Functional requirements

In a collaborative process with medical experts of the Klinik und Poliklinik für Kardiologie, Universitätsklinikum Leipzig, physicists from the Gesellschaft für Angewandte Medizinische Physik und Technik (GAMPT mbH, Merseburg, Germany) and scientists from the Innovation Center Computer Assisted Surgery (ICCAS, Leipzig, Germany), a system for automated measurement of blood flow in the CCA during resuscitation scenarios was proposed. The system should provide the following features:

• The device needs a complete hands-free approach for sensor attachment to the neck.
• After attaching the sensor, a fast detachment, repositioning and reattachment should be possible.
• Sensor positioning on the neck should be easily relatable to neck landmarks for documentation and comparable repositioning.
• Possible positions should cover most of the frontal neck area of one side (left or right).
• Attachment pressure should be applicable in an adjustable manner.
• No adhesive should be used to ensure sensor‑skin contact.

System setup

Towards that purpose, we developed a new addition to an already existing neck brace system, the Laerdal Stifneck Select Collar™ (Laerdal Medical GmbH, Stavanger, Norway) (Fig 1).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. A specially developed ultrasound system (GAMPT mbH) with Stifneck Select Collar™ (Laerdal) for sensor attachment.

a) Left: US probe inside sensor mount on rail, fixed to the neck brace. The image is a cutout of the image area of the larger image to the right, with the cutout region marked with green lines. Right: System with open Stifneck Select Collar™ and inserted US probe and the processing unit in the background; b) System applied to a dummy with simulated measurements.

https://doi.org/10.1371/journal.pone.0358589.g001

The final system for automatic acquisition of carotid blood flow consists of a Stifneck Select Collar™, a custom 3D-printed sensor mount, a proprietary ultrasonic probe and a small computer with three ultrasonic control modules. The probe is fixed on a custom-made sensor mount, which in turn is fixed on the Stifneck Select Collar™, both shown in Fig 1a. Fig 1b shows the Stifneck Select Collar™ applied to a phantom for resuscitation exercises, the Resusci Anne® (Laerdal Medical GmbH, Puchheim, Germany). The sensor mount’s components are the sensor holder, an inflatable pillow (TR-Band™, Terumo, Elkton, MD, USA), a rail and the rail holder. The sensor holder can fit a sensor with a base area of 25 mm x 41 mm. The sensor used in this setup has a height of 15 mm and two additional extensions at one end that fit under grooves in the holder and ensure the sensor does not fall out. The casing of the probe is made of disinfectable polymer with rounded corners and edges, to minimize risks of tissue damage.

For security measures, the cables of the sensor probe are held by a cable relief opposite the sensor mount, consisting of a strain relief bracket and the strain relief fixture. The sensor mount rail can be adjusted in its position, enabling the user to change the position of the sensor probe along the neck of the proband. To account for different-sized necks and adjust the necessary contact pressure, the pillow below the sensor holder can be inflated up to 18 ml, pressing the sensor probe towards the neck.

The ultrasonic probe used for measurements contains three different piezo-ceramics. Each ceramic has a different angle towards the probe surface, which ultimately connects to the proband’s skin. The ultrasonic angles, results of the ceramic’s orientation, toward the probe surface’s normal are 25°, 0° and −12°. The probe is connected to the computer responsible for ultrasonic transmission and reflection measurement at each transducer ceramic. Besides the main computing unit, it features three ultrasound modules, one for each ceramic. The modules have synchronized communication, thus enabling the main unit to trigger an ultrasonic pulse via one ceramic by its respective ultrasound modules and measure the response of targeted tissue via all ceramics/modules [23,24].

The probe is secured with a specially designed holder printed from polylactic acid (PLA). The material is characterized by high stability and low water absorption and is resistant to most common disinfectants. To enable repeatable measurements due to taking off the collar for medical interventions or changing the position of the sensor probe, the sensor can be adjusted along a rail (Fig 2a). The rail features markings, enabling documentation of the sensor position in relation to the holding add-on, which is in a fixed place on the neck brace. To ensure sufficient contact between the skin and sensor surface, the sensor is pressed against the skin surface using a small, inflatable air cushion between the probe mount and sensor rail (Figs 2b and 2c).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Neck brace features for comparable repositioning and sensor pressure.

a) sensor rail with notches for fixation on the sensor holder via an extra bolt, 5 mm increments in notch difference allow for comparison of sensor position between patients and attachments; b) pillow (TR-Band™) for pressure build-up between sensor rail and sensor backside in deflated state; c) pillow (TR-Band™) in completely inflated state.

https://doi.org/10.1371/journal.pone.0358589.g002

Study

The study was conducted with healthy volunteers between 10/10/2023 and 28/02/2024. Study conductance was in accordance with the 1964 Declaration of Helsinki and received approval from the local ethics committee of the Leipzig University, Medical Faculty (reference number 183/23-ek). Volunteers provided written informed consent before participating in the study.

