Conductive Thread-Based Stretchable and Flexible Radiofrequency Coils for Magnetic Resonance Imaging

Abstract

Objective:

We propose a novel flexible and entirely stretchable radiofrequency coil for magnetic resonance imaging. This coil design aims at increasing patient comfort during imaging while maintaining or improving image quality.

Methods:

Conductive silver-coated thread was zigzag stitched onto stretchable athletic fabric to create a single-loop receive coil. The stitched coil was mounted in draped and stretched fashions and compared to a coil fabricated on flexible printed circuit board. Match/tune circuits, detuning circuits, and baluns were incorporated into the final setup for bench measurements and imaging on a 3T MR scanner. A fast spin echo sequence was used to obtain images for comparison.

Results:

The fabricated coil presents multi-directional stretchability and flexibility while maintaining conductivity and stitch integrity. SNR calculations show that this stretchable coil design is comparable to a flexible, standard PCB coil with a 13–30% decrease in SNR depending on stretch degree and direction. In vivo human wrist images were obtained using the stitched coil.

Conclusion:

Despite the reduction in SNR for this combination of material, there is a reduced percentage of SNR drop as compared to existing stretch coil designs. These imaging results and calculations support further experimentation into more complex coil geometries.

Significance:

This coil is uniquely stretchable in all directions, allowing for joint imaging at various degrees of flexion, while offering the closest proximity of placement to the skin. The materials provide a similar level of comfort to athletic wear and could be incorporated into coils for a variety of anatomies.

I. Introduction

Magnetic resonance imaging (MRI) is a noninvasive modality with the ability to reveal contrast between a broad range of tissues and joint components including muscle, cartilage, bone, ligaments, and tendons [1]. Radiofrequency (RF) receive coils are used to detect the localized NMR signal that is subsequently processed into the MR image. Typically, to enhance the signal-to-noise ratio (SNR), receive coil arrays are shaped to encompass a generalized form of the anatomy of interest. Fixed coil arrays are durable and designed for specific applications, e.g., brain imaging; however, multiple sizes are often desirable to facilitate a closer fit and maximize SNR [2]. Researchers have identified flexible coils as a means to improve SNR, with recently-proposed flexible or bendable planar RF coil designs fabricating coil conductors using screen-printing technology [3], copper braid [4], [5], and custom coaxial cable [6], [7]. These designs are well-suited for wrapping coil elements around an elliptic cylinder, e.g., the torso or abdomen. However, unidirectionally flexible coils do not facilitate MRI of joints such as the shoulder, knee, and elbow.

Millions of patients suffering from joint injuries and degenerative joint diseases, such as arthritis, have difficulty fully extending their joints [1], [8]. With increased life expectancies, it is estimated that more than 78.4 million individuals in the United States will be diagnosed with arthritis by the year 2040 [8]. Commercial joint-imaging RF coils are rigid and require the subject to lie in a pronated or supine position with the joint fully extended. Notably, a stretchable knee and wrist coil has been developed utilizing copper braid solder wick on a bandage designed to be stretched as it is wrapped around an area [4], [5]. Copper braiding is bendable but does not stretch, making a universal fit difficult. Other stretchable designs include liquid metal on neoprene [9], [10]. While neoprene can stretch, it is a heavy, less breathable fabric. There is a need for a flexible and stretchable coil that enables comfortable positioning, particularly during lengthy MRI scans ranging from 20–60 min [11]. In addition to improving patient comfort, RF coil placement nearer the skin could enhance image quality since the coil is closer to the signal source.

In this work, we introduce a flexible and omnidirectionally stretchable RF coil. By utilizing conductive thread on athletic fabric, as opposed to copper or inks, the RF coils are able to stretch and bend like an athletic sleeve or compression band. Preliminary versions of this work have been reported in conference proceedings [12], [13]. We present results of both bench measurements and MRI experiments. Single-loop stretchable coils are evaluated at draped and stretched coil positions and compared with a flexible copper-trace surface coil. Lastly, we discuss the feasibility of employing this fabrication technique for arrays.

