Introduction
Elevated intracranial pressure (ICP) is a serious condition in neurocritically ill patients.Reference Pickard and Czosnyka1 If not diagnosed and treated promptly, it has been shown to lead to secondary brain injury due to decreased cerebral perfusion pressure.Reference Schreiber2,Reference Badri, Chen and Barber3 Invasive intracranial monitoring is currently the gold standard for diagnosis of intracranial hypertensionReference Carney, Totten and O’Reilly4; however, it can be expensive, requires neurosurgical expertise, and is associated with a risk of bleeding and infection.Reference Ragland and Lee5
Sonographic assessment of optic nerve sheath diameter (ONSD) as a non-invasive means to assess ICP has become increasingly popular in the literature.Reference Robba, Santori and Czosnyka6 Acute increases in ICP have been shown to strongly correlate with increases in ONSD.Reference Geeraerts, Merceron, Benhamou, Vigué and Duranteau7–Reference Chen, Wang, Hu, Jiang, Wang and Xing10 Recently, ultrasound measurement of ONSD has been investigated for use as a dynamic real-time ICP monitoring tool in critically ill patients with promising results.Reference Chen, Wang, Hu, Jiang, Wang and Xing10 Ultrasound itself is becoming increasingly available due to decreasing costs, the production of portable pocket ultrasound systems, and the dissemination of ultrasound expertise. As access to ultrasound spreads, sonographic assessment of ONSD will continue to have increasing uptake as a diagnostic test. Ultrasound is particularly useful in less well-equipped and austere environments where other tools such as invasive monitors and neurosurgical expertise, CT scanners, or MRI are not readily available, but where rapid assessment of ICP is still essential. Furthermore, ultrasound can be used as an adjunct to the physical exam in order to help justify the use of more expensive tests. As with any ultrasound measurement, ONSD sonography has been shown to be operator dependent with small to moderate inter-observer and intra-observer variation reported in the literature.Reference Ballantyne, O’Neill, Hamilton and Hollman11–Reference Oberfoell, Murphy, French, Trent and Richards16 It has been previously shown that this variance can be improved with practice.Reference Potgieter, Kippin, Ngu and McKean15,Reference Zeiler, Ziesmann and Goeres17
There is growing evidence that simulation is effective for developing and assessing both technical and non-technical skills.Reference Munshi, Lababidi and Alyousef18 It allows for practice in areas of weakness, rarity, critical importance, or procedures that are uncomfortable to the patient. However, simulation can be very elaborate or cost prohibitive, thus defeating its purpose of increasing opportunities to practice otherwise inaccessible skills. Previously, we have described a standardized model using a gelatin globe with a 3D printed disc to generate a simulated ONS of consistent dimensions in order to test inter- and intra-observer variability.Reference Zeiler, Unger and Zhu12,Reference Johnson, Zeiler, Unger, Hansen, Karakitsos and Gillman14,Reference Zeiler, Unger, Kramer, Kirkpatrick and Gillman19 However, as this measurement technique relied on a shadow generated by the disc that may be distinct from the size of the disc itself, we did not have an adequately reliable physical parameter to measure, in order to directly compare our sonographic ONS measurements. Here, we describe the design and development of a new model to simulate the ONS using cost-effective tubes and demonstrate that it can be used to acquire ultrasound images that can be measured accurately and are similar in appearance to those acquired from in vivo sources. This model can be used not only as a low-cost instructional tool to teach the method of ONSD measurement but also has the potential for use as a research tool for determining accuracy in measurement of predetermined ONS sizes in a reproducible and controlled setting.
Materials and Methods
ONS models were constructed as an improvement to previously described methods.Reference Zeiler, Unger, Kramer, Kirkpatrick and Gillman19–Reference Murphy, Oberfoell, Trent, French, Kim and Richards21 They are made from: 237 ml (8 oz) white disposable coffee cups (Dart Container Company, Mason, MI, USA), white glue (Elmer’s School Glue, Elmer Products, Inc. Highport, NC, USA), straight pins, unflavored gelatin (Knox-brand, E.D. Smith Foods, Ltd., Winona, ON, Canada), sugar-free psyllium powder (Metamucil Sugar Free Dietary Fiber Supplement, Procter & Gamble, Cincinnati, OH, USA), and assorted sizings of synthetic polymer medical tubing. We used four different types of tubing in seven diameters obtained from excess/expired material in our intensive care unit to simulate the ONS: a 6 mm disposable portex tracheostomy inner cannula (Smiths Medical, Minneapolis, MN, USA), 4.2 mm sheath from a spring-wire guide (Arrow brand, Teleflex Inc., Wayne, PA, USA), 3.5 mm polyvinyl IV tubing (B. Braun Medical Inc., Englewood, NJ, USA), and various sizes of gastric sump tubes (Salem sump tubes, Medtronic, Dublin, Ireland). Tubing was selected for their varying diameters simulating the range of ONSD found in nature (3–8 mm).
