Ultrasound-guided fascia plane blocks have gained popularity in recent years due to the advancement in ultrasound technology, allowing for muscle and fascia identification and needle enhancement.¹ Fascia blocks target intermuscular fascia layers in which peripheral nerves lay; the correct identification of musculoskeletal anatomy and needle tip placement within a specific layer is vital to the success of a block.²

One type of fascia block is the transversus abdominis plane (TAP) block.³ TAP and its variants provide analgesia for abdominal surgeries^4,5 and have played an important role in opioid-reduced surgery and enhanced recovery after surgery (ERAS) protocols.⁶ The introduction of ultrasound techniques has allowed for the performance of peripheral blocks with greater accuracy and safety.^7,8 While this is a relatively safe block to perform compared to a central neuraxial block, ultrasound-guided peripheral nerve blocks have their limitations. Distribution of sensory fibers in the lower abdominal area is quite variable and may require using a different regional anesthesia technique (eg, the quadratus lumborum block).³ Technical limitations,^9,10 including needle visualization, artifacts, and probe positioning, may result in needle tip misplacement. The success rate of TAP catheter placement rate is estimated to be 63%–88%.¹¹ Some therapeutic failure rate is estimated to be as high as 30%.¹² Training to improve hand-eye coordination is essential for the success of ultrasound-guided blocks.¹³ For inexperienced physicians, failure to visualize the needle may lead to failed blocks¹⁴ or severe complications (eg, nerve damage,¹⁵ inadvertent peritoneal puncture, and intravascular injection). Furthermore, local anesthetics may be adversely injected into surrounding muscles,¹⁶ thus decreasing efficacy of the block.

Optical coherence tomography (OCT) is an imaging modality based on low-coherence interferometry. OCT can provide axial image resolution of approximately 10–15 μm—10 times better than the resolution of typically acquired clinical ultrasound images—and provides microscopic anatomical confirmation from the needle tip. In our previous studies, we were the first to propose side looking fiber-needle swept-source OCT (SSOCT) imaging for the identification of epidural space with a high success rate¹⁷ and high sensitivity (97.5%), specificity (95%), and accuracy (96.2%).¹⁸

The transversus abdominis fascia plane has not been previously identified in real-time by OCT imaging. In training to perform a TAP block, needle “tip” visualization, in addition to hand-eye coordination and needle visualization, is the most challenging. Attention to this fine detail is essential for the success of the block.¹⁹ In light of that, OCT may be used in fascia blocks to increase tissue recognition, which could potentially increase accuracy. In this study, we aimed to assess the feasibility of the needle-probe SSOCT system to identify the TAP and the surrounding tissues.

SSOCT System and the Needle Probe

A schematic diagram and an experimental photograph of the SSOCT system are shown in Figure 1. A swept laser (Axsun, Billerica, MA) at 1310 nm with a bandwidth of approximately 100 nm, a sweep frequency of 100 kHz, and an output power of 30 mW were used as the light source. The input light was split by a 1 × 2 coupler (splitting ratio 50:50) into 2 arms. One arm passed through a circulator and connected with a fiber probe for side-looking imaging. Figure 1C shows the OCT fiber probe within a puncture needle. The design of our fiber probe has been previously described in detail.¹⁷ In brief, a fiber probe was connected to a rotary motor, covered by a plastic catheter (with a 0.9-mm outer diameter), and placed into a 17-gauge Hustead needle with a 1.07-mm inner diameter and a 1.47-mm outer diameter (Arrow FlexTip Plus Epidural Needle, Teleflex, Wayne, PA). The rotary motor drove the fiber probe with a rotational rate of 80 rounds per second. Another arm passed through a circulator and was converted into a collimated beam and reflected by a mirror. In the detection arm, the reflected light, which came from the mirror and the fiber probe, were sent into a balanced detector (PDB470C, Thorlabs, Newton, NJ), and the signals were collected by a 12-bit analog to digital converter (ADC, ATS9360, Alazartech, Montreal, Canada). By sweeping the laser wavelength and detecting the intensity using a single detector, this method shows improved sensitivity and thus enables the rapid 3-dimensional imaging of tissues. The SSOCT algorithm was processed by field-programmable gate array (FPGA) real-time data acquisition and displayed at a rate of 80 frames per second with 1280 A-lines in each frame. Figure 1D shows a representative OCT image acquired by circumferentially scanning the fiber probe with a rotational motor.

