Technological advancements continue to redefine upper-limb prosthetic solutions and potential outcomes. It is the choice of the clinician to either ignore or resist these new innovations or embrace and welcome them. Many of the technological advancements remain in research and development stages for extended periods before being released into the general population for clinical use. Once these advancements become commercially available, widespread adoption requires still more time. Several of these currently developing advancements involve surgical intervention beyond the initial amputation and postsurgical management. Targeted muscle reinnervation (TMR) and osseointegration (OI) are two surgical procedures that will have an increasing impact on clinical fitting of upper-limb prostheses. Pattern recognition (PR) and imbedded electrodes offer additional improvements in the functional control of externally powered devices and will require clinical understanding and acceptance. Three-dimensional (3D) printed solutions represent another technology movement that is beginning to have substantial presence in the upper-limb prosthetics community and must also be considered. This article summarizes the clinical considerations associated with these recent and continued technological advancements on upper-limb prosthetic care.
TARGETED MUSCLE REINNERVATION
TARGETED MOTOR REINNERVATION
Targeted muscle reinnervation is an elective surgical procedure that has been performed on humans since 20021 and has even been referred to as “routine” in recent publication.2 There were several initial goals of the TMR procedure and others that have since been discovered. Primarily for higher level upper-limb amputations (transhumeral or shoulder disarticulations), the concept was intended to create additional electromyographic (EMG) sites that would provide intuitive, physiologic, improved control of myoelectric prostheses. These additional sites were gained via surgical denervation of nonutilized muscles with subsequent coaptation of the peripheral nerves into alternate muscle bellies that can then provide signals for distal limb movements.
For example, TMR for the transhumeral user typically involves denervating the short head of the biceps brachii muscle by removing its branch of the musculocutaneous nerve and reinnervating that same motor point with the terminal end of the median nerve. This reinnervated short head of the biceps brachii muscle would then contract when the user's intent was to close his/her hand (performing contractions germane to normal median nerve function). Likewise, the lateral head of the triceps brachii would be denervated by removing the proximal radial nerve and then reinnervated with the distal branch of the radial nerve. Similarly, this lateral head of the triceps brachii would then contract when the user's intent was to open the hand (performing contractions germane to normal distal radial nerve function). Collectively, the EMG signals from the newly reinnervated muscles (short head of the biceps and lateral head of the triceps) can be used to close and open a prosthetic terminal device. Because the native long head of the biceps, along with the long and medial heads of the triceps, had not been altered, they remain available to be used for physiologically controlled EMG input for prosthetic elbow flexion and extension, respectively. Ultimately, the user has now gained a means of controlling the elbow and hand functions similar to preamputation physiology.
Conceptually, at shoulder-level amputations, the surgical procedure is analogous to that described above, with the major differences being that the denervated and reinnervated muscles are of the pectoralis region. This may vary slightly depending upon whether the individual truly had a shoulder disarticulation or is being fit with that style of prosthesis with a limb that is actually a forequarter (shorter) or humeral neck (longer). The exact muscles chosen for reinnervation may differ, but with similar expected outcomes.
Novel clinical considerations associated with both the transhumeral and shoulder disarticulation level fittings include the increased number of electrodes and electrode sites and the ability to train the individual to use those sites to control multiple motors or degrees of freedom (DOFs) simultaneously. They further include the utilization of newly designed components capable of receiving control input from multiple signal sources and the ability to instruct individuals with higher levels of amputation to use intuitive actions (via EMG) to control corresponding hardware such as elbow, wrist, and hand components.
Unplanned sensory reinnervation was observed in the first TMR recipient as sensory components of the reinnervated nerves spontaneously grew into the individual's subcutaneous tissue. After that discovery, some of the subsequent TMR surgeries also replicated this targeted sensory reinnervation, a characteristic that created the possibility of being used for closed loop feedback from the prosthesis.1 Although preliminary data suggest that TMR may also prevent neuroma formation by providing a location in which the distal, cut peripheral nerve can grow,2 additional research in this area is ongoing.
In addition, TMR for more distal amputation levels has also begun. For example, TMR at the transradial level is being explored as a means of better differentiation of control signals at wrist and terminal device components. Traditional “direct control” of transradial prostheses has been via wrist extensors and wrist flexors controlling the opening and closing of the prosthetic terminal device, respectively. With the acceptance of multiarticular hands in the field of prosthetics, an improved means of selecting hand grasps is warranted. This may be accomplished through the use of transradial TMR or other implantable devices that can decipher more specific hand motions rather than just the groups of wrist flexors and extensors.