II. Materials and Methods

A. Coil Design

Two 71-mm (~3-in) single-loop receive coils were stitched on 90% polyester, 10% spandex athletic fabric with Lyofil thread (silver-coated, p-phenylene benzobisoxazole (PBO), Syscom Advanced Materials Inc., Columbus, OH) and standard polyester bobbin thread (Fig. 1). The dimensions were chosen to fit over the patella of the knee or other joint. Silver coating was chosen for the thread due to its higher conductivity over copper. In order to facilitate stretching, the preset zigzag stitch on a Brother JX2517 sewing machine (Brother International Corporation, Bridgewater, NJ) was used. The stretched-to-draped coil radius ratio is 42.5:35.5 mm. A comparison coil was etched on flexible, 1-oz copper-clad, 0.18-mm thick, FR-4 printed circuit board (PCB) in a 76-mm (3-in) circle design with a trace width of 2.8 mm (Fig. 1). This coil was mounted to a custom 3D-printed half-cylindrical shell, with an inner diameter of 73 mm, outer diameter of 79 mm, and thickness of 6.0 mm designed to fit over a 1-L, 73-mm diameter, phantom bottle with an adhesive hook and loop fastener. This setup can be seen in Fig. 1. Based on the desired degree of stretch, hook-and-loop fasteners were used to mount the stretchable coil in either a draped, horizontally-stretched, vertically-stretched, or omnidirectionally-stretched position. In all stretched positionings, the coil was distorted to the maximum size allowed by the length of thread utilized.

Fig. 1.

Side-by-side comparison of PCB coil (left) and a draped, stitched stretchable coil (right).

All coils were segmented on opposite sides allowing for the attachment of a match/tune board and variable capacitor (BFC280832659, Vishay Intertechnology, Malvern, PA). A 27-pF capacitor was added adjacent to the match/tune board. This board also functions as a current trap, utilizing a 5-mm diameter tunable, unshielded inductor (165–00A06L, Coilcraft Inc., Cary, IL) and PIN diode (MA4P7470F-1072T, MACOM, Lowell, MA). The circuit diagram and board are shown in Fig. 2. One quarter-wavelength of coaxial cable, including a balun, connected the coil upstream via a BNC connector.

Fig. 2.

(a) Schematic for the coil, match/tune board, and upstream connection to the balun. The dashed region identifies the trap comprised of L, C₁, and D. C₃ is utilized for tuning and C₂ for matching. (b) Detailed view of the match/tune board with the incorporated current trap and capacitor, C₂, adjacent to the board.

B. Coil Testing- Bench Measurements

A network analyzer (E5071C, Keysight Technologies, Santa Rosa, CA) was utilized to tune the trap and coil and measure input impedance. Prior to tuning the trap, an S₁₁ measurement was employed to match and tune the coil to the Larmor frequency of hydrogen at 3T (128 MHz) by adjusting the tuning capacitor, C₃. This tuning capacitor was then removed to create an open circuit, and an S₂₁ measurement was taken using a small, single-loop probe lightly coupled to the inductor and a second sniffer probe placed above the diode to actively measure the trap tuning during the off and on switching of the PIN diode (Fig. 3). A dc power supply was connected to a series 10-Ω resistor to drive the PIN diode circuit. 5 V was supplied to forward bias the diode and the trap was tuned to 128 MHz. Subsequently, the tuning capacitor was reattached and adjusted so that, with a bias of −5 V, the coil was matched and tuned. These measurements were repeated for all coil types and positions and are henceforth referred to as MT (matched/tuned) configurations.

Fig. 3.

Single-loop and sniffer probe placements for S₂₁ trap tuning.

To address the variability in frequency shift due to coil stretching, measurements were taken by first matching and tuning the stretchable coil in the draped position. Then, the coil was stretched horizontally and remounted using the hook and loop fastener. No additional matching and tuning were performed. This process was repeated for the vertical and omnidirectional positionings. These coils with no additional matching and tuning are referred to as NMT (not matched/tuned).

After the matching and tuning of the coils were completed for the MT coils and subsequent stretching for the NMT coils, the coil quality factor (Q-factor) was calculated using the following formula for S₁₁ measurements, where f₀ is the resonant frequency and Δf_-7dB is the bandwidth measured at the −7 dB crossings to the left and right of f₀ [14], [15]:

$$Q=\frac{f_0}{\mathrm{\Delta }{f}_{-7\text{dB}}}$$ (1)

Q-factors were calculated in both unloaded and phantom-loaded states. The phantom was a 1-L bottle comprised of 16.7 g of NaCl and 538 g of sugar dissolved in 700 mL of water, as determined using the Dielectric Phantom Recipe Generator based on [16] to simulate the dielectric properties of human muscle, which has a relative permittivity of 63.5 and a conductivity of 0.719 S/m as determined through the IT’IS Foundation Tissue Properties Database [17]. 0.7 g of CuSO₄ was added to shorten the T₁ and facilitate a faster repetition time (T_R) of the MRI pulse sequence. Muscle tissue was chosen due to the intent to use these coils for joint imaging. To consider the influence of variable proximity to the load for a rigid coil, the PCB coil was positioned in two locations: wrapped tightly around the phantom and mounted 4.2 cm (approximate distance between one subject’s joint and a rigid commercial volume coil) above the phantom. The stitched coil was evaluated while draped and while stretched omnidirectionally, horizontally, and vertically.