Ex Vivo OCT Identification of TAP Protocol

Swine shares many anatomic and physiologic characteristics with humans; therefore, this study is based on a swine model.²⁰ For ex vivo experiments, a slice of pork belly was used (Figure 1B, left). First, the OCT probe was placed within a 17-gauge Hustead needle, then ultrasound-guided needle insertion using the in-plane technique was performed with a linear probe (M-turbo, Sonosite, Bothell, WA). OCT images were obtained in real-time as the ultrasound-guided needle trajectory to TAP plane. OCT tomograms from the muscle and fascia layers were acquired to evaluate whether the characteristic features of different tissues are visible in our needle probe-based SSOCT images. As the OCT needle probe progressed through the tissues, the location of the needle tip can be identified using near-infrared (NIR) imaging (ElectroViewer 7215, Electrophysics, Fairfield, NJ) (Figure 2A, B). The needle was inserted within 2 different angle ranges (0°–30° and 30°–60°). Fifty images were acquired from each puncture. The procedure flow chart is shown in Supplemental Digital Content 1, Figure 1, https://links.lww.com/AA/D260. The ultrasound-guided needle insertion images and corresponding OCT-needle probe images are shown in Figure 2C–F.

In Vivo OCT Imaging and Real-Time Identification of TAP

For in vivo experiments, the procedures were performed on 3 female 6-month-old native Chinese Landrance Duroc pigs with an average weight of 25 kg. The study was approved by the Institutional Animal Care and Use Committee of Taipei Veterans General Hospital (Taipei, Taiwan). Animals were intubated after the induction of general anesthesia with Zoletile 50 (Virbac, Carros, France) and mechanically ventilated with isoflurane. The animals were placed in the right lateral position for needle placement. Ultrasound imaging of the OCT needle probe was also performed at the same time for needle trajectory guidance. The linear ultrasound probe was placed transversely between the lower costal margin and the anterior superior iliac spine of the pigs. A total of 10 punctures were performed, including needle insertion within 2 different angle ranges (0°–30° and 30°–60°). The procedure flow chart is shown in Supplemental Digital Content 1, Figure 1, https://links.lww.com/AA/D260. Our SSOCT probe presents and obtains “live” 2-dimensional (2D) images during needle movement. The insertion procedure will stop until the fascial layer was identified by directly observing fascial structures in the OCT image. After obtaining the images, the OCT probe was removed, and hydrodissection with 5 mL of normal saline was injected via the needle for final fascia plane confirmation (Figures 2F and 3C). Twenty independent raters were enrolled for the fascia tissue identification to test for the OCT image criteria. All are experienced anesthesiologists who were uninformed of the needle placement. They were asked to categorize the in vivo OCT images into muscle or fascia on a laptop computer after a short briefing about the criteria.

Extraction of Quantitative Optical Properties From SSOCT Images and Statistical Analysis

Attenuation coefficients have been used for various diagnostic and classification purposes for biomedical applications, and they can be measured by optical methods (eg, OCT in this study). To objectively discriminate whether the needle tip was positioned inside the fascia tissue and to quantitatively assess the influence of different insertion angles within the 0°–30° and 30°–60° ranges on OCT images, the attenuation coefficients were extracted from the lateral averaged depth-profile signal in each of the 2D OCT images.

Furthermore, for tissues in the near-infrared spectral range (eg, 1300 nm in this study), the absorption coefficient is too small and can be ignored. Thus, based on a single-scattering model, the power of a light beam propagating through a turbid medium is attenuated along its path mainly due to the scattering effect and is described by the Lambert-Beer law.²¹

I(r) is the irradiance of the beam after traveling through the medium over a distance (r). I₀ is the irradiance of the incident light beam, and μ_t is the attenuation coefficient. The attenuation coefficient (µ_t) was determined from OCT data by linearly fitting the logarithm of an OCT depth profile. A more rapid decline of the irradiance of light with depth indicates a more significant attenuation coefficient and thus implies the tissue with a higher scattering property.

