Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

A biosynthetic thumb prosthesis

A biosynthetic thumb prosthesis npj | biomedical innovations Article https://doi.org/10.1038/s44385-025-00031-z Check for updates 1 1 2 2 2 2 Sachi Bansal ,GraciaV. Lai , Marcus Belingheri , Amber Q. Kashay ,JinyoungKim ,AlyssaTomkinson , 3 4 4 2 5 Samantha Herman , Keval Bollavaram , Brian K. Zukotynski , Sean Thomas , Nirbhay S. Jain , 4,5 4 1,2,4 Kodi K. Azari , Lauren E. Wessel & Tyler R. Clites In cases of severe damage to the extremities, the function and structure of compromised tissues must be replaced. If biological reconstruction using autologous tissue is not feasible, amputation and replacement with a synthetic prosthesis is often the next best option. Synthetic prostheses are limited, especially in their ability to restore skin sensation. Here we show a biosynthetic prosthesis that combines the versatility of titanium with the rich sensory capabilities of biological skin. The prosthesis recreates opposition pinch by linking motion of the prosthetic joint to that of the residual biological joint, and is enclosed in neurotized skin from the patient’s own body. We validated the biosynthetic thumb’s mechanical function on the benchtop and in a cadaver, and showed viability of the skin interface in an animal model. These results provide a framework for functional reconstruction of amputated digits using a combination of synthetic materials and biological tissues. The thumb is considered clinically to be the most important digit on the For some of these functions (e.g. actuation, mechanical strength), synthetic 1–4 hand, responsible for approximately 40% of hand function .The thumb’s components (motors, titanium) outperform the biological systems (mus- 5 11–13 core role is opposition, which is critical for grasp and pinch .Natural cles, bone) that they are intended to replace . However, other functions, opposition requires an ability to trace the tip of the thumb along the tips of such the skin’s ability to sense the environment and protect from infection, the remaining four fingers. In the intact hand, this ability is a direct result of have proven particularly challenging to replicate with synthetic 5 14–16 thumb’s position on the hand and complex anatomical structure .The components . As a result, it has been difficult to create synthetic digits 17–19 interphalangeal (IP) joint, which is the thumb’smost distal joint,plays a that look,feel,and behave like their biological counterparts . crucial role in functional opposition pinch, accounting for up to 31% of With no good option for autologous reconstruction, thumb prostheses 6,7 pinch strength and ensuring fingertip-to-thumb-tip contact .The rich face the challenges of restoring functional opposition pinch, attaching to the sensory information from the skin at the thumb tip allows for robust closed- hand, and providing natural touch sensation. Conventional external pros- loop control of pinch and grasp force, as well as perception of texture, slip, theses, such as the GripLock Finger (Naked Prosthetics, Olympia, WA, and temperature . The skin also acts as a barrier to external insult such as USA), are effective in restoring opposition pinch but cannot provide cuta- 20,21 infection. neous feedback to the nervous system . This type of prosthesis typically Catastrophic injury to the thumb is fairly common: approximately 6 straps to the hand, which does not provide a rigid attachment and adds bulk million digital amputations occur annually worldwide, one third of which that can obstruct hand function. Percutaneous osseointegrated implants include loss of at least one thumb joint . Because the thumb plays such a provide a low-profile option for rigidly attaching prosthetic thumbs to the critical role in hand function, its loss can have tangible effects on lifestyle . residual bone; however, osseointegrated thumb prostheses do not have a This is especially true in populations who rely on their hands for manual functional IP joint, and therefore cannot restore opposition pinch. Without labor, whichincidentally predisposes them to these very amputations; loss of neural integration these devices also cannot restore skin sensation, and the the thumb can inhibit their ability to perform work functions, creating percutaneous abutment creates a chronic infection risk .Even the most financial strain or necessitating a shift in career. More broadly, partial hand advanced of current clinical devices offer only partial restoration of function, amputation is linked directly to decreases in mental health and overall and none have successfully replicated the critical tactile feedback and pro- quality of life, with a large portion of the deficit related to functional