• Mission           
• Laboratory      
• Facilities &       Resources      

• Intravascular Imaging
• Molecular Imaging
• Imaging at Microscopic Resolution        
• Laser-Induced Cavitation Bubble    
• Deep Vein Thrombosis
• Image-Guided     Photothermal Therapy
• Magneto-Acoustic Imaging           
• Multi-Modal Imaging
• Real-Time Elasticity Imaging          

• 2009               
• 2008               
• 2007               
• 2006               
• 2005              
• 2004               
• 2003              

• Current         Members       
• Collaborators
• Industrial         Partners        
• Alumni            
• Undergraduates

• Journals         
• Databases & Indexes        
• Societies & Conferences
• University of Texas            
• Austin &         
  Texas         

Ultrasound-based Multifunctional Imaging to Detect and Stage Deep Vein Thrombosis

Most clinically important pulmonary emboli (PE) originate from proximal deep vein thrombosis (DVT) of the leg (popliteal, femoral, or iliac veins). As the population ages, the medical and economic impact of DVT and associated PE is expected to increase, with one of every hundred cases being fatal – most often the death is related to PE where part of DVT breaks free, travels through the veins and reaches the lungs. Treatment goals for deep venous thrombosis are to stop clot propagation and to prevent the recurrence of thrombus and the occurrence of pulmonary embolism. Currently, all imaging methods to assess DVT including venous ultrasonography, contrast venography, and MRI cannot distinguish between an old (chronic) clot and a new (acute) clot. Therefore, there is a need for a definite and urgent clinical need for a technique that is widely available, is simple to perform, is safe, and can reliably diagnose and adequately stage DVT. We are developing a multifunctional, ultrasound-based method for remote, reliable and non-invasive detecting and staging of DVT.

An acute clot, formed in a relatively static environment, tends to contain more enmeshed erythrocytes and is, therefore, known as a red or stasis thrombus (Fig. 1, top). As the blood clot matures, it is typically firmly adherent to a vessel wall and becomes gray-white-yellow, composed of a tangled mesh of platelets, fibrins, and degenerating leukocytes. Such reorganization of the clot may result in decrease of a hemoglobin concentration and, therefore, reduced optical absorption of DVT (Fig. 1, middle and bottom). Therefore, the photoacoustic imaging – an imaging technique capable of remote measurements of optical absorption of tissue – could be used to stage the detected thrombi. Photoacoustic imaging relies on absorption of a short pulse of an electromagnetic radiation and a subsequent emission and detection of an acoustic wave.

A round or cylindrical object with even distribution of optical absorption, irradiated by a short pulse of laser light, generates an N-shaped photoacoustic radiofrequency (RF) signal. An ultrasound transducer acts as band-pass filter with a finite central frequency and bandwidth modifying this broadband RF signal. Therefore, a resulting signal will outline only the boundaries of the absorber.10 Since the magnitude of a photoacoustic response is proportional to the light absorption coefficient, an acute blood clot should emit stronger ultrasonic signals compared with a chronic DVT. Indeed, the magnitude of photoacoustic signal monotonically decreases with the reduced concentration of red blood cells (RBC) suggesting that photoacoustic imaging has the potential to stage DVT.

To test the photoacoustic imaging technique initially, the artificial clots mimicking two borderline cases –acute and chronic clots were imaged. These clots were made out of 7% polyvinyl alcohol (PVA) with 1% silica particles to achieve ultrasound scattering in clots. PVA was dissolved in saline to mimic the chronic clot and in solution of RBC to mimic an acute clot. The same RBC solution was also used to mimic whole blood in the remaining vein. The 10% gelatin with 0.8% silica particles (ultrasound scatterers) and 20% low-fat milk (optical scatterers) mimics environment tissues.

A focused single-element ultrasound transducer with the central frequency of 7.5 MHz, the F-number of 4, the focusing distance of 2 inches, and the bandwidth of 60% was used in experiments for both ultrasound and photoacoustic imaging. A pulsed laser operated at 532 nm was employed as light source of 5 ns pulses. The photoacoustic and ultrasound images of the same cross-section of phantoms with artificial clots are shown in Fig. 3 and Fig. 4. The phantoms were illuminated from bottom while the ultrasound transducer was placed on the opposite side of phantoms. Each image consists of the plotted together corresponded RF signals after Hilbert transform and rescaling.