Although the ability to provide simultaneous myoprosthetic control after TMR has been shown, there is little evidence that this is used spontaneously by prosthetic users outside of the therapy or laboratory settings. Even if the user does not perform actions simultaneously, a considerable benefit from the surgery is that it provides additional data in the form of EMG signals that are specific to actions of the distal amputated limb, which would not have been present otherwise. These points of data may be beneficial when used in conjunction with advanced control algorithms that can now differentiate signals more specifically to the intended action (motor movement).
There are limited drawbacks to TMR surgery for individuals with upper-limb amputation seeking improved control of their myoprosthesis(es). As in any surgical procedure, there are risks of infection, postoperative pain, and paresthesia. In addition, phantom pain may temporarily increase but subsequently returns to its baseline level.1 Those muscles with nerves being deinnervated are no longer performing mechanical function to the body (i.e., moving a segment), and thus, the use of these as hosts for EMG signals has no associated functional deficits. Having TMR surgery does not prevent the user from switching to a body-powered design if that was a plausible option presurgically. Educating the patient, especially in the case of acute TMR, is essential for understanding the expected outcomes.
Algorithms for detecting differences in muscle contraction patterns have been in existence for many years, as has the investigational use of these algorithms in applying these patterns to EMG control for prosthetic use. However, the use of these PR algorithms has had little attention until the past decade. The advent of TMR, commercially available multiarticular hands, and development of multiarticular arms from federally funded grant research efforts has led to an increased plausibility of using PR for myoelectric control of prostheses outside of the laboratory.
Direct control of multiarticular systems requires that the user perform a means of “mode selection” to change from one DOF to another (i.e., wrist to hand). This is often time consuming and cumbersome to the user. With PR, the user is able to directly select the DOF and direction of control. This feature provides more intuitive prosthetic control. Pattern recognition is able to do this by examining the signals from multiple EMG sources and classifying them into motions within the prosthetic system (i.e., wrist pronation/supination and hand close/open) in addition to a “no motion” classification. To do this, the system requires that additional electrodes are placed around the residual limb for the necessary data input. As previously discussed, TMR surgery provides additional data points, which may help further separate the signals into their distinct classification.
From the standpoint of current clinical considerations, there is only one commercially available PR system for prosthetic use. When using this system, distinguishing which classification pertains to which motion begins with the user of the prosthesis providing his/her signals to the system. The user is thus teaching the PR system which EMG patterns they wish to elicit to perform particular prosthetic movements. Once the system is able to differentiate the intended signals into distinct classifiers, the user is able to perform tasks without the need to select the desired DOF. “Calibration,” as it is called for this commercially available system, can be quickly performed whenever necessary and can include up to five different sets of input. This is beneficial when the prosthesis will be used in different positions in-space, thus creating different force-couples on the arm/body and possibly requiring different EMG signals for these DOFs.
Other claims with respect to the benefits of PR include the following:
- Decreased cognitive load for the patient
- Reduced consequence to electrode location
- Enhanced proportional control
- Decreased adjustment time
- Fewer return visits for patients
- Viable control strategies for individuals with low amplitude contractions
- Previous non-myoelectric-candidates can be users of myoelectric prostheses
It is important to note that there is little published evidence regarding these claims. Furthermore, these claims may be interpreted differently depending upon the clinician and his/her particular approach to fitting.
Although more than 350 fittings have reportedly occurred to date, there are a few obstacles that must be addressed regarding the use of PR. Simultaneous control of multiple DOFs is not yet perfected within these systems. That said, owing to their enhanced processing speed, switching from one DOF to another is sometimes unnoticeable and has been termed “seamless sequential control.”
Furthermore, selection of grasp/grip patterns cannot yet be directly controlled via PR on all multiarticular hands. At present, with some of the hands, grip patterns can be directly entered using PR while with others; the mode/grasp selection schemes are those that are designed by the specific manufacturers of the multiarticular devices and rely on traditional signal triggers (i.e., co-contraction, hold open, etc.).
Additionally, the size of the hardware, including the number of electrodes required, power consumption, and cost are all considerable. This too will be improved upon as second generations of systems develop and competition arises in the market. In addition, with the absence of existing L-codes, PR may not be viable for all users. It is well known that the etiology of upper-limb amputations is different from that of lower-limb amputations, equating to variability in funding resources and access to upper-limb prosthetic technology. Ultimately, the benefits of PR systems should be an available consideration for all upper-limb prosthetics users.
