Smart needle electrical bioimpedance to provide information on needle tip relationship to target nerve prior to local anesthetic deposition in peripheral nerve block (USgPNB) procedures

USgPNB is one example of a procedure that could benefit from a smart needle, a hypodermic needle integrated with an impedance sensor, to determine needle tip location within the body. Limitations remain with the inability to objectively identify the relationship of the needle tip to the nerve prior to injection during USgPNB as mentioned previously. To address this limitation, a hypodermic needle integrated with an impedance sensor was developed that could potentially identify tissue at the needle tip thereby providing objective information indicative of needle tip location. Electrical impedance was identified as a feasible parameter, as impedance has previously been used successfully to detect tissue types in diverse clinical scenarios. The possibility of distinguishing intra-neural from extra-neural tissue based on bioimpedance has already been demonstrated and Kalvøy’s group have made novel investigations regarding needle guidance using bioimpedance [3]. However, additional data is needed to confirm the sensitivity and reproducibility of the technology.

The ratio of neural tissue to non-neural tissue between the different nerves and between different needle positions in the same nerve could be a substantial source of variation in bioimpedance trends for at-nerve and in-nerve test sites. Moayeri et al. quantified the neural and non-neural components of the brachial plexus and the sciatic nerve and investigated how the ratio of neural to non-neural tissue changed from proximal to distal. As the nerve leaves the spinal cord, the density of an epineurium (composed of stroma and connective tissue) decreases, but the total volume increases. Thus, the ratio of neural to non-neural tissue is approximately 1:1 in proximal plexuses, 1:2 in distal plexus and a peripheral nerve might contain up to 70% of connective tissue [5, 6].

This means more stroma and connective tissue is present within the epineurium of the brachial plexus the further away the nerve travels from the spinal cord. As the vagus nerve was used to obtain bioimpedance data from in this animal study, the proximal ratio of neural to non-neural tissue is more relevant for this data set as the vagus nerve lies proximal to the brachial plexus.

The sciatic nerve is similar to other peripheral nerves with outer epineural tissue surrounding fat and connective tissue and internal nerve fascicles surrounded by perineurium. The sciatic nerve is divided into tibial and common peroneal compartments in humans. Each compartment is composed of fascicles enveloped by their own epineural tissue. The two compartments, within the sciatic nerve, are separated by a connective tissue-adipose septum, known as the Compton Cruveilhier septum, which runs internally, along the length of the sciatic nerve [4]. The sciatic nerve is flat and large in shape proximally (closer to the spinal cord) and becomes more oval and smaller distally (as it moves further away from the spinal cord). This is related to the size and architecture of the brachial plexus, which changes from an oligofascicular (few fascicles) to a multifascicular pattern as it travels distally [4]. The sciatic nerve is multifascicular throughout and has a different microanatomy. The absolute amount of neural tissue decreased from proximal to distal whereas the non-neural tissue remained about the same. This decrease is attributed to branching off nerve tissue to the upper leg muscles. Consequently, neural to non-neural tissue ratio changed, from proximal to distal, from 2:1 to 1:1 [7, 2].

The first indication of nerve bioimpedance was obtained during the pig carcass study performed at a local abattoir. Nerve bioimpedance could not be obtained during in vitro studies as nerve tissues disintegrate and are absorbed by surrounding tissues within approximately 2 hours of death of the animal. Therefore, nerve tissue is not present in meat. For this reason, it was necessary to travel to a local abattoir and gain access to a pig carcass as soon as possible post-slaughter. The animal used for the pilot study was made available to us within 10 minutes of slaughter. All standard operating procedures for slaughter of the animal was adhered to and remain unaffected.

Prior to device validation in an anesthetised animal model, in vitro bioimpedance measurements were performed on pork meat samples. Results demonstrated a clear distinction between muscle and subcutaneous tissues based on bioimpedance recordings. In pigs, femoral nerve, sciatic nerve, and brachial plexus (vagus nerve) sites were used to obtain bioimpedance recordings from subcutaneous tissue, muscle and at-nerve and in-nerve tissues. Data was obtained from 7 animals initially using the frequency range of 10 – 100 kHz, voltage at 10 mV, current range 100 nA – 1 mA on the first two animals and a change in frequency range to 100 Hz – 1 MHz for the other five animals. The frequency range for bioimpedance measurement was widened to ensure as much useful data as possible was collected thus increasing the probability of identifying unique bioimpedance trends for tissues at certain frequencies within the range. If necessary, the optimum frequency range could be narrowed again after identification and characterisation of unique tissue bioimpedance profiles with the wider range.