tective properties of native skin. In fact, the shortcomings of existing impairment . While biological reconstruction using native tissues is typi- prosthetic technologies are significant enough that some patients choose to cally favored, it is not always possible. In such cases, we face the task of sacrifice another of their digits to restore thumb function (e.g. toe-to-thumb reproducing the function of the lost biological digit via synthetic prostheses. transfer). While this purely biological approach restores both opposition 1 2 Department of Bioengineering, University of California Los Angeles, Los Angeles, CA, USA. Department of Mechanical and Aerospace Engineering, University of 3 4 California Los Angeles, Los Angeles, CA, USA. Department of Physics and Astronomy, University of California Los Angeles, Los Angeles, CA, USA. Department of Orthopaedic Surgery, University of California Los Angeles David Geffen School of Medicine, Los Angeles, CA, USA. Department of Plastic Surgery, University of California Los Angeles David Geffen School of Medicine, Los Angeles, CA, USA. e-mail: [email protected] npj Biomedical Innovations | (2025) 2:22 1 1234567890():,; 1234567890():,; https://doi.org/10.1038/s44385-025-00031-z Article pinch and sensation, toe-to-thumb transfer is a technically demanding implanted component to be entirely covered in living skin. Complete cov- surgical procedure, provides a limited range of motion compared to the erage of the implant (in contiguity with the skin of the residual thumb) native thumb, and can lead to changes in gait due to donor site morbidity . ensures a sealed skin envelope, which is crucial to keeping the transposed Other patients forego prosthetic reconstruction altogether and instead adapt skin alive and preventing infection. This surgical context became the starting 24 1–4 to life without a thumb , which comes at significant functional cost . point for our design process; we deliberately chose to design the implant Synthetic “skin” and direct neural interfacing have shown promise around the surgery (and not vice versa), to increase the likelihood that the in experimental devices as a means of restoring cutaneous sensation, but prosthesis would be clinically viable and work synergistically with the body still cannot compete with the density, precision, or versatility of sensory to restore function. Through a series of cadaver dissections, we developed 15,25,26 endorgansinbiologicalskin . In addition to sensing, commu- and demonstrated a new surgical approach that combines established flap nication of high-density information from external sensor arrays to the techniques in a new way such that they can be used to cover our fully- nervous system remains a persistent challenge .Direct nerve stimu- synthetic implant core (Fig. 1a). lation has been useful in restoring touch sensation from a single sensor The intramedullary canal of the residual proximal phalanx was first that localizes to a region (e.g. the whole thumb-tip), and some pre- identified as the primary site for bony anchoring, with the driving link liminary evidence has been provided that such stimulation can evoke secured to the volar aspect of the metacarpal using two biocortical 28–30 natural touch perception . While this approach has been shown to screws (M2) through the metacarpal plate. We selected a combination improve task performance, direct nerve stimulation faces drawbacks of two cutaneous flaps to cover the device with innervated, vascularized including low resolution and precision, uncomfortable or unnatural autologous skin. The first of these is the RRF flap, which provides 31,32 sensation, and sensory adaptation . Current nerve stimulation surface area to surround the implant. A RRF flap measuring 8 cm by approaches also require either implanted electronics or percutaneous 4 cm was harvested from the forearm ipsilateral to the thumb ampu- transmission of power, which can pose challenges for long-term via- tation. The proximal end of the radial artery within the flap was clipped bility. Cutaneous sensory information has also been delivered directly and ligated; in a living patient, this would induce retrograde flow in the to the brain via intracortical microstimulation; however, this approach radial artery, allowing the flap to be perfused from its distal end without has faced challenges with stability and longevity and requires invasive compromising flow to the hand . The RRF flap on its vascular leash was 33,34 surgery that is difficult to justify for persons with thumb amputation . then passed through a subcutaneous tunnel in the wrist, tubularized As an alternative to purely-synthetic reconstruction, we propose an around the thumb implant, and sutured to the distal edge of