Hemodynamic stability and absence of cardiovascular symptoms were ensured at the time of measurement. Vital signs were recorded before the measurement. Exclusion from the study was limited mobility of the cervical spine, neck area injuries, or a known high-grade stenosis of the internal carotid artery.

During the study, the sensor contact surface was covered with ultrasound gel before each measurement. For measurements, the probands were in a supine position on an examination bed. After putting the Stifneck Select Collar™ around the proband’s neck and securing the hold with a hook-and-loop fastener, the cushion was slowly inflated using a syringe, while a valve ensured that the air volume inside the cushion remained stable. Maximum volume was 18 ml, which limited the possibility of clamping off vessels or tissue damage due to excessive pressure. Afterwards, the position was documented, and the US device was used to check for audible and visual blood flow information using Doppler data. If verified, the position was used for an iterative measurement, the cushion was deflated, the rail bolt was unsecured, and the rail was moved to the next position. The bolt was fastened, and the pillow inflated again. Measurement and adjustment were repeated for as long as the sensor mount did not touch the lower jar.

Scaled assessments

After execution of semiautomatic measurement workflows, volunteers and study conductors (users) filled out questionnaires. Participants provided an estimation along a scale between 1 (lowest) and 10 (highest) to report how much pain (1 = none, 10 = intense pain) they felt due to the neck brace sensor mount combination and how comfortable (1 = very uncomfortable, 10 = very comfortable) the measurement was overall. The sensor attachment during measurements was assessed by the user and documented on a scale as well (1 = no stable connection, 10 = very robust connection). During our study, volunteer comfort and avoidance of harm were prioritized. Communication and purposefully slow system application increased setup time. Durations that represent only Stifneck Select Collar™ attachment were not recorded.

Binary assessments

During the measurement, the user evaluated whether a blood flow signal was audible and the flow curve on the real‑time display of the device was usable. After measurement completion and removal of the neck brace, the user checked the measurement area for skin lesions as well. Signal audio, flow curve usability and lesion existence were documented with negative/positive checkboxes.

Statistical analysis

Statistical analysis was done with MATLAB R2024a (The MathWorks, Inc., Natick‚ Massachusetts, USA). Descriptive analysis employed Matlab’s functions for mean, standard deviation, median, minimum and maximum values. Tests for normal distributions were done with the Shapiro‑Wilk test addon of Ahmed BenSaïda [25].

System safety

No reports indicated that the Stifneck Select Collar™ or probe attachment system negatively affected circulation and/or respiration.

The construction

The presented device addressed this issue with an approach that is not reliant on dry skin for good adhesion, allows for easy sensor removal and reattachment, and limits head movement. It has little injury risk since all rapid prototyping components feature smoothed edges and corners. Apart from the smoothed sensor probe, only the sensor case had skin contact if it was pushed until the lower jaw. Hook-and-loop fastener strips allowed a fast neck brace detachment in under two seconds, and the sensor rail and pressure pillow enable efficient sensor detachment, repositioning and reattachment, even with the neck brace on. The attachment pressure is manually adjustable and measurable via air volume or an additional barometer. The sensor mount can be attached to any already procured standard Stifneck Select Collar™ neck brace. The attachment can also be done without removing the neck brace from the patient.

The application

Pain or any skin lesion, irritation, or trauma were never reported, which attributes to the safety and efficient combination of the sensor mount and the neck brace. Wear duration of up to 70 minutes provided no issues and didn’t require any measurement termination, although a higher measurement duration originated in the device’s operating system and has been addressed during the study. The combination of the Stifneck Select Collar™ with the sensor holder and the inflated cushion (TR-Band™, Terumo) for contact pressure did not cause any notable restriction in blood circulation or respiration either.

The system had extensive stability over time, which has been verified by medical users. Sensor fixation was resilient to moisture, e.g., sweat, and the relative alignment with the vessel had little risk of shifting, due to the fixed head position. Pressure adjustments or complete deflation are done fast and easily, to keep risks to a minimum. While neckbands are similar in their adjustability, they do not provide comparable stability and skin contact. Proper contact pressure of neckbands can cause breathing problems, occlude vessels and cause skin trauma, while straps and bands can interfere with invasive methods of medical personnel [12].

The measurement

Proband assessments of pain and comfort showed that the system was generally accepted and did not cause noticeable pain or intense discomfort. While some probands noted a slight discomfort associated with the pain grading 2 and 3, the majority felt nothing worth mentioning. The main concern is lower levels of comfort during the measurement, which were represented by the distribution of comfort ratings below level 7. The probands in question noticed the discomfort of the neck brace while resting their head on the bed. Considering the circumstances, the discomfort might be worth the advantages the system provides, since a) most patients of the OHCA use case will be unconscious, b) the risk of actual harm is low, and c) the advantages of CPR evaluation via a stabilized sensor attachment are far greater than potential discomfort after ROSC.