C. Coil Testing- MRI Experiments

Coil testing was performed on a whole-body 3T MR scanner (Discovery MR750, GE Healthcare, Chicago, IL). Each coil was connected individually to a receiver gateway box (16xRx, Clinical MR Solutions, Brookfield, WI) and positioned as previously described around the aforementioned muscle phantom (Fig. 4). The insulation distance between the coil and phantom was that of the shell thickness for all setups aside from the 4.2 cm spaced set up (giving an insulation distance of 4.8 cm). A fast spin-echo (FSE) sequence was executed with the following preset parameters: sequence repetition time (T_R) 4.01 ms, echo time (T_E) 120 ms, slice thickness 10 mm, number of signals averaged (NSA) of 1, pixel size 0.7 mm × 0.7 mm, grid 256 × 256. The built-in body coil was used to transmit, while the stitched coils and PCB coil were receive-only. All phantom images were acquired during the same scanning session and included a localizer, calibration, and FSE.

Fig. 4.

Scanning setup with the stretchable coil loaded/draped on phantom and connection to receive gateway.

After scanning, phantom data were imported into MATLAB (MathWorks, Natick, MA). Using in-house code, SNR was calculated following the alternate single-image SNR measurement procedure prescribed by section 2.5 of the NEMA MS 6 standard for single-channel non-volume coils [18]. A 7×7-pixel region of signal and 11×11-pixel region of noise were selected (Fig. 5). To evaluate SNR at various depths inside of the phantom, three regions of signal were selected, with regions denoted A, B, and C spaced 5 mm apart. The in-house code was validated by equivalent SNR calculations using OsiriX Lite (Pixmeo SARL, Geneva, Switzerland).

Fig. 5.

Positions of signal: A (purple), B (red), C (blue) and noise region: D (green) for calculating mean signal and noise standard deviation. The black lines denote the boundary of the phantom, and the coil is depicted in tan.

D. Coil Testing- Temperature Evaluation

Prior to in vivo use, evaluation of heating from components near the skin was performed using fiber-optic probes (AccuSens, Opsens Solutions, Montréal, QC, Canada). Placement of the fiber-optic probes is illustrated in Fig. 6. Absolute temperature was recorded while running a 52-min, 23-s preset GE protocol for knee and joint imaging.

Fig. 6.

Placement locations of four fiber-optic probes to assess heating of coil elements and verify safety of coil before in vivo use.

Human in vivo wrist MRI were acquired using the stitched, stretchable coil on the same scanner. The coil was centered over the top of the wrist, palm down, with the subject in the prone position. There was no added insulation between the subject and the coil. The sagittal FSE image was obtained using the same scan parameters listed above and the axial image was obtained using a T₁-weighted sequence with the following parameters: T_R 539 ms, T_E 9.19 ms, slice thickness 2.5 mm, NSA of 3, pixel size 0.2 mm × 0.2 mm, grid 512 × 512. Consistent with the phantom scans, the built-in body coil was used to transmit, while the stitched coil was receive-only. The experimental procedures involving human subjects described in this paper were approved by the local Institutional Review Board (protocol 19030219).

E. Coil Testing- Decoupling

After coil validation through phantom and human scanning, the geometric overlap decoupling performance of these stretchable coils was analyzed. Plastic embroidery hoops with a diameter of 12.7 cm were used to stabilize the coils in a draped/unstrained or stretched position (Fig. 7). The metal screw used to tighten the hoop was replaced with a plastic cable tie.

Fig. 7.

Stitched coils mounted in an omnidirectionally-stretched configuration using plastic embroidery hoops, overlapped for decoupling measurements. Note that the coil in the green hoop is stretched to fit over the inner hoop while the coil in the orange hoop is stretched to fit under the inner hoop.