We can easily identify the background tissue and the clots on the ultrasound images [Fig. 2(a)]. The artificial vein is represented by two longitudinal lines along the images. The RBC solution looks black since erythrocytes do not scatter ultrasound at this relatively low frequency. The differences between ultrasound images of these two phantoms are attributed to differences between the phantoms and the imaging planes. Overall, the acute and chronic clots cannot be distinguished in ultrasound image.

However, acute and chronic clots can be easily identified in photoacoustic images of the same cross-sections of these phantoms shown in Fig. 2(b). The RBC concentration in an acute clot is close to the rest of the veins filled by RBC solution and the strong photoacoustic signal is generated in both the acute clot and the hypoechoic lumen of the vessel so the photoacoustic imaging cannot differentiate between them. In contrast, the photoacoustic response in the area of the chronic clot is reduced, up to disappearance – this is due to almost no optical absorption in the chronic clot. The top vein-tissue interface is darker due to light attenuation in the clot and in the RBC solution – the boundary exposed to a greater light intensity exhibits the photoacoustic signal of a larger magnitude. Overall, a chronic clot produces a relatively small photoacoustic response in comparison with an acute one.

The developed approach was evaluated on 3-day and 9-day rat clots (Fig. 1) embedded together in similar phantom and imaged across. An anatomy of this phantom is visualized by the ultrasound image [Fig. 3(a)]. Diameters of the 9-day clot located on the left and the 3-day clot located on the right measured 3.5 mm and 2 mm respectively in imaged cross-sections. The dashed circles show the positions of the clots in the phantom. The horizontal line in Fig. 3(a) represents the boundary between two layers of gelatin.

The photoacoustic image of the same cross-section of the phantom is shown in Fig. 3(b). These two clots are completely different. As expected, the 3-day clot is brighter indicating the higher light absorption than the 9-day clot. A photoacoustic profile shown in Fig.6 makes it clear. Each point in Fig.6 corresponds to maximum value of RF signal after Hilbert transform. Magnitude of photoacoustic signal from younger clot is more than ten times greater. In addition, signal from older clot is greater than the apparatus noise defined from this profile in the area between clots.

The younger clot in Fig. 3(b) is represented by two photoacoustic signals corresponding to its boundaries. It indicates the more or less even distribution of light absorbers in a 3-day rat clot. In contrast, the photoacoustic signal is generated inside of a rat 9-day clot signaling the presence of separated absorbing areas in the bulk of a clot surrounded by non-absorbing tissue. Note, that the boundaries of clot are still visible. Finally, the smaller 3-day clot looks bigger than 9-day clot which is attributed to the relatively wide directivity diagram of the transducer and modest blood proliferation between gelatin layers.
In addition to changes of optical properties the maturation of thrombi results also the changes in its biomechanical characteristics. In the early stages of thrombosis, endothelial cell damage exposes collagen attracting a platelet plug. Soon fibrin appears forming a meshwork that traps blood formed elements. It is well accepted that DVT harden as they age and, therefore, ultrasound-based elasticity imaging assessing the hardness of tissues directly is considered as a very promising technique for estimating thrombus age.

Elasticity imaging is a reconstructive imaging technique where tissue motion in response to mechanical excitation is measured using modern imaging systems, and the estimated displacements are then used to reconstruct the spatial distribution of Young’s modulus. Usually hard tissue regions are less deformed in comparison with soft regions and, therefore, contrast in strain images is influenced by the tissue elasticity. Using inverse problem formulations, the elasticity (Young’s modulus) distribution in tissue is evaluated based on the distribution of the strain tensor components. In ultrasound-based elasticity imaging motion of the speckle is tracked over large range of tissue deformation.

