font-size: 14px;" target="_blank">Please click here to view a larger version of this figure.</a>
Future Applications:
The 3DUS approach provides an imaging tool that can be used for various purposes and settings in sports and clinics. In clinical interventions effectiveness is related to the physical fitness level52. Using 3DUS for monitoring patients who are at risk of losing muscle mass is important (e.g., references53,54,55) and potentially allows for adjustment of the treatment. Another potential application of 3DUS lies in monitoring the morphological adaptation of muscle in response to intervention (training) and/or injury.
This protocol described a cost- and time-effective method of measuring soft tissue structure of the human body based on freehand 3DUS sweeps. Moreover, assessment of the meaningful morphological parameters of m. vastus lateralis and m. gastrocnemius medialis proved to be valid and reliable.

In healthcare and sports, information on the morphology of muscles is very valuable in diagnostics and/or follow-up evaluations after treatment or training1. Ultrasound (US) imaging is a tool commonly used for visualization of soft tissue structures in muscle diseases2, critical illnesses3,4, cardiovascular diseases5, neurological disorders6,7,8, and effects of physical training6,9,10. US imaging enables quantification of muscle size, fascicle length, and angle of pennation. These morphological variables are important determinants of muscle force and length range of force exertion11,12,13,14,15.
Currently, US imaging measurements are mostly performed in 2D images, with the examiner choosing a presumably, suitable orientation and location of the ultrasound probe. Such 2D methods restrict morphological measurements to one image plane, while the parameter of interest may not be present within this plane. Morphological analysis requires a 3D approach, providing out-of-plane measurements using 3D reference points. Such a 3D morphological representation of soft tissues is known to be provided by Magnetic Resonance Imaging (MRI)16,17,18,19,20. However, MRI is expensive and not always available. Also, visualization of muscle fibers requires special MRI sequences, such as diffusion tensor imaging (DTI)21. A cost-effective alternative to MRI is 3D ultrasound (3DUS) imaging. The 3DUS approach provides several advantages over MRI techniques, e.g., it imposes less space limitations for positioning the subject during an examination. 3DUS imaging is a technique sequentially capturing 2D (B-mode US) images and positioning them into a volume element (voxel) array22,23,24. The process of 3DUS image reconstruction consists of five steps: (1) Capturing a series of freehand 2D US images; (2) Tracking the position of the US probe, using a Motion Capture (MoCap) system; (3) Synchronizing the MoCap position and US images; (4) Calculating the location and orientation of the ultrasound images within the voxel array using a calibrated system of reference; and (5) Placing these images into this voxel array.
The 3DUS approach has been successfully applied for assessment of morphology of skeletal muscle15,25,26,27,28,29. However, previous approaches7,15,25,30 have proved cumbersome, time consuming and technically limited, as only small segments of large muscles could be reconstructed.
To improve the 3DUS approach, a new 3DUS protocol has been developed that allows reconstruction of complete muscles within a short period of time. This protocol article describes the use of 3DUS imaging for morphometry of the m. vastus lateralis (VL) and m. gastrocnemius medialis (GM).