Since this fitting approach requires tissue with a relatively uniform attenuation coefficient over a certain depth, much data needs to be averaged before a reliable fit can be obtained. In this study, the OCT depth profile was obtained by lateral averaging along the θ direction in each of the 2D OCT images within the tissue’s subsurface structure using the following equation:

Here, r denotes the depth, θ represents the probe’s rotation angle, and N is the number of pixels along the θ direction at a specific depth (r) of an image. Finally, I is the intensity of each pixel.

Statistical data for µ_t are presented as the mean and standard deviation (SD). Comparisons between the 2 groups were assessed using the linear mixed-model analysis. P values <.01 were considered statistically significant. The receiver operating characteristics (ROC) curves were used to show the capacity for using this μ_t value to discriminate between the 2 sample populations. The area under the ROC curve (AUC) was used to provide a measurement of the diagnostic discrimination of a test. All statistical analyses were performed using SPSS software (SPSS 24, IBM, Armonk, NY).

RESULTS

A total of 200 ex vivo 2D SSOCT scans were obtained; half were obtained using a needle insertion angle within 0°–30°, and the other half were obtained with an angle within 30°–60°. Each group included 50 images from the sites identified as fascia, and the other 50 were from the muscle tissue. The representative results of the muscle and fascia are shown with NIR imaging (Figure 2A, B) and OCT imaging (Figure 2C, D). During the acquisition of the OCT images, 3 layers of abdominal muscle (ie, the external oblique, internal oblique, and transversus abdominis) were identified by ultrasound image (Figure 2E). When the needle tip was estimated to be within the fascia layer, the conventional hydrodissection technique was also used to confirm the tip location (Figure 2F).

After the ex vivo experiment, 10 insertions with different angles were performed (each insertion obtain 50 in vivo 2D images of the fascia and muscle tissue), and a total of 1000 images were obtained from 3 piglets. The representative ultrasound-guided needle insertion images and corresponding OCT images are shown in Figure 3 and Supplemental Digital Content 2, Video 1, https://links.lww.com/AA/D261. The laterally averaged intensity of each image was calculated simultaneously. The muscle tissue (Figure 3B) had a layered structure and dispersed adipose tissues, which were identified as nonuniform signal distribution in the OCT images. OCT images of the fascia layers (Figure 3D) have a shallower depth due to the fascia being a strong-scattering and homogeneous tissue, which reduces light penetration.

To know whether a different puncture angle will influence the OCT image or not, Supplemental Digital Content 3, Figure 2, https://links.lww.com/AA/D262, presents OCT images with different puncture angle ranges. Supplemental Digital Content 3, Figure 2A, B, https://links.lww.com/AA/D262, shows the puncture angle of a needle probe within 0°–30° and 30°–60°, respectively. Different angles of puncture produce different image depths in muscle tissue (Supplemental Digital Content 3, Figure 2C, E, https://links.lww.com/AA/D262). However, for the fascia layer, the image depth remains the same despite the different angle of needle entry (Supplemental Digital Content 3, Figure 2D, F, https://links.lww.com/AA/D262).