An ultrasound imaging of rat in thrombus area with 5-day clot was performed using a Siemens “Elegra” ultrasound scanner with a linear 12MHz array transducer (VFX13-5). The ultrasound image of this area is shown in Fig. 4 (left). The area of interest is underlined by dashed square and includes vein, part of artery and environment tissues. To compress the rat’s abdominal wall and underlying tissue in a transverse orientation the transducer itself was used. It was attached to a manual deformation device to achieve continuous compression. The deformation lasted approximately 6 seconds, while phase sensitive ultrasound frames were collected in real time. Consecutive frames were processed offline using a 2D correlation-based phase-sensitive speckle tracking algorithm to derive the strain image of the DVT and surrounding tissue. Based on this data an iterative elasticity reconstruction was performed for the area of interest and corresponding strain image is shown in Fig. 4 (right). An artery and vein contents are the most deformable tissues appearing black. On a contrary, surrounding tissues are harder and less deformable appearing white. Corresponding stress in Fig. 4 indicates a significant, up to 18%, deformation of clot tissue compared with close environment.

Based on these data, the distribution of Young’s modulus can be restored using model-based reconstructive elasticity imaging procedure. In this animal model, 3-day to 4-day old thrombus represents an acute DVT, 6-day to 7-day old thrombus represents a subacute DVT, and 10-day old thrombus represents a chronic DVT. As the thrombus develops, its elasticity increases. The results of in-vivo elasticity reconstruction were compared with the results of ex-vivo direct elasticity measurements. A total of 59 Sprague-Dawley rats were studied. On days 3, 6, 10, 12, and 14, the group of animals (9–13 animals) was euthanized for the direct measurements of Young’s modulus of the blood clots. Thrombi were removed from the dissected vein and transported immediately to the next room for mechanical load-displacement measurements to accurately assess the Young’s modulus of the thrombi.

The rate of thrombus growth is different in humans, but the process of clot formation has the same stages. The ultrasound image of 7-day right femoral vein thrombus in a 77-year-old female patient is shown in Fig.5 (left). The ultrasound image discovers the anatomy of artery, clot and surrounding tissues. The corresponding normalized strain image of the same area is shown in Fig.5 (right). The color in each pixel in the thrombus represents the magnitude of the strain at that point divided by the average strain between the skin surface and the back wall of the vein. The higher the strain magnitude corresponds to the softer the region. Presence of black areas in clot image indicates that a 7-day human thrombus (acute) is soft and well-deformable.

Overall, the results of our studies indicate that the model-based reconstruction approach is applicable in ultrasound in-vivo elasticity imaging and can provide needed information about biomechanical properties of tissues. The proposed approach was tested both on animal model and clinically, and results were in excellent agreement with direct measurements of Young’s modules ex-vivo.

In the future, a calibration method to measure absolute values of Young’s modulus in-vivo has to be developed. In addition, the elasticity imaging has to be combined with photoacoustic imaging which would require developing a new ultrasound array with embedded light delivery system and allowing to press tissues precisely.

 

Our Publications

A.B. Karpiouk, S.R. Aglyamov, S. Mallidi, W.G. Scott, J.M. Rubin, S. Y. Emelianov, “Combined ultrasonic and photoacoustic imaging to age deep vein thrombosis: preliminary studies,” proceedings of IEEE Ultrasonics Symposium, 399-402, 2005 PDF

J.M. Rubin, H. Xie, K. Kim, W.F. Weitzel, S.Y. Emelianov, S.R. Aglyamov, T.W. Wakefield, A.G. Urquhart, M. O’Donnell, “Sonographic elasticity imaging of acute and chronic deep venous thrombosis in humans,” J Ultrasound Med 25, 1179–1186 (2006) PDF

S.R. Aglyamov, A.R. Skovoroda, H. Xie, K. Kim, J.M. Rubin, M. O’Donnell, T.W. Wakefield, D. Myers, S.Y. Emelianov, “Model-based reconstructive elasticity imaging using ultrasound,” International Journal of Biomedical Imaging, Article ID 35830, Vol. 2007, 11 pages (2007) PDF

A.B. Karpiouk, S.R. Aglyamov, S. Mallidi, J. Shah, W.G. Scott, J. Rubin, and S.Y. Emelianov, "Combined ultrasound and photoacoustic imaging to detect and stage deep vein thrombosis: phantom and ex vivo studies," Journal of Biomedical Optics, 13(5), 054061, (2008) PDF