First, to simulate the ONS, the medical tubing was cut into a 1-cm long segment and placed with the long axis perpendicular to the base of the coffee cup. A small pin was placed through the bottom of the cup vertically through the center of the tubing in order to keep it in place. Next, the bottom end of the tubing was fixed to the cup base with a generous portion of glue and allowed to dry for an hour with the pin in place to keep the tubing upright. After the glue had dried, the pin was removed from the base of the coffee cup with the tubing now suspended upright by the glued bottom (Figure 1). The resultant pinpoint hole in the bottom of the cup was covered externally with a piece of tape. Using a variation of a previously described recipe for gelatin ultrasound phantoms,Reference Bude and Adler20 we suspended three packets of gelatin in 100 ml of room-temperature water. We brought a separate 150 ml to a boil and gradually mixed in the gelatin suspension with the boiling water over medium heat until the gelatin was completely dissolved (1–2 min). Next, one tablespoon of psyllium powder was added and whisked slowly until it completely dissolved (~1 min). Using this method reduces bubbles or clumping of the gelatin compared to Bude and Adler’s originally described method where the gelatin packets were added directly to the boiling water.Reference Bude and Adler20 Finally, the hot mixture was poured slowly into the coffee cups to a few millimeters above the upper level of the tubing and placed in the refrigerator for at least 1 h to allow the gelatin to congeal. Models could be stored in sealed sandwich bags or covered plastic containers in a refrigerator at 4 °C for up to 3 weeks before desiccating or becoming moldy.
Figure 1: Model making process. (A) Intravenous tubing held upright by pin and glued to the base of a coffee cup. (B) Nasogastric tube after pin is removed held upright by dried glue alone.
As an optional step in order to simulate the optic globe above the ONS, the molten gelatin mixture can be poured into a 1-inch spherical baking mold (Chicago School of Mold Making, Oak Park, IL, USA) as previously described.Reference Zeiler, Unger, Kramer, Kirkpatrick and Gillman19 A semi-circular defect can then be carved into the hardened gelatin above the tubing in the coffee cups, lubricated with a generous portion of ultrasound gel, and the gelatin balls placed in the defect (Figure 2).
Figure 2: Finished models of tubes suspended in gelatin and psyllium (A) with optional gelatin eyeball and (B) without eyeball.
The models were evaluated utilizing a 13–6 MHz linear array ultrasound transducer (L25x transducer, Sonosite Corp., Bothell, WA, USA) and a portable ultrasound (Sonosite M-Turbo, SonoSite Inc., Bothell, WA, USA) using the ophthalmic manufacturer preset. Measurement of ONSD was conducted using the “black-stripe” method described in the literature. This is performed by placement of the probe over the orbit and identification of the ONS as a black line behind the globe. This black line is the optic nerve, and its diameter is measured 3 mm behind the optic disc (Figure 3).Reference Geeraerts, Merceron, Benhamou, Vigué and Duranteau7–Reference Soliman, Johnson and Gillman9,Reference Geeraerts, Launey and Martin22,Reference Blaivas, Theodoro and Sierzenski23 Measurements were conducted by three expert point-of-care ultrasonographers (LG, TJ, AD) in a single session. The diameters were 3.5 mm (polyvinyl intravenous line tubing), 4 mm (nasogastric tube), 4.2 mm (spring wire guide sheath), 4.6 mm (nasogastric tube), 5.5 mm (nasogastric tube), 6 mm (tracheostomy cannula), and 7.9 mm (nasogastric tube). Sonographers measured each model in random order a total of three times. Operators were blinded to the identity of the model, to the actual diameters of the tubing, and to each other’s measurements.
Figure 3: Ultrasound images of model optic nerve sheaths (left) compared to in vivo human eye (right). (A) Simulation model made from 3.5 mm polyvinyl intravenous tube without gelatin globe, (B) in vivo human eye close-up of optic nerve with measurement technique highlighted. Optic nerve sheath diameter is taken 3 mm behind the globe. (C) 16 French nasogastric tube model with gelatin globe, (D) entire in vivo image of 5 mm human ONS.
All statistical analyses were performed using SPSS Statistics version 25 (IMB, Armonk, NY, USA). The measurements obtained by the ultrasound devices were compared to the physical diameter of the tubes as measured with digital calipers by a separate investigator (GJ) and their measurement error was calculated.