To provide an objective judgment tool which can help a physician to discriminate whether the needle tip was positioned inside the fascia tissue or not and to quantitatively assess the influence of different insertion angles on OCT images, Figure 4A, D shows the representative OCT depth-profile after the lateral averaging of intensities within 1 image from the in vivo results. The arrows indicate the first interface (ie, the surface) of the tissue. Figure 4A demonstrates the needle probe insertion within 0°–30°, whereas the 30°–60° results are shown in Figure 4D. The attenuation coefficients of the in vivo fascia and muscle tissue were extracted by linear fitting the logarithm of an OCT depth-profile intensities within 900 µm, shown in the dotted line in Figure 4A, D. The attenuation coefficients from 500 fascia images and 500 muscle images with different needle entry angles (all were within 0°–30° and 30°–60°) are calculated in Figure 4B, E, respectively. With a 0°–30° needle angle, the means and SDs of the attenuation coefficient values calculated from the in vivo fascia and muscle images are 0.084 (0.01) (mean [SD]) and 0.059 (0.01), respectively. With a 30°–60° needle angle, the means and SDs of the attenuation coefficient values calculated from the fascia and muscle tissue were 0.088 (0.01) (mean [SD]) and 0.037 (0.01), respectively. In both 0°–30° (Figure 4B) and 30°–60° (Figure 4E) needle angle, there was a significant difference in attenuation coefficients between the fascia and muscle (P < .001). No significant difference in the average μ_t was noted among the distinct repeated measures in either the fascia or muscle tissues. The ROC curves (Figure 4C, F) demonstrate the prospective use of the attenuation coefficients for discriminating fascia tissue. The AUC for the ROC curve indicates that the attenuation coefficients have a high discriminatory capacity for discriminating between the fascia and muscle tissues either the needle entry angle was within 0°–30° (AUC = 0.93) or 30°–60° (AUC = 1.00) (Figure 4C, F).

The above visual details regarding the morphological information obtained from the ex vivo and in vivo OCT images are summarized in Table 1 and were used to establish the OCT image criteria for distinguishing the fascia from the outside tissues. The independent validation of this OCT image criterion for the muscle/fascia differentiation by 20 OCT readers for the in vivo data demonstrated the sensitivity = 0.91 (range 0.73–1) and specificity = 0.90 (range 0.75–1), as shown in Table 2. Diagnostic accuracy was 0.89, with a range of 0.69–1.

DISCUSSION

TAP block is a fascia block and was first introduced in 2001.²² The success of the ultrasound-guided technique depends in part on the correct identification of muscle layers and placement of local anesthetic in the fascia plane between the internal oblique and transversus abdominis muscles.^23,24 Needle visibility during ultrasound-guided procedures is often limited by the dispersion of the needle’s reflections away from the probe.²⁵ Numerous needle enhancement algorithms have been developed and incorporated in ultrasound machines over the years.²⁶ However, needle tip visualization is still limited, despite improvements in needle and ultrasound design. Several new technologies to confirm needle tip positioning have been proposed, including adding a piezo crystal to the needle, fiber-optic hydrophone, electromagnetism,²⁷ and microultrasound.²⁸

Turbid materials and biological tissues greatly scatter light. The primary cause of scattering is a refractive index mismatch between the scattering particles (eg, cell membranes, extracellular media, collagen and elastin fibers, and background medium).²⁹ The fascia is of a dense sheet of connective tissue made of mostly collagen, which causes a diffuse scattering of light with OCT imaging and blocks light penetration into deeper tissues. While the needle travels from the muscle layer to the fascia, interfascial tissue may first not be visible due to this blockade. However, as the needle penetrates the fascia, the interfascial fat can be identified by bright speckled reflections on OCT images. In this study, the muscle tissue, fascia tissue, and interfascia adipose tissue were distinguishable with the side-looking fiber needle SSOCT system. Both ex vivo and in vivo SSOCT images show consistent characteristic features (Table 1). We found that the OCT imaging penetration depth in muscle tissue is more significant than in the fascia (Figures 2 and 3). This is because muscle tissue has a layered structure and a nonuniform signal distribution, whereas fascia tissue includes fat and nerves, which leads to strong-scattering and homogeneous signal distribution. Therefore, fascia and surrounding muscle tissue can be easily identified with SSOCT images after brief training with regard to the criteria and images. With these criteria, we demonstrated the sensitivity of this technique to be 0.91 (range 0.73–1) and specificity = 0.90 (range 0.75–1) with a diagnostic accuracy of 0.88.

The angle of needle insertion can affect imaging morphology (Supplemental Digital Content 3, Figure 2, https://links.lww.com/AA/D262). The more parallel to skin the needle entry, the more prominent the layer structures become, and the intensities decayed more quickly in the fascia layer compared to the muscle tissue when the insertion angle of the needle probe was beyond 30°; this insertion angle is usually used in clinical settings for fascial blocks. A statistical difference in the attenuation coefficients indicates differentiability of images between the fascia and muscle tissue (P < .001) in both 0°–30° (Figure 4B) and 30°–60° (Figure 4E) needle angle. The AUC (Figure 4C, F) also indicates that the attenuation coefficients have a high discriminatory capacity for discriminating between the fascia and muscle tissues. Thus, OCT imaging with the quantitative calculation of the attenuation coefficients can further aid objective identification.