Whereas TMR and PR represent advancements that are now commercially viable treatment options, implanted electrodes represent an advancement in upper-limb prosthetic rehabilitation that is still confined to research settings. In contrast to TMR, where peripheral nerves are routed to muscle bellies where their signals can be recorded by surface myoelectrodes, implanted electrodes are surgically brought directly to the target muscles themselves. By doing so, this strategy seems to address many limitations associated with surface electrodes, including external signal interference, the deleterious effects of socket migration, and the inability to harvest signals from deeper muscle bellies.3 Signals harvested from such electrodes will need to be transferred to the external prosthesis through percutaneous leads or wireless signals. Additional clinical considerations will continue to emerge as this technology progresses toward commercialization. However, at a very fundamental level, any individual pursuing such implantable technology will need to have access to the appropriate externally powered prosthetic systems to derive any benefit from the procedure.
In addition to implanted motor electrodes (i.e., electrodes that record signals that direct the motor control of the prosthesis), efforts are under way to use implantable cuff electrodes to translate sensory data from the prosthesis to the body's sensory pathways.4 As with implantable motor electrodes described above, this technology has yet to reach commercialization, but promising reports of extended human trials have been published. With regard to clinical considerations, the effect of sensory transmission from prosthesis to end user could have tremendously beneficial effects on the use, compliance, and functional benefits associated with external limb prostheses. However, as described above, such surgical interventions would need to be reserved for those patients with guaranteed access to externally powered prosthetic systems capable of obtaining the sensory data to transmit to the targeted sensory nerves.
Direct skeletal attachment, as it was formerly known, was being performed in the post–World War II era for treatment of lower-limb amputations.5 However, the technology that is being applied today is more attributable to the direct results of work performed in Sweden throughout the 1950s, 1960s, and 1990s. This work on OI, sometimes referred to as percutaneous or transcutaneous implantation, continues in Sweden today, with several other regions including the United Kingdom, the Netherlands, Australia, Germany, and the United States becoming involved after the turn of the century.6 Although most of OI procedures have occurred in relation to lower-limb applications, particularly at the transfemoral level, there have been numerous reports of OI procedures performed on the upper limb. Cases involving thumb absence, partial hand, transradial, and transhumeral amputations have all been reported.7 Prosthetic systems for these patients have included passive, body-powered, and externally powered designs.7
Until very recently the US Food and Drug Administration (FDA) had not approved of OI in human subjects in the United States. Most of the individuals in the United States who have had OI procedures have had to travel to other countries to receive the implant. Many of these have been at the user's expense, which is quite a costly procedure.8 For these individuals, the possible benefits of the procedures, including a socketless fitting, improved comfort, improved proprioceptive control, and increased function, far outweigh the associated risks of infection around the percutaneous exit of the implant, a possible lack of robustness of the associated components, bone fracture, and lengthy rehabilitation processes.9
With regard to future clinical considerations, for the passive fitting of upper-limb prostheses, all indications are that the devices will provide outstanding results. Finger prostheses at the middle or proximal phalanx, as well as the thumb at the metacarpal phalangeal joint, will provide the benefit of a solid attachment with the coupled movement of the proximal joint(s). Passive devices at higher amputation levels will provide similar benefit and should also decrease lost motion and discomfort historically associated with the limb/socket interface.
Body-powered fittings of upper limbs with OI have also been described as a viable option. Because these cable-driven devices rely on precise capturing of the limb motion, OI would seem to represent optimal excursion recovery. The prosthetist will likely need to give particular attention to provision of the appropriate attachments for the customary baseplate/retainer, etc., and this will most likely be accomplished in a frame-type design rather than on the outside of the hard exoskeletal forearm or humeral section.
Externally powered fittings have also attracted considerable attention in academic publication. An integrated OI prosthesis controlled through a combination of TMR and PR has been described.10 A separate center has described a case study in which a transhumeral OI case has been fit with an externally powered prosthesis controlled via implanted motor electrodes routed through the percutaneous abutment and processed using PR algorithms.11 The system also included a single implanted cuff electrode providing sensory input from the device.11
Externally powered and hybrid systems have the detriment of being heavier than their body-powered and passive counterparts. It will take time to determine if such variables as increased range of upper-limb motion, velocity of movement, augmented force couples applied in many positions in space, and the lifting of objects will have deleterious effects on the integrity of either the surgical abutments or residual bony structures of the upper-limb user.