The matching and tuning of the coils were verified in isolation using S₁₁ measurements prior to overlapping. To measure decoupling performance, S₂₁ measurements were acquired while gradually moving the coils to the optimal overlap. This process was executed for various stretching configurations, loaded and unloaded on a muscle phantom with the same dielectric properties as the previous (Fig. 7).

III. Results

A. Coil Testing- Bench Measurements

Impedance measurements for all stretched/draped configurations in MT and NMT comparisons with the flexible PCB coil are summarized in Table I. From this table, it can be seen that stretching the stitched coil impacts the impedance, with the most significant change being with the omnidirectionally stretched coil. The stretched-coil radius is approximately 26% larger than the radius of the draped coil.

TABLE I.

Loaded Impedance Measurements Z = R + jX

Coil	MT (Ω)	NMT (Ω)
PCB (Spaced 4.2 cm)	32.6 + j15.8	-
PCB	50.0 + j0.104	-
Stitched (Draped)	50.4 + j0.172	-
Stitched (Stretched- Omnidirectional)	50.3 − j0.200	487 + j 10.7
Stitched (Stretched- Horizontal)	50.0 + j0.175	149 + j93.4
Stitched (Stretched- Vertical)	50.0 + j0.00936	122 + j53.8

Q-factor calculations from both the loaded and unloaded measurements are summarized in Table II. The loaded Q-factors of the stretched coil configurations achieve body noise dominance, with loaded Q values approximately half of the corresponding unloaded values.

TABLE II.

Q-Factor Calculations

Coil	Q, loaded	Q, unloaded
PCB (Spaced 4.2 cm)	116	153
PCB	45.8	153
Stitched (Draped)	35.4	53.0
Stitched (Stretched- Omnidirectional)	37.5	65.3
Stitched (Stretched- Horizontal)	20.8	38.6
Stitched (Stretched- Vertical)	27.3	55.7

B. Coil Testing- MRI Experiments

Fig. 8 shows the resulting MR images gathered from the single-slice FSE sequence. All images have been set to have the same contrast and brightness for comparison. The resulting SNR calculations at varying phantom depths for the MT coils and NMT coil at region A, depicted in Fig. 5, are presented in Table III. Because the PCB coils only flex, no additional matching and tuning was needed, so NMT data were not collected. It appears that the PCB coil has a greater signal value than the stitched coils. However, the calculated SNR values are similar between the draped and PCB coils and, based on the comparison of the PCB coil oriented on versus spaced 4.2 cm above the phantom, there is a clear advantage of proximity of placement. With a larger coil, there is an increase in the total noise volume, yielding a lower SNR. All images in Fig. 8 exhibit the same artifacts at the top and bottom extremes nearest the coil, where the coils’ fields include a significant z-directed component insensitive to signal.

Fig. 8.

FSE phantom images using MT (A) PCB coil, (B) PCB coil spaced 4.2 cm above phantom, (C) Stitched-draped coil, (D) Stitched omnidirectionally-stretched coil, (E) Stitched horizontally-stretched coil, and (F) Stitched vertically-stretched coil.

TABLE III.

SNR Calculations

Coil	MTA	MTB	MTC	NMTA
PCB (Spaced 4.2 cm)	717	503	349	-
PCB	1070	964	712	-
Stitched (Draped)	923	771	545	-
Stitched (Stretched- Omnidirectional)	744	694	538	426
Stitched (Stretched- Horizontal)	935	840	645	595
Stitched (Stretched- Vertical)	769	735	579	677

To verify the coil size impacted loaded SNR values in stitched stretched and draped positions, comparison was performed between the flux densities, B₁, using the Biot-Savart law. The flux density ratio should be equivalent to that of the loaded SNR measurements. This was confirmed, as the ratio of draped to omnidirectionally stretched for both B₁ and SNR_loaded within region A is 1.24. With the increased coil size, whether horizontally, vertically, or omnidirectionally stretched, there was a lesser reduction of SNR calculations using signal regions deeper in the phantom. From selection regions A to C, seen in Fig. 5, and the resulting values presented in Table III, the PCB coil had an SNR reduction of 51%, whereas the vertically stretched had a decrease of only 28%, horizontally 25%, and vertically 31%.

C. Coil Testing- Temperature Evaluation

To further assess whether retuning was required when the coils were stretched, SNR calculations for the NMT coils are also included in Table III. As expected, there was a reduction in SNR when compared to the MT configurations. This indicates that maintaining tuning at the operating frequency is crucial for improving image quality.