Results
Sonographic images of our models shown next to an in vivo image of ONSD from a live person are shown in Figure 3. Measurement error (observed by ultrasound minus predicted by caliper measurement) for each model size is summarized in Figure 4. The mean error across all measurements is 0.10 mm and the error variance is 0.26 mm.Reference Schreiber2
Figure 4: Box and whisker plot of measurement error (observed minus predicted) stratified by predicted size of model optic nerve sheath diameter as measured by digital calipers. o = outlier, * = extreme outlier.
Discussion
Our simulation model for sonographic ONSD measurement provides a low-cost realistic opportunity for trainees to practice this relatively difficult task without discomfort to a patient. The images obtained of the tubing by ultrasound can be measured and provide realistic simulation of the ONS (Figure 3). Furthermore, these measurements can be made with a low degree of error (Figure 4), similar to those previously published.Reference Zeiler, Unger and Zhu12
Simulation of the orbit and its surrounding structures has been described in the literature in various models in the past. To our knowledge, there are only two models that simulate the ONS.Reference Zeiler, Unger, Kramer, Kirkpatrick and Gillman19,Reference Murphy, Oberfoell, Trent, French, Kim and Richards21 In a previous study, our group used a gelatin globe with a 3D-printed disc that generated a sonographic “shadow” to simulate the ONS. While this produced good visual results, and the disc could be created in any desired dimensions, the shadow was not necessarily the same size as the disc itself and served as a poor standard.Reference Zeiler, Unger, Kramer, Kirkpatrick and Gillman19 Printed discs required access to a specialized 3D printer, which is cost prohibitive, limiting the utility of this model as a low-cost training aid. Our current model addresses both of these deficiencies. Medical tubing was used, as excess is readily salvageable from hospital supplies (e.g., expired, opened, used, etc.) at low cost. The diameter of the simulated ONSD is a direct measurement of the tube reflection itself, not its ultrasound shadow, so there is an objectively verified diameter for comparison. The tubes come in a variety of sizes relevant to ONSD measurement.
More recently, Murphy et al. used a water-filled ping-pong ball and vinyl tubing submerged in gelatin to simulate the orbit and ONS. While this method produced excellent still images that were indistinguishable from in vivo images by emergency medicine residents, it had a number of limitations. Very low gain was needed in order to minimize reverberation artifact from the ping-pong balls.Reference Murphy, Oberfoell, Trent, French, Kim and Richards21 This was inconsequential when study personnel obtained still photographs. However, if this model were to be used as a training aid where novices would be expected to attain the images themselves, this could decrease fidelity of the simulation. While not difficult to learn,Reference Potgieter, Kippin, Ngu and McKean15,Reference Zeiler, Ziesmann and Goeres17 we believe that a significant challenge of sonographic determination of ONSD is acquiring the best image quality. Image acquisition skills must be demonstrated and practiced along with image interpretation skills. Moreover, using a gelatin globe instead is inexpensive, easier to produce, and has less reverberation artifact than the ping-pong ball method. For even clearer images, psyllium can be removed from the recipe as previously described,Reference Zeiler, Unger, Kramer, Kirkpatrick and Gillman19 and construction of a spherical globe can be abandoned, allowing one to practice measuring the diameter of a tube suspended in gelatin alone.
Murphy et al. also used polyvinyl tubing to simulate the ONS.Reference Murphy, Oberfoell, Trent, French, Kim and Richards21 Search from various manufacturers however shows the smallest diameter readily available is approximately 6 mm, which only simulates the upper end of possibilities of ONSD found in vivo. Medical tubing such as that used in our model comes in a large range of diameters across the normal and abnormal ONS measurement spectrum.
Further benefits to our model as an educational tool include the potential ability to assess learners’ progression as they develop the skill of ONSD measurement. The diameter of the tubing used in our models can be preselected to simulate a variety of normal or pathologic ONSD. The actual physical diameter of the tubing is measured rather than a shadow; thus, an instructor can know the actual diameter of the ONS by pre-measuring it in advance in order to more objectively assess a learner’s ability to determine the diameter. Previously, agreement with expert sonographer,Reference Potgieter, Kippin, Ngu and McKean15 agreement with other imaging modalities,Reference Ohle, McIsaac, Woo and Perry24,Reference Bäuerle, Schuchardt, Schroeder, Egger, Weigel and Harloff25 or a decrease in intra-observer variation over timeReference Zeiler, Ziesmann and Goeres17 would be the only ways to monitor learner progression of ONSD measurement.
