Anatomically realistic aortic dissection simulator as a potential training tool for point-of-care ultrasound
, , , and | Feb 27, 2024
Article Category: Research paper
Published Online: Feb 27, 2024
Page range: -
Received: Apr 18, 2023
Accepted: Aug 23, 2023
DOI: https://doi.org/10.15557/jou.2024.0002
Keywords
© 2024 Mutiah Rahmah et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Aortic dissection causes propagated damage (proximally or distally) due to blood entering the intima-media space(1). The mortality associated with aortic dissection (AD) is high; in Malaysia, the number of cases of aortic aneurysm which can cause AD is predicted to reach 15,299 cases among the total population of 23 million people(2). The two most common anatomic classification systems for AD are the Stanford and the DeBakey. The Stanford system is further classified into two affected areas, type A is in the ascending aorta, which is the case for most patients, but there are exceptions, as in type B, which involves the descending area. The DeBakey classification is divided into type I, type II, and type III. Type I dissection is affecting the entire aorta, type II involves the ascending area only, while type III is usually found in the descending area, excluding the ascending and aortic arch(3). Stanford type A and DeBakey type II, which include ascending areas in the affected area, are associated with higher mortality rates than type B. Symptoms such as chest pain have been reported in a higher percentage (85%) of type A than type B (67%) patients, but hypertension in type A patients’ condition occurs in fewer cases than type B. It causes AD patients type A to increasingly be transferred to the aortic specialist department for further assessment to determine emergent surgery needs(4).
Several types of imaging modalities are employed to diagnose AD, with US typically used in emergency departments because it is noninvasive and can diagnose disorders in a short time(5). The utility of ultrasound (US) encompasses a wide range of medical conditions involving the cardiovascular, respiratory, and vascular systems. US also shows high accuracy for the detection of aortic aneurysm (100%) if the diameter size is 3 cm(6). The advantage of US to utilise as a point-of-care ultrasound (PoCUS) is that it is easy for handheld devices, easy to perform the assessment, and not time-intensive(7).
Physicians lack practical experience and have fewer opportunities to acquire expertise in clinical settings directly in patients due to safety reasons(8). Therefore, nowadays, simulators are used for training and education purposes(9). The training simulation method can produce clinical scenarios, help physicians to encounter and respond to abnormal conditions, and reduce the potential for provider deficiencies during the assessment(8). Training simulators rely on 3D printing technologies for rapidly prototyping and converting digital objects into physical objects which were applied in this study to extrude a mould(10). In medical and clinical research, tissue mimicking materials (TMM) are usually used, including biopolymers (agarose, gelatin, and gellan gum) and synthesised polymers (polyvinyl alcohol (PVA), silicone, and polyvinyl chloride). The advantages of PVA include non-toxicity, biocompatibility, and biodegradability(11). The purpose of aluminium oxide (Al2O3) in tissue mimicking is to act as a scatterer during ultrasound assessments(12). Glycerol (C3H8O3) serves as a plasticiser and simulates better mechanical properties(13). Silicone owes its benefits to durability, stability over time, and strongly hydrophobic properties. Additive materials such as graphite have roles as a scattering agent, thinner to reduce viscosity, and slacker to provide a human tissue-like self-sealing property(14).
The present study aimed to create an ascending aorta (AAo) and DeBakey type II AD simulator synthesised by polyvinyl alcohol (PVA) and silicone (Si), and study its surface morphology and ultrasound features (diameter, wall thickness, and echogenicity) as a potential PoCUS training model.
The chemical materials used included PVA with a molecular weight of 89,000–98,000 and 99+% hydrolyzed (Sigma-Aldrich), aluminium oxide (Al2O3) (Sigma-Aldrich), glycerol (QRec), silicone eco flex 00-10 (Smooth-On), slacker (Smooth-On), thinner (Smooth-On), graphite (Sigma-Aldrich), and agarose (Promega). The equipment utilised included a 3D printer machine (ender3pro), TM3000 TableTop SEM (Hitachi), and ultrasound Toshiba Aplio 300. The software used included Blender, Cura, IBM SPSS statistic 27, and ImageJ.
The 3D image of healthy AAo and AD was acquired from an anonymous patient who was examined through CT scan provided by open source in
Fig. 1.
Final design of the mould in Cura software after the slicing process; the hours indicate the time estimation for completing the printing process and the weight mass (g) for the required filament.