the native approach in which the prosthetic thumb is reimagined as a hybrid system skin surrounding the residual thumb. The portion of this RRF flap over that selectively leverages both biological and synthetic subsystems where the thumb tip was then replaced with an innervated 1.5 cm × 2 cm first- each is most advantageous. Specifically, we describe a biosynthetic pros- dorsal metatarsal flap (FDMA flap) from the dorsal aspect of the index thesis that combines a titanium linkage with the patient’s own skin to both finger; this flap is used to provide coverage and restore sensation to large 42,43 enable opposition pinch and restore skin sensation. At the core of the device wounds on the tip of thumb . A complete depiction of the surgical is a four-bar linkage, which couples motion of the prosthetic IP joint to that technique in the context of the full thumb prototype is shown in Sup- of the biological metacarpophalangeal (MCP) joint; this design provides plementary Fig. 1. opposition pinch without the complexity of powered actuators and does not In these dissections, both flaps were successfully harvested, rotated on require attachment of tendon to synthetic material, which remains a sig- their respective vascular pedicles, and reattached to the residual skin at the 35–37 nificant unsolved problem . Although fully-implanted articulating base of the thumb. The skin flaps were large enough to cover a proxy implant 38,39 prostheses are common in orthopedics —and have even been proposed without excessive tension in the transposed skin. The defects created on the as a solution for leg amputation —their movement depends on attachment forearm and wrist were sufficiently small to be sutured closed by primary of tendon to metal or other synthetic material. In our system, the metal intention (Fig. 1a); the defect on the index finger from the FDMA flap is portion of the biosynthetic thumb anchors only to bone, providing robust usually closed via a split-thickness graft from the forearm . Because these long-term attachment and actuation. The biological portion of the pros- flaps are not typically used to cover synthetic implants, extrusion through 45–49 thesis comprises the patient’s own vascularized, innervated skin, sustainably the reconstructed skin was a significant concern . We addressed this by harvested from elsewhere on the arm and hand. Current endoprosthetic covering the non-joint portions of the device, which do not move relative to 38,39 devices are designed to be buried beneath a robust soft tissue envelope , the surrounding skin, in porous high-density polyethylene (HDPE) to 50,51 and therefore cannot be applied to treat pre-existing or traumatic promote skin ingrowth during the healing process . In order to prevent amputations . In contrast, our biosynthetic thumb incorporates transposed soft tissue impingement or infiltration that could disrupt link movement autologous skin onto the prosthesis itself, eliminating the need for existing and joint motion, we determined that future iterations will include flexible native skin at the site of reconstruction. This transposed skin becomes the silicone covers over all articulations (Fig. 1b). Silicone was selected for its outer surface of the prosthesis, protecting the body from infection and flexibility and history of use in orthopedic applications (e.g. Swanson leveraging innate cutaneous end organs to provide natural touch and Flexible Finger JointImplant,Stryker,Kalamazoo,MI, USA),and in temperature sensation. pacemakers to protect the device from the surrounding biological This manuscript describes comprehensive preclinical validation of the environment. biosynthetic thumb, addressing surgical, mechanical, and clinical feasibility. We first use a cadaver to establish and validate a new surgical paradigm for Optimization and validation of mechanical linkage transposition of the patient’s skin onto the surface of the device. We then The synthetic portion of the implant is a crossed four-bar linkage that present an optimization and evaluation of the device’s mechanics in free couples motion of the prosthetic IP joint to that of the biological MCP joint. space and in the context of opposition-relevant loading. We describe an Importantly, this motion is enforced through bone-anchored metal com- animal study of the proposed synthetic architecture in living skin, through ponents, which avoids the need for tendon-metal interfaces; such soft-tissue 35–37 which we establish long-term clinical viability of the skin-implant interface. interfaces are notoriously prone to failure in the body .This linkage was The manuscript concludes with a surgical and kinematic evaluation of the optimized to match a target trajectory derived from a published relationship