Blood would affect OCT imaging. When the light is shining into the blood, the OCT image shows a completely white screen due to highly reflected blood. The absorption coefficient in water is only 1.2 per cm at a 1310-nm wavelength³⁰ (ie, the wavelength used in OCT light source). Therefore, the image quality would not be severely influenced when our needle is filled with anesthetic fluid or saline. However, injecting normal saline will change the tissue structure causes changes in the OCT image. In the same regard, the use of saline or local injection to confirm needle position under ultrasound posed problems and increase difficulty in identifying the correct tissue layer. For example, injectate may blur ultrasound imaging of surrounding tissue, especially if the air bubble was injected along with saline or if target depth was increased, thus hindering ultrasound penetration. We hope that by utilizing OCT, which can provide microscopic anatomical imaging from a needle tip, to confirm needle tip position, saline or local hydrodissection would not be necessary.

There are still several limitations to our study. First, the penetration of OCT imaging is approximately 1.5 mm, which is the depth range from the OCT probe tip that signals from tissues are recorded. Since in needle puncture procedure, the tissue is just around the needle tip; thus, 1.5 mm OCT imaging depth is large enough to obtain tissue-specific patterns. However, ultrasound guidance is still needed to identify the anatomy correctly. Second, the fascia space can be identified with human interpretation for OCT images but with high interexpert variation and makes the decision greatly divergent. Artificial intelligence aided interpretation may be used in the future to increase diagnostic accuracy. Third, a small number of study animals were used. However, needle punctures were deliberately distanced from each other to minimize potentially false reproducibility. Moreover, the linear mixed-model analysis indicated that there was no significant difference in the OCT signals obtained among the 10 different punctures in 3 pigs. Fourthly, the current system required a large gauze needle for OCT imaging. Further technological improvement is needed for an all-in-one system to incorporate OCT into small gauze needle stylet (eg, 22 gauze).

In conclusion, this study introduced a novel needle imaging probe method to identify the transversus abdominis fascia layer in real-time. Quantitative calculation of the attenuation coefficients can further aid identification, providing direct confirmation of the tip position, and decreasing the need for needle repositioning. Although the angle of needle insertion affects imaging morphology, it is easy to distinguish fascia and muscle tissue with OCT images once the insertion angle is beyond 30°; this insertion angle is usually used in clinical settings for facial blocks. Although our findings indicate that the feasibility of the SSOCT system to identify the TAP and the surrounding tissues and such methods may be used in clinical practice, additional studies with other fascia plane blocks or peripheral nerve blocks and human subjects are still required.

ACKNOWLEDGMENTS

The authors thank Chien-Hsun Chen for his technical support.

DISCLOSURES

Name: Wei-Nung Teng, MD, PhD.

Contribution: This author helped conceive the presented idea, perform the experiments, discuss the results, and contribute to the final manuscript.

Name: Meng-Chun Kao, PhD.

Contribution: This author helped perform the experiments and computations and contribute to the final manuscript.

Name: Chien-Kun Ting, MD, PhD.

Contribution: This author helped conceive the presented idea and contribute to the final manuscript.

Name: Wen-Chuan Kuo, PhD.

Contribution: This author helped perform the computations, verify the analytical methods, discuss the results, and contribute to the final manuscript.

This manuscript was handled by: Thomas M. Hemmerling, MSc, MD, DEAA.

Supplemental Digital Content

• AA_2020_10_14_KUO_AA-D-20-00643R2_SDC1.jpg; [Image] (149 KB)
• AA_2020_10_14_KUO_AA-D-20-00643R2_SDC2.wmv; [Video] (5.73 MB)
• AA_2020_10_14_KUO_AA-D-20-00643R2_SDC3.jpg; [Image] (586 KB)