The first simulator of PVA composition was made up of 5% (2.5 g) and then dissolved in 50 mL of distilled water. Aluminium oxide (2%) and glycerol (12%) were also added to the PVA solution(11). The mixed solution was then poured into the mould and solidified by using a freeze-thaw process. The freeze-thaw cycle was done by putting the mould into the freezer with a temperature −20°C for 12 hours and continuing to thaw in the room temperature for 12 hours. Then, the procedure was repeated for three cycles, which made up a total of 72 hours for the whole process(12). The second solution was composed of part A and part B of silicone in a 1:1 ratio. Then, 53% of part A + part B silicone, 15% thinner and 7% of slacker was added to reduce the viscosity of the silicone and as the property of self-sealing process(14). To enhance the echogenicity during US evaluation, 5% graphite was added as well. The solution was allowed to solidify at room temperature after mixing homogeneously. The demoulding process of the simulator was done manually to remove the simulator from the mould. To evaluate the simulator by US, a low-fidelity agar-based simulator was fabricated. A low-fidelity simulator was prepared by using distilled water (250 mL) and agarose (1%), then stirred for 30 minutes at 180°C until it completely dissolved. The agarose was allowed to cool down for three to five minutes before being poured into the container containing the simulators in order to prevent the melting of the PVA simulator. Finally, the agarose was allowed to solidify for an hour, and the simulator was ready for further testing.
The characteristics studied included morphology (surface roughness and 3D plot surface) and ultrasound features (diameter, wall thickness, and echogenicity measurement). The SEM test was conducted using a TM3000 TableTop system to evaluate the surface morphology of the simulators. The sample was prepared by cutting the healthy AAo and the AD simulator into 1 cm × 1 cm pieces with a thickness of not more than 5 mm at the ascending part, which was the area of interest. The SEM magnification was set to 100×, as a standardisation from the previous study, to allow for an ideal comparison between the results of this study(15). The samples were then coated with gold sputter to improve image quality.
Surface roughness was evaluated using imageJ, based on the SEM images. The roughness parameters were calculated using the root mean square (RMS), with (Rq) obtained from Fiji ImageJ and evaluated using the SurfCharj tools available in the windows menu. The images were converted to the 32-bit format and then sharpened to improve clarity. The data was analysed with the SurfCharJ tools from the plugins window(16). The RMS value was compared between the healthy AAo and AD for the purpose of quantitative analysis. The 3D surface plot was measured through a plugin in the windows of imageJ menu to analyse the intensity of light from a surface plot of each pixel in the SEM images(17). The surface plot profile in the menu shows the grey value of each pixel and the raw data used to analyse the variance by IBM SPSS statistic 27.
The simulators were assessed by using a phased-array cardiac transducer type PST-65 AT (6.5 MHz) from the Toshiba Aplio 300 ultrasound system. The probe was put above the low-fidelity simulator and started with the transverse and long-axis view assessment. The image was freeze-captured in the transverse view to visualise the aortic wall, and the distance between the inner and outer walls of the aorta was used to calculate the thickness (
The simulator synthesised by PVA appeared transparent, and white crystalline precipitate of Al2O3 was observed (Fig. 2) on the wall of the healthy and AD’s and also in the region of the flap. The rough surface on the ascending area of AD was visible due to the design of the mould.
Fig. 2.
Synthesised simulator based on PVA.
In addition, the synthesised Si was black in colour due to the addition of graphite (Fig. 3). The Si simulator had a more precise structure than the PVA simulator due to the higher density of silicone, which makes it easier to mimic the detailed shape of the mould.
Fig. 3.
Synthesised simulator based on silicone.