complete implant in a human cadaver model. between MCP and IP motion in the biological thumb during opposition pinch , while accounting for geometric constraints of the anatomical Results envelope. An initial optimization yielded an “idealized” linkage design, with Surgical approach to covering the prosthesis with living skin a mean absolute percent error (MAPE) of 0.73% relative to the target angle To create a biosynthetic prosthesis with real translational potential, it was relationship (Fig. 2). However, the two links that formed the IP joint first necessary to develop a surgical approach that would allow the extended beyond the device envelope, posing a high extrusion risk. As such, npj Biomedical Innovations | (2025) 2:22 2 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 1 | Surgical approach and device design. a (left to right) Biological thumb with closed, and the system is allowed to heal. b The implant is made up of four parts: the intact native joints and soft tissue. The IP joint is amputated, and native skin covers thumb tip, structural link, driving link, and metacarpal plate. The linkage connects the residual proximal phalanx. The device is attached to the local bony structures the MCP and IP joints, such that flexion of the biological MCP joint creates flexion of including the residual proximal phalanx and metacarpal. The reverse radial forearm the synthetic IP joint. Static (non-joint) portions of the implant are covered in (RRF) and first dorsal metacarpal artery (FDMA) flaps are harvested and rotated porous HDPE, and the synthetic IP joint is protected by a flexible cover. A prototype about their pedicles to cover the implant. The donor and recipient sites are sutured was manufactured in Ti-6Al-4v. we constrained the links to remain within the diameter of the “structural Validation of skin-implant interface in an animal model link” (Fig. 1b) and re-optimized. This constrained optimization produced Although the proposed surgical techniques are part of established clinical the “clinically viable” linkage design, which had an MAPE of 15.2% from the practice, these flaps are not typically used to cover synthetic implants in an ideal angle relationship. Specifically, the slope of the MCP-IP relationship area of high motion. This introduces concerns of flap viability and potential was shallower, resulting in an IP joint that was 3.4° less flexed at the extrusion. To address these concerns, a rodent model was used to validate minimum MCP joint angle, and 4.4° more flexed at the maximum MCP long-term viability of a proxy implant in the context of a large tubularized angle (Fig. 2b, blue line). rotation flap. In a total of 23 rats, vascularized skin flaps were elevated, A prototype of the clinically viable linkage design was manufactured in tubularized, and rotated to a new orientation in a high-motion area near the Ti-6Al-4V, and attached at its distal end to a simple mechanical hinge hind leg (Fig. 3a). In 12 rats, the flap was wrapped around a proxy implant (proxy MCP joint). This prototype was then affixed to an adjustable loading (experimental group); the remaining 11 did not have implants (control frame, and loads were applied to a hook on the dorsal aspect of the thumb tip group). Proxy implants were two-part components shaped analogously to (Fig. 2a). In the absence of external loads, the proxy MCP joint and the the thumb implant and made of the same HDPE material (Supplementary device’sIPjoint were able to flex over a range of 70.3° and 29.7° from their Fig. 3). Two rats from each group were lost due to flap detachment caused by rest positions, respectively (Fig. 2b, green dots), and matched the predicted chewing. In the remaining 19 animals, flaps were allowed to heal for kinematic relationship with 2.84% MAPE. 3.5–11.25 months. After the sutures in each animal had dissolved, we 53–55 Thumb-tip loads during a typical grasp activity can exceed 80 N . assessed strain in the flapsasthe rodentswalkedfreelyacrossatransparent With the linkage locked at seven different proxy MCP angles (Fig. 2c), cage (Supplementary Movie 1); these trials served to ensure that the flaps ranging from −9.0° (slight extension) to 37.1° (flexion), we observed a deformed each time the animal walked, in a way that adequately represented maximum of 3.4 mm of deflection at the thumb tip under 82.4 N of force the large amount of skin motion in the human thumb. All flaps exhibited (Supplemental Fig. 2a). Angular stiffness about the base of the implant large strains, with averages ranging from 69% to −17% (Supplementary changed as a function of proxy MCP angle, with a maximum effective Fig. 4). deflection of 9.4° under 82 N of force perpendicular to the thumb tip, cor- Our primary measure of success