Fig.2:

Fig.3:

The vagus nerve (a nerve lying closer to the spinal cord than the brachial plexus) at the side of the neck was easier to access than the brachial plexus. For this reason, the side of the neck was open dissected to the vagus nerve instead of exposing the brachial plexus. Therefore, bioimpedance data was obtained from the different tissues around the vagus nerve on day 4 – 7 of the animal study.

Femoral, sciatic, brachial and vagus nerves were excised from the nerve locations where the bioimpedance measurements were analysed to validate the nerve type. These slides were prepared in histopathology at Cork University Hospital (CUH) with Haemogylin Van Goss Stain on cross section and longtidutional peripheral nerves. These stained peripheral nerve slides were observed under an electron microscope in life sciences laboratory at Tyndall National Institute (TNI). ImageJ software was used, and clear images of peripheral nerves were exported and analysed. Sketching around the parameters of connective and neural tissue separately the ratios of connective to neural tissue in each peripheral nerve could be calculated.

Bioimpedance results from the ultrasound guided tissue identification procedure in the pig carcass did not match previous in vitro experimental results. In particular muscle bioimpedance under ultrasound guidance was uncharacteristically high. The tissues during this procedure were identified from the ultrasound image by a consultant anesthetist, an expert in interpretation of ultrasound images of human anatomy.

This was the first opportunity for the expert in human anatomy to ultrasound image porcine anatomy and attempt to locate and identify different tissue structures. The bioimpedance results on open dissection of the tissues fig. 4 were much closer to what was expected from previous in vitro results where muscle had low bioimpedance.

Fig.4:

This was the first and only attempt to date at obtaining bioimpedance of porcine tissues in a freshly slaughtered animal. During bioimpedance measurement of nerve tissue it was noted how fragile and friable the nerve structure was. The nerve was already beginning to disintegrate within the surrounding tissues as measurement was performed thus the bioimpedance results obtained should only represent a rough estimate of the bioimpedance of nerve in live tissue. This pilot study however was not only allowed an indication of bioimpedance of near to live porcine tissues but also allowed testing of a study protocol before use on a live anaesthetized pig. Insights gained from this pilot study enabled refinement of the protocol for bioimpedance measurement of in vivo tissues and was an important step in progressing to a live anaesthetized pig model [11]. The decision to continue using open dissection method on the in vivo pig models was to ensure accuracy of data for this preliminary study as the needle prototype is small in length.

On observation of the tissue bioimpedance data from live tissues in the first two animals evaluated, it was found to be necessary to assess the repeatability and reproducibility of data provided by the smart needle in live tissues.

Results from these studies, along with pre- and post-testing impedance measurement of NaCl solutions, demonstrated that the data obtained using the smart needle was repeatable and had good reproducibility for most tissues. However, some variation between puncture sites was observed during reproducibility testing [11]. This was expected due to the heterogeneity of biological tissues and was accounted for by taking three bioimpedance measurements at three different puncture sites within proximity to each other, from each tissue at each test site [2].

With respect to this research and the intended application of the technology described, the bioimpedance profiles of the in-nerve and at-nerve locations were of most significant interest. It was found that at times no differentiation between these two locations was provided by bioimpedance data. When differentiation between these locations was observed the femoral and vague nerve sites displayed higher EI for at-nerve than in-nerve while the opposite was found for the sciatic nerve site (at-nerve had lower EI than in-nerve). It must be noted there were exceptions to this trend at all sites also [11]. Interestingly, despite the sciatic nerve data obtained indicating the at-nerve site had lower bioimpedance than in-nerve, which was the inverse of what was found at the femoral and brachial nerve sites. This data correlates with the findings of a previously published study on distinguishing intra-neural from extra-neural needle placement using a nerve stimulator to measure impedance [1, 2].