Although performing ocular ultrasound on humans is safe,Reference Shankar and Pagel26 phantom models such as ours are convenient, easily accessible, and allow for prolonged scan time without risk of discomfort or injury to the retina. This could be particularly important for novice sonographers who may be developing their skills in image acquisition and adjusting device settings. Furthermore, practice on a low-cost model can allow trainees to perform dedicated practice of this technique on their own time, which may be important for skills acquisition and maintenance given strict duty hour restrictions imposed on trainees in many jurisdictions.Reference Nataraja, Webb and Lopez27 In addition, having an ONSD of a known size allows for an evaluation of competence and accuracy of trainee measurements. While volunteers can be free and even more realistic, they are often in short supply and could be reserved for those trainees that have already gained some degree of competence on a practice model.
Although standard portable ultrasound units have been extensively tested and have been determined to be reasonably accurate for measurement of ONSD and determination of ICP,Reference Ballantyne, O’Neill, Hamilton and Hollman11–Reference Bäuerle, Lochner, Kaps and Nedelmann13,Reference Potgieter, Kippin, Ngu and McKean15,Reference Oberfoell, Murphy, French, Trent and Richards16 newer handheld units are continuously being added to the market. New pocket ultrasound devices are not identical to previously validated portable ultrasound, nor are they identical to the pocket Vscan unit that we have previously found to be accurate.Reference Johnson, Zeiler, Unger, Hansen, Karakitsos and Gillman14 These ultrasound units can have small screen sizes, different probe frequencies, employ new touch screen technology and software presets than machines that have been previously validated for, and are currently not approved for ocular use.28–30 These differences can be reasonably assumed to affect how accurately an operator can measure a subcentimeter structure such as the optic nerve sheath. Our simulation model with known reproducible sizes of ONSD could be an invaluable tool for testing novel ultrasound units as they hit the market. This could provide manufacturers with the rationale to apply for and implement the required safety presets onto their pocket devices so that they could be used for this use in the future. Sonographic assessment of simulated ONSD can be compared for agreement with actual tube diameter to determine the accuracy of measurement of new units as they become available.
Despite the expected benefits of our novel model, there are some limitations and areas for future improvement. Most notably, our model can only simulate the popular “black stripe” method for measurement of OSND. Recently, new criteria have been proposed to improve image quality of sonographic measurement of ONSD and standardize measurements across different observers and scans.Reference Sargsyan, Blaivas, Geeraerts and Karakitsos31 Early data suggest that these criteria may lead to a change in normal values for ONSD measurement.Reference Soliman, Johnson and Gillman9 These criteria emphasize bilateral ocular sonography and avoidance or identification of landmarks within and around the globe, which are not simulated in our model. To our knowledge, no model to date has been developed which is able to simulate all structures necessary to apply the new quality criteria.
Our model in its current state lasts approximately 3 weeks if kept in a sealed container in a refrigerator. If kept for longer, or improperly stored, it will desiccate rapidly or become colonized with fungus. Alternative materials such as agarose, ballistic gelatin, silicone, polyacrylamide, polyvinyl alcohol, polyurethane, oil gel, animal and human cadaveric tissue, and others have all been studied,Reference Culjat, Goldenberg, Tewari and Singh32 but none are accessible by a simple trip to the grocery store (in the case of knox gelatin and psyllium) and many are more expensive and require specialized equipment or a laboratory environment to make. Further study on easily accessible preservatives that can be added to the gelatin without affecting its sonographic properties or user safety in order to extend the life of the model would be helpful, though preservation may ultimately not be important with affordable and easily reproduced models.
Conclusions
Our simulation model for sonographic ONSD measurement provides a low-cost option to practice this relatively difficult task. The images obtained of the tubing by ultrasound can be measured accurately, with low error, and provided realistic simulation of the ONS for purposes of training or further research.
Ethics Approval and Consent to Participate
This study was deemed not to require research ethics board review by the University of Manitoba Health Research Ethics Board.
Statement of Authorship
All authors were involved in study design, data analysis, and manuscript composition. LMG, GGRJJ, TJ, and AD were involved in data collection. All authors read and approved the final manuscript.
Availability of Data and Materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Funding
FZ’s research program is supported through the University of Manitoba Thorlakson’s Chair in Surgical Research Establishment Grant, University of Manitoba VPRI Research Investment Fund (RIF), Winnipeg Health Sciences Center Foundation, and the University of Manitoba Rudy Falk Clinician–Scientist Professorship.
BU’s research is funded in part by unrestricted grant from advanced bionics which is being administered in conjunction with Mitacs.
This research was supported by a General Operating Grant from the Health Science Centre Foundation and a GFT Research Grant from the Department of Surgery, University of Manitoba, both in Winnipeg, Manitoba, Canada.
Disclosures
The authors have no conflicts of interest to declare.