<table>
	<tbody>
		<tr>
			<td>Ultrasound device (Technos MPX)</td>
			<td>Esaote, Italy</td>
			<td>NA</td>
			<td></td>
		</tr>
		<tr>
			<td>Linear array probe (12.5 Mhz, 5 cm)</td>
			<td>Esaote, Italy</td>
			<td>NA</td>
			<td></td>
		</tr>
		<tr>
			<td>Workstation (HP Z440)</td>
			<td>HP, USA</td>
			<td>http://www8.hp.com/us/en/workstations/z440.html</td>
			<td></td>
		</tr>
		<tr>
			<td>Framegrabber (Canopus, ADVC 300)</td>
			<td>Canopus, Japan</td>
			<td>ADVC 300</td>
			<td></td>
		</tr>
		<tr>
			<td>Motion Capture System (Certus)</td>
			<td>NDI, Canada</td>
			<td>http://www.ndigital.com/msci/products/optotrak-certus/</td>
			<td></td>
		</tr>
		<tr>
			<td>Synchronisation device</td>
			<td>VU, NL</td>
			<td>Contact corresponding author</td>
			<td></td>
		</tr>
		<tr>
			<td>Calibration frame</td>
			<td>VU, NL</td>
			<td>Contact corresponding author</td>
			<td></td>
		</tr>
		<tr>
			<td>Thermometer</td>
			<td>Greisinger, Germarny</td>
			<td>GTH 175/PT</td>
			<td></td>
		</tr>
		<tr>
			<td>Examination table</td>
			<td>NA</td>
			<td>NA</td>
			<td>Any examination table</td>
		</tr>
		<tr>
			<td>Inclinometer</td>
			<td>Lafayette instrument, USA</td>
			<td>ACU001</td>
			<td></td>
		</tr>
		<tr>
			<td>Adjustable Footplate</td>
			<td>VU, NL</td>
			<td>Contact corresponding author</td>
			<td></td>
		</tr>
		<tr>
			<td>Torque wrench</td>
			<td>VU, NL</td>
			<td>Contact corresponding author</td>
			<td></td>
		</tr>
		<tr>
			<td>Extendable rod</td>
			<td>VU, NL</td>
			<td>Contact corresponding author</td>
			<td></td>
		</tr>
		<tr>
			<td>Goniometer (Gollehon)</td>
			<td>Lafayette instrument, USA</td>
			<td>1135</td>
			<td></td>
		</tr>
		<tr>
			<td>Triangular shaped beam</td>
			<td>NA</td>
			<td>NA</td>
			<td>Made out a piece of stiff foam</td>
		</tr>
		<tr>
			<td>Lashing straps</td>
			<td>NA</td>
			<td>NA</td>
			<td>Any lashing strap</td>
		</tr>
		<tr>
			<td>Surgical skin marker</td>
			<td>NA</td>
			<td>NA</td>
			<td>Any surgical skin marker</td>
		</tr>
		<tr>
			<td>Ultrasound transmission gel</td>
			<td>Servoson</td>
			<td>NA</td>
			<td>A sticky gel type is recommended</td>
		</tr>
	</tbody>
</table>

3d ultrasound imaging: fast and cost-effective morphometry of musculoskeletal tissue

The developmental goal of 3D ultrasound imaging (3DUS) is to engineer a modality to perform 3D morphological ultrasound analysis of human muscles. 3DUS images are constructed from calibrated freehand 2D B-mode ultrasound images, which are positioned into a voxel array. Ultrasound (US) imaging allows quantification of muscle size, fascicle length, and angle of pennation. These morphological variables are important determinants of muscle force and length range of force exertion. The presented protocol describes an approach to determine volume and fascicle length of m.&#160;vastus lateralis and m.&#160;gastrocnemius medialis. 3DUS facilitates standardization using 3D anatomical references. This approach provides a fast and cost-effective approach for quantifying 3D morphology in skeletal muscles. In healthcare and sports, information on the morphometry of muscles is very valuable in diagnostics and/or follow-up evaluations after treatment or training.

The described 3DUS technique was used to collect morphological data of the GM and VL in four male human cadavers, age at death 76.8 &#177; 7.9 years (mean &#177; SD). The cadavers were obtained via the donation program of the Department of Anatomy and Neuroscience of the Vrije Universiteit Medical Center (VUmc), Amsterdam, the Netherlands. The bodies were preserved using an embalming method aimed at maintaining the morphological features of the tissue39.
Prior to dissection, a 3DUS image was made of the GM and VL according to the methodology described. During dissection, skin, subcutaneous tissue, and fasciae overlaying the GM and VL were removed. A mid-longitudinal section was cut, taking the orientation of the distal aponeurosis into account. Using a caliper, the fascicle length was measured, halfway between the origin and the distal end of the muscle belly. Subsequently, after tenotomy, the muscle belly was dissected and submerged in a calibrated water column. Using ImageJ, the volumes were measured on photographs of the water column with and without the muscle belly, and muscle volume was calculate from the difference40. Fascicle length and volume were measured three times and the mean and standard deviation values were calculated. Criterion validity between the 3DUS method and dissection measurements was tested using a Pearson&#39;s correlation for average fascicle length and muscle volume. Intra-rater reliability of the 3DUS method derived fascicle length and volume measurements was quantified using a two-way mixed model intra-class correlation coefficient (ICC3,3)41, and after logarithmic transformation of the data, the coefficient of variation (CV) was calculated. The validity of fascicle length and muscle volume measurements were confirmed by significant and high correlations (r = 0.998, p &#60;0.01 and r = 0.985, p &#60;0.01, respectively). Intra-rater reliability of the 3DUS method derived measurements of fascicle length and volume was high (ICC3,3 0.983, CV 7.3%, and ICC3,3 0.998, CV 5.4%, respectively). It is concluded that the 3DUS approach presented is a valid and reliable tool for volume and fascicle length assessment of human VL and GM (Table 1).
<img alt="Table 1" src="/files/ftp_upload/55943/55943table1.jpg" /> 
Table 1:&#160;Cadaver Validation Data. C# is Cadaver number, GM is&#160;m. gastrocnemius medialis, VL is&#160;m. vastus lateralis. &#34;Dissection&#34;shows results from the cadaver dissection, and &#34;3DUS&#34; shows the results from the 3DUS image analysis of the cadavers.