The system is less prone to placement errors due to the fixed mount on the Stifneck Select Collar™. The sensor rail allows only one degree of freedom, and preferred sensor positions can be configured quickly. Adhesive solutions require a complete certainty regarding the CCA position before sensor placement, an unlikely circumstance when facing a new patient. Neckbands are more easily corrected in their position, but the initial placement is just as complex. As long as no clear guidelines exist that define probable vessel depth and best sensor placement in relation to neck size, sensor position adjustments will most likely be necessary.

To our knowledge, no research group has addressed the question of the optimal sensor placement depending on the patient anatomy conclusively. Finding the best spot is usually done by holding the patch and using US audio and visual feedback during blood flow times, requiring multiple people before adhesion [30,31]. The RescueDoppler guiding patch shows the apparent optimal sensor placement in relation to the larynx, yet patient‑specific anatomy could make this placement method invalid [26]. Formulating conclusive statements requires measurements with a sensor attachment method that provides a high degree of repeatability and reproducibility across different patients. Both solutions, adhesive and neckband, struggle with the realization of comparable results due to their respective disadvantages. Comparable sensor placements, quick probe adjustment, and stable measurement conditions are imperative and could be provided with the presented system.

Limitations and future works

The proposed solution has been evaluated during a study for comparable US Doppler measurements with different neck positions on different probands. The largest limitation of the solution is the necessity to manually adjust the sensor probe position along the rail. This is due to the inconclusiveness of the current state‑of‑the‑art regarding the best measurement position. The system, therefore, needs further research and development to draw connections between significant patients’ characteristics and the neck position that is most likely to guarantee an alignment of US SV and the target vessel.

Forceful impacts during CPR applied in the medical use case will add noise to the US measurement and increase the risk of SV misalignment. The impact of those interferences on the measurement has to be researched more thoroughly. Additionally, under current circumstances, the pillow is held to the sensor rail at only one point, creating a rotational axis during inflation, shown in Fig 2. The resulting contact pressure between the sensor probe and skin might therefore be uneven, allowing gaps and corrupting US transmissions further. Displacement over time and hazard of sensor dislocation are minimized, but have yet to be evaluated under conditions comparable to the use case.

The system is also limited in the way that it can not approximate the entire NFT period duration, since NFT periods can occur before system application. The automatic identification of LFT periods is yet to be researched, due to the fact that there is no scientific consensus that defines what amount of flow volume constitutes as sufficient or too low for effective CPR. Further studies have to be conducted to analyze the blood flow volumes of multiple CPR efforts and the respective long‑term medical outcome.

Data collection and analysis also have to include cases with impaired cerebral perfusion. While measured values during CPR might indicate an effective blood flow through the CCA, injuries or sicknesses can still prohibit the actual brain perfusion. Any system feedback about sufficient blood flow might mislead the user about the therapy efficacy and result in severe hypoxic brain damage. At the current moment, the system cannot recognize these problems. Further research and development with a dedicated database is necessary to evaluate if and how the system can gather indications about cerebral flow issues.

The system placement usability in emergency scenarios, i.e., the fastest possible setup without causing harm, has not been addressed in this work either and needs to be analyzed further. At the moment, the device is applied during interviews with volunteers and with a priority on comfort and the prevention of harm, which increases the duration of positioning time and prevents conclusive statements about the fastest valid application time. Without the interview process and with a standard sensor position that could provide high sensor‑vessel alignment probability, the time to start measurements would decrease significantly from the current value of one minute. System application and measurement initiation times could then hypothetically be reduced to the standard duration of Stifneck Select Collar™ fixation, when used by a proficient user. In a real resuscitation scenario, the cervical collar also needs to be temporarily removed during endotracheal intubation, as it may impede airway management and interfere with the intubation procedure. This is quickly achieved with the hook-and-loop system of the collar, keeping intubation delays to a minimum. After airway control has been established, it can be reapplied quickly, while the fixed sensor mount guarantees an almost identical sensor position as before detachment.

Finally, the used Stifneck Select Collar™ has come under critical review in recent years, since it partially covers neck areas that are important for intravenous injections. While the additional use of the Stifneck Select Collar™ as a platform for wearable sensor technology could increase appeal, the higher priority of easy access to jugular veins would require an alternate solution. An alteration to the current neck brace or a completely new solution would be required. The substitute would need to provide similar stability and usability, since most medical personnel are trained to use the Stifneck Select Collar™.