The results of the SEM test in Tab. 1 show the differences between the AAo of the healthy aorta and the dissected area. The healthy aorta simulator synthesised by both PVA and Si showed a uniform line on the surface of the ascending area. The abnormal condition in the dissect area of both synthesised simulators was disorganised surface structure. The SEM image of an actual aorta was used as a reference (healthy AAo from Pasta
Surface texture morphology of PVA and silicone simulators in SEM test (healthy AAo and AD compared to the actual aorta)
The RMS result of the healthy AAo simulator synthesised by PVA was
Root mean square (RMS) value obtained from (Rq) in imageJ between the healthy aorta (AAo) and AD simulator synthesised from PVA and silicone
| 10.26 | 23.07 | 13.11 | |
| 20.28 | 33.8 | 38.83 |
The 3D surface plot of SEM images shown in Tab. 3 consists of pixels or xyz-triples. The uniform line of the sharpening shape in the 3D plot of healthy AAo is consistently similar between the simulators and the actual aorta. The standard deviation was calculated, showing a small range between the AAo simulators
3D surface plot of SEM images analysed with imageJ between the healthy AAo and AD simulator (PVA and silicone) and the actual aorta
The surface plot shows the variance (
Surface plot of grey value in each pixel obtained by imageJ between the healthy AAo and aortic dissection, compared by simulators and the actual aorta
| 40.46 | 41.78 | 58.23 | 60.18 | 92.89 | 64.35 | |
| 52.01 | 59.62 | 66.78 | 91.54 | 98.23 | 109.59 | |
| 2.04 | 5.35 | 2.14 | 9.59 | 0.92 | 13.68 | |
| 4.16 | 28.57 | 4.56 | 91.97 | 0.84 | 187.36 |
The low-fidelity simulators of the healthy aorta and aortic dissection were immersed in the agarose, as shown in Fig. 4. The anatomical evaluations of both healthy AAo and aortic dissection simulator were conducted through the US in the UTM ultrasound clinic. The US image visualised the aorta with the echoic that exists around the simulator due to the agarose. The simulator itself was able to reflect the image and reveal the anatomically ascending area of the aorta in the transverse and long-axis views.
Fig. 4.
Low-fidelity simulators immersed in agarose
The PVA synthesised simulator was detected in the transverse view and shown in the monitor; then the measurement of the diameter (
The B-mode image of the US was analysed to identify the echogenicity of the intimal flap. In the transverse and long-axis views of the ultrasound image, the AD simulator was able to visualise the intimal flap due to the dissection and the false lumen, as shown with the arrow in Tab. 5. The intimal flap echogenicity was then measured using imageJ to calculate the average grey value (
Ultrasound images of the healthy aorta and aortic dissection simulators (PVA and silicone) and the actual aorta (comparison of short-axis and longitudinal views)
|
In the short-axis view, the diameter of the aorta is indicated by the letters A and B, while its thickness by the letter C. The arrow marks the intimal flap in the AD images. In the long-axis view, the healthy aorta images show the ascending aorta in its normal condition with no flap, whereas the AD indicated by the arrow shows the presence of the flap. |
||
Average grey value of the intimal flap (
| 170.89 | 85.27 | |
| 99.36 | 77.84 | |
The physical characteristics of the simulator’s surface structure imitated the ascending area of the human tissue. It was the results of printing a 3D mould that was designed based on the CT scan image obtained from the AD’s patient. The aortic dissection simulator appeared physically rougher on the separation surface than the healthy aorta simulator. To extract information about the surface structure and different roughness of the healthy aorta and the AD simulator, the RMS value was measured. The total surface roughness from the RMS value for the PVA-based AD simulator was (
The material used in this study has an impact on the echogenicity of ultrasound image. For example, PVA exhibits a high level of hydrolysis and aluminium oxide acts as a scatterer, which helps the simulator to reflect during the ultrasound assessment(12). The coated bubbles inside polymers that make small air pockets inside result in the echogenicity in the ultrasound image. In addition, the freeze-thaw method applied, which causes the presence of dissolved gas, also generates echogenicity by trapping air in the bilayer(20). The echogenicity of Si allows the simulator to be identified in US images due to the presence of an additive material (graphite), serving as a scatterer agent in Si(21). Additionally, the higher echogenicity expected for a soft-tissue which has a rougher or varied dimple surface rather than a flat surface(22).
During the US examination, the differences between the healthy AAo and AD simulators were able to be visualised. A comparison of the diameter and wall thickness between the simulator and the actual aorta is shown in the Tab. 4. The healthy adult aorta has
For the purpose of the reported study, we created in-house designed aortic dissection and healthy aorta simulators and tested them to determine whether they exhibited realistic anatomical sonographic features and thus had a potential for future applications in PoCUS training tools. The results from the US assessment of the synthesised simulators were comparable to the open source-retrieved US images and measurements done in actual patients. The AD simulators were able to replicate the intimal flap, which indicates hyperechoic properties that can be seen and differentiated between the healthy aorta and aortic dissection by using PoCUS. As a result, based on simulator characteristics that are relevant to the actual tissue, emergency medical physicians are able to plan and simulate actions that will be performed on patients using the simulator, as well as provide training tools for them to practise.