Watch this Scientific Journal Video about 3D Ultrasound Imaging: Fast and Cost-effective Morphometry of Musculoskeletal Tissue at JoVE.com

3D Ultrasound Imaging: Fast and Cost-effective Morphometry of Musculoskeletal Tissue

3D ultrasound imaging (3DUS) allows fast and cost-effective morphometry of musculoskeletal tissues. We present a protocol to measure muscle volume and fascicle length using 3DUS.

The developmental goal of 3D ultrasound imaging (3DUS) is to engineer a modality to perform 3D morphological ultrasound analysis of human muscles. 3DUS ...

Research

JoVE Journal

Medicine

Murine Echocardiography and Ultrasound Imaging

Rodent models of cardiac pathophysiology represent a valuable research tool to investigate mechanism of disease as well as test new therapeutics.1  ...

A Novel Application of Musculoskeletal Ultrasound Imaging

Ultrasound is an attractive modality for imaging muscle and tendon  motion during dynamic tasks and can provide a complementary methodological  approach ...

Diagnostic Ultrasound Imaging of Mouse Diaphragm Function

Function analysis of rodent respiratory skeletal muscles, particularly the diaphragm, is commonly performed by isolating muscle strips using invasive ...

State of the Art Cranial Ultrasound Imaging in Neonates

Cranial ultrasound (CUS) is a reputable tool for brain imaging in critically ill neonates. It is safe, relatively cheap and easy to use, even when a ...

High-frequency Ultrasound Imaging of Mouse Cervical Lymph Nodes

High-frequency ultrasound (HFUS) is widely employed as a non-invasive method for imaging internal anatomic structures in experimental small animal ...

In Vivo Dynamics of Retinal Microglial Activation During Neurodegeneration: Confocal Ophthalmoscopic Imaging and Cell Morphometry in Mouse Glaucoma

In Vivo Dynamics of Retinal Microglial Activation During Neurodegeneration: Confocal Ophthalmoscopic Imaging and Cell Morphometry in Mouse Glaucoma

Microglia, which are CNS-resident neuroimmune cells, transform their morphology and size in response to CNS damage, switching to an activated state with ...

3D Whole-heart Myocardial Tissue Analysis

Cardiac regenerative therapies aim to protect and repair the injured heart in patients with ischemic heart disease. By injecting stem cells or other ...

Generation of Human 3D Lung Tissue Cultures (3D-LTCs) for Disease Modeling

Generation of Human 3D Lung Tissue Cultures 3D-LTCs for Disease Modeling

Translation of novel discoveries to human disease is limited by the availability of human tissue-based models of disease. Precision-cut lung slices (PCLS) ...

Use of 3D Robotic Ultrasound for In Vivo Analysis of Mouse Kidneys

Use of 3D Robotic Ultrasound for In Vivo Analysis of Mouse Kidneys

Common modalities for in vivo imaging of rodents include positron emission tomography (PET), computed tomography (CT), magnetic resonance imaging (MRI), ...

Drug Treatment by Central Venous Catheter in a Mouse Model of Angiotensin II Induced Abdominal Aortic Aneurysm and Monitoring by 3D Ultrasound

Since pharmaceutical treatment options are lacking in the clinical management of abdominal aortic aneurysm (AAA), animal models, in particular mouse ...