Fig. 9 shows the change in temperature over a 41-min, 32-s, scan consisting of 12 series from the preset GE protocol for imaging a knee. With the pauses between each sequence, the total scan time was 52 min, 23 s. The durations of most pauses between scan sequences were 26–59 s. The longest pause between the 12 series was 3 min, 34 s at the beginning of the calibration series and did not produce a discernable cooldown of the components. The other two longer pause were 1 min, 39 s and 2 min, 21 s. Of the 11 periods between the series, 7 are perceptible, primarily at the fixed capacitor junction and match/tune board. The greatest temperature rise occurred during the 7-min, 37-s axial, proton density, adiabatic spectral inversion recovery sequence (Ax PD ASPIR) near 25 min, followed by a 2-min, 21-s gap before the 3-min, 35-s Ax PD FSE fat saturation (FS) sequence which generated the highest temperature. The maximum 6-min average SAR was 2.6 W/kg (out of a 3.2 W/kg limit for a 68.04 kg input weight). The maximum 10-s average SAR was 3.1 W/kg (with a corresponding limit of 6.4 W/kg). Both max SAR values occurred during the Ax PD ASPIR sequence. All probe/channel data are summarized in Table IV.

Fig. 9.

Radiofrequency coil and component temperature change over time. The heating in the area of probe placement, as identified by the legend and illustrated in Fig. 6, is displayed during the course of a 52-min, 23-s (3143-s) scan.

TABLE IV.

Temperature Probe Data

Temperature (°C)	Match/Tune Board	Tuning Capacitor	Thread	Fixed Capacitor Junction
Average	19.8	18.9	18.1	19.1
Standard Deviation	0.977	0.203	0.221	1.04
Minimum	17.8	18.5	17.5	17.1
Maximum	21.8	19.3	18.5	21.5

While synthetic fabrics, such as spandex, are flammable, the overall heating did not lead to any indication that this material would pose a burn risk to the patient, as the maximum temperature of 21.8°C was well below the sticking point of 175°C and melting point of 250°C for spandex [19]. The melting point of polyester is 260°C [20]. Due to the zigzag stitching method, the conductive thread is anchored only to the top of the fabric, while the only thread that would potentially come in contact with a subject is nonconductive bobbin thread.

D. Coil Testing- Human Scanning

Select in vivo human wrist images are shown in Fig. 10. Clear distinction of the wrist anatomy and fine details, such as the porosity of the bone, can be observed. The subject was able to comfortably place their wrist at a 30° bend for the scan. Due to the coil overlapping past the top surface of the wrist (the subject had a wrist circumference of 13.8 cm, allowing the coil to wrap beyond half of the volume), some artifacts can be seen, more so in Fig. 10a. These scans demonstrate the imaging quality and potential for these coils, notably while scanning joints at a bend.

Fig. 10.

Human wrist images using the stitched coil (wrist positioned at 30° bend): (A) Axial T₁-weighted image, (B) Sagittal FSE image.

E. Coil Testing- Decoupling

The degree of coil decoupling resulting from optimal overlap changed with the stretching of the coil. The inter- element coupling S₂₁ measurement for select stretching scenarios is summarized in Table V. These stretch configurations were chosen based on expected distortion patterns when imaging a variety of joints at various degrees of flexion.

TABLE V.

S₂₁ Coil Decoupling Measurements (dB)

Stretch Configuration: Coil 1/Coil 2	Loaded	Unloaded
Draped/Draped	−10.7	−21.9
Omnidirectionally Stretched/Draped	−18.9	−30.8
Horizontally Stretched/Draped	−17.7	−28.6
Horizontally Stretched/Omnidirectionally Stretched	−20.9	−30.1
Omnidirectionally Stretched/Omni. Stretched	−17.6	−31.9

Given the loaded and unloaded S₂₁ measurements, the decoupling performance remained satisfactory in the event of coil shape distortion with those coils showing greater degrees of distortion exhibiting greater decoupling in both the loaded and unloaded states. As expected, the phantom loading affected the mutual inductance, resulting in greater coupling.

IV. Discussion

Given the tight-fitting nature of a stretchable coil, separation of conductors and heat sources must be assured. The thread and fabric are washable, providing a hygienic aspect to the utilization of these coils, assuming any lumped elements and accessory boards are designed to be detachable. The stitch pattern and thread have maintained their integrity across several stretched configurations. For illustrative purposes, Fig. 11 shows a 76.2 mm × 76.2 mm, square-stitched coil stretched to 101 mm × 101 mm around a 170-mm spherical phantom. There is limitation to the amount of stretching possible with this design. Since the zigzags can only be lengthened to be fully taut. A wider zigzag would allow for better utilization of the stretchable fabric.