One distinct drawback at the time of publication is the lack of a consistent, clear position from the FDA with regard to where the implanted device terminates and whether the components of external limb prostheses will need to meet the requirements of an FDA Class III medical device. Although several prosthetists in the United States have proceeded with fitting individuals who have gone abroad for an OI procedure or who have received with a custom implant outside of FDA oversight, the liability of fitting such a device is still unknown. Currently, these systems have their own coupling piece, not unlike the pin used in a locking mechanism, which are designed to connect the implant to the external prosthesis. However, the loading limitations of these custom domestic implants and internationally obtained devices are seldom established, limiting the ability of a clinical prosthetist to design and provide a prosthetic solution that is known to be safe for the user.
To date, the FDA has issued a single humanitarian device exemption and approved a single feasibility trial with regard to OI procedures, both of which pertain to transfemoral applications. However, feasibility trials for transhumeral applications have been submitted to the FDA and upper-limb procedures will reach the United States in the near future.
3D PRINTED SOLUTIONS
One of the most rapidly advancing forms of technology has also found new applications to the field of upper-limb prosthetics. As a method of manufacturing, 3D printing has broad potential benefits across the health care field. For the upper-limb prosthetics community, 3D printing has been a topic of much discussion. However, until late summer of 2014, much of this discussion had been without the input of the prosthetist.
Most of the publicity surrounding 3D printing has been that of devices that were made for children at a very low cost in comparison with the higher-cost devices that are provided by a certified prosthetist. Although the media has conveyed its position on 3D printed devices and their utility, the manufacturers of these devices have stated otherwise. There is no direct comparison between what certified prosthetists have been providing for their patients and the 3D printed devices that are being primarily provided for children. Most of these 3D printed prostheses have been for children with transcarpal-level deficiencies that still have wrist motion to drive the fingers shut in a tenodesis-style motion. Often, these devices come in a kit and can be assembled by the child and his/her family, but sometimes, these completed devices are provided by institutions or celebrities.
The challenge to certified clinicians is whether and where to express concerns over the unregulated provision of such devices. For example, in certain states, licensure laws established to ensure patient safety assert that one must be licensed to provide a medical device. Alternatively, experimental devices may be provided to research participants who have provided informed consent as part of an experimental protocol that has institutional review board approval and all other appropriate regulatory research approvals. However, if licensed prosthetists exercise their rights to discourage the provision of unregulated and potentially unsafe devices by untrained, noncertified, and unregulated personnel, their actions may be frowned upon by the community at large as self-interested. Thus, a delicate balance must be made to educate the general population, ensure the safety of end users, and proliferate the potential benefits of these developing, creative but largely untested devices.
There are no reported studies on the robustness of the devices, although it has been reported that many of them will fail at significantly lower forces and cycles than any commercially available prosthetic components. The extent to which such designs represent a true threat to established prosthetic care models is uncertain. However, this will certainly change with time, and the prosthetic community must become involved in the development of these devices for the benefit of patients with upper-limb amputations and limb differences.
The materials and applications of 3D printing are ever evolving. Many of the principles for shape capture are well understood by prosthetists. From as early as the late 1800s, people have taken measurements and tracings of their residual limb(s) and sent them via mail to a manufacturer of their prosthesis. In return, they received a device made to measurement. While shape capture has evolved into methods far superior to a paper tracing and tape measure, this historic technique is still the case for some lower-limb sockets, upper-limb finger prostheses, and lattice-style coverings.
Computer-generated models have been used in prosthetics for decades. These same types of models can be used for the manufacturing of 3D printed components. The field has yet to define what types of prosthetic components will best use this technology. Semicustom and custom socket designs are being created via computer-generated models. Mechanical hands and externally powered hand/arm systems are being designed. “One-off” components are also being made and tested for “proof of concept.” This is viable due to the low cost of some 3D printing techniques. Unlike the proposed use of magnetic resonance imaging techniques for limb shape capture; scanning is much more benign, cost-effective, and viable to the patient. In addition, these concepts have great potential in developing countries with limited resources.
Many of the major innovations within upper-limb prostheses originate outside the field of prosthetics. With the exception of TMR, the innovations of PR, OI, and 3D printed solutions are technologies that are advancing in other professions with growing applications in prosthetics. All of these have shown merit in upper-limb prosthetic rehabilitation and will continue to do so with advanced surgical techniques, improvements in signal acquisition and rendering, innovative manufacturing processes, and material applications. Prosthetists that fit individuals with upper-limb prostheses must embrace these and other emerging technologies in order to stay current with the demands of the population they serve.