in these experiments was lack of responding to a moment of 3.1 Nm about the proxy MCP (Supplementary extrusion; according to this metric, we observed a 100% success rate in the Fig. 2b). Stiffness of the device varied as a function of proxy MCP angle and animals with implants. All flapsinboth groupshealed well, andshowed no was highest with the proxy MCP and IP near the middle of their ranges of gross signs of necrosis or tissue damage (Fig. 3b). Qualitative histological motion (Fig. 2d). Deflection of the thump tip decreased the effective IP angle analysis in the experimental group revealed an implant surrounded on all by up to 5.87° under 82 N, shifting the MCP-IP relationship toward less IP sides by healthy dermal tissue, with some pockets of subcutaneous fat flexion (Fig. 2e). (Fig. 3c). By volume, tissue filled 94.3% ±3.85% of the implant voids. We did npj Biomedical Innovations | (2025) 2:22 3 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 2 | Benchtop mechanical testing results. a Experimental setup, in which the d Stiffness of the device as a function of MCP angle. Stiffness represents the amount prototype thumb is clamped into an adjustable load frame and loaded with the proxy of perpendicular force required at the thumb tip to produce 1 mm of linear deflection MCP joint locked at different angles. b MCP-IP angle relationships showing the of the thumb tip. e Experimental MCP-IP angle relationship in the presence of target trajectory, the idealized linkage design, the clinically viable linkage design, and different loads. Dots show average measured angle over 1.5 s, and solid lines show the experimental performance of the physical prototype (Supplementary Table 1). first-order linear regression fits. The dashed line represents model predictions for the c Deflection of the thumb-tip marker across seven locked proxy MCP angles. clinically viable device (identical to the solid blue line from (b)). not observe any relationship (R = 0.02) between ingrowth and post-op time flexion movement at both the biological MCP and synthetic IP joints prior to tissue harvesting (Fig. 3d). We found no significant difference in (Fig. 4b, Supplementary Movies 2 and 3) that was sufficient for functional skin thickness (p = 0.825, Fig. 3d, right) between the experimental flaps and opposition pinch (Fig. 5). We opted to characterize the transmission ratio of skin samples collected from the contralateral side of the epigastric region both the biological and biosynthetic systems according to kinematics (input (Fig. 3c, bottom), suggesting that neither the presence of the implant nor tendon excursion to output thumb-tip displacement) rather than kinetics tubularization of the skin flap caused thinning of the surrounding flap tissue (input tendon force to output thumb-tip force). Assuming that neither (Fig. 3d, bottom). We also evaluated tissue encapsulation, which we defined system exhibits meaningful frictional losses and that the principle of virtual as the proportion of the implant’s circumference covered by thick fibrous work therefore holds, any scaling of the kinematic relationship between connective tissue; by this metric we observed 100% encapsulation in all tendon excursion and thumb-tip motion across systems would imply an implants (Supplementary Figs. 5, 6). inverse scaling of the force relationships. When actuated by manually excursing the FPL tendon, the IP joint in Performance of the complete biosynthetic thumb in a cadaver the biosynthetic thumb moved 1.08° per 1° of MCP joint flexion (Fig. 4c, We used a cadaver to characterize opposition pinch after attachment of the green), compared to 1.12° per 1° of MCP flexioninthe cadaver’sbiological implant to the residual proximal phalanx. In the cadaver (Fig. 4), thumb thumb (Fig. 4c, blue). Across a range of 9 mm of FPL excursion, the MCP function was compared between two conditions: 1) the native biological and IP in the biosynthetic thumb produced 72.9% and 77.2%, respectively of thumb, actuated by pulling on the flexor pollicislongus(FPL) in itsnative the range of motion observed in the biological thumb. The cadaver’sbio- insertion (Fig. 4a, left); and 2) the biosynthetic thumb, comprising the logical MCP moved between 34.0° and 47.6° of flexion (13.6° range), and the uncovered implant anchored to the residual bone and actuated by pulling on IP moved between 24.8° and 40.7° of flexion (15.9° range). Meanwhile, the FPL in its post-amputation, transposed insertion (Fig. 4a, middle; across about 9 mm of tendon excursion, the biosynthetic MCP moved Supplementary Table 1). We also completed the reconstruction by covering between 45.5° and 54.8° of flexion (9.3° range), and the IP moved between the thumb implant with the proposed flaps (Fig. 4