Fig. 11.

Stitched square coil stretched over a spherical phantom.

In these experiments, all coils were mounted on a half-cylindrical shell that fits over the phantom. The main purpose of this mounting was to maintain a consistent stretch and coil size for the stitched coils. This maintenance of size was crucial for matching and tuning, as well as maintaining such during the scans. As previously mentioned, the coil was matched and tuned on the subject prior to scanning to minimize the shift in operating frequency when placed on the subject for scanning Without this additional matching and tuning, a frequency shift of around 7 MHz was seen with a reduction in attenuation. In order to address variability in operating frequency due to coil stretching, strategies may be employed to dynamically match the coil, e.g., piezoelectric actuators [21] or PIN diode circuit switches [22]. The stitched coils, both draped and stretched, exhibited similar Q-factor values to the comparison PCB coil when loaded. There is a substantially higher unloaded Q-factor for the PCB coil when unloaded, which was expected given the lower resistivity of the copper trace compared to the conductive thread. Despite the lower SNR values for the stitched coils, there is the advantage of greater retained SNR at deeper phantom depths due to the increase in coil size. Reduction of coil resistance, through alteration of thread composition, could improve the SNR further.

When comparing the horizontally and vertically-stretched coil configurations to the omnidirectionally-stretched setup, a greater SNR reduction was seen in the horizontally-stretched coil. This is most likely due to the position of the signal region for SNR calculations. An alternative SNR calculation method could be explored for future comparison.

The Q-factor and SNR calculations for the PCB coiled spaced 4.2 cm above the phantom are important in illustrating that the coil proximity of placement to the skin has an impact on image quality. Even with a higher Q-factor than the stitched, stretchable coils, the SNR (Table III) was lower for the spaced PCB coil when compared to all stretchable coil configurations (horizontal, vertical, and omnidirectional). This supports the notion that coils designed to custom fit the anatomy of interest would yield better images over rigid counterparts because they allow closer placement to the skin.

Current stretchable coil designs have exhibited similar decreases in SNR, mostly likely due to the higher resistance of the coil. Mehmann et al. demonstrated a 33% reduction in SNR from their copper reference coil to their unstrained liquid metal coil [9]. Comparatively, the presented stitched, draped coil (unstrained) only showed an SNR reduction of 14%. This is due to the lower resistance of the thread.

An RF coil that is both stretchable and flexible may increase the utility of MRI for a broader range of applications beyond joint disease and injury. Given its lightweight and body conforming construction, this coil design is advantageous in imaging curved anatomies such as breasts and areas in close proximity to the bore, such as shoulders or hips. The stretchable coils presented herein may be utilized to fabricate an array of receive elements. This is supported by their ability to maintain decoupling under various coil distortion scenarios (Table V). The more plausible combinations for in vivo applications, such as horizontally stretched and omnidirectionally stretched (where the joint would be bent to greater degrees), had a lesser measure of coupling than the draped coils of uniform size and shape. orthogonal geometric decoupling between elements may be exploited for low-channel count arrays, while numerous techniques may be employed for high channel count arrays, e.g., element overlap [23], induced current elimination [24], resonant inductive decoupling [25], or self-decoupling through mutual impedance cancellation [26].

V. Conclusion

This study indicates that conductive thread and athletic material is a viable combination for an entirely stretchable coil that allows for increased patient comfort through a variety of available positionings. The stitched coil had a reduction in both Q-factor and SNR yet showed less SNR reduction at greater phantom depths as compared to the PCB coil. Furthermore, a lesser degree of SNR reduction was seen when compared to current printed liquid metal stretchable designs. These stretchable coils pose an advantage in joint imaging of shoulders and hips, where traditional coil setups require close, cramped placement of the coil on the patient within the MRI bore. This omnidirectional stretchable coil not only provides a significant advantage in the variety of positioning available across anatomy, but also for patient comfort given the incorporation of fabrics commonly used in athletic wear. While the zigzag stitch is a conventional stretch stitch, design and testing of new stretch stitches is an area of future interest. Coil decoupling was achieved and maintained through various coil distortion combinations supporting the fabrication of a receive array coil and comparison against current industry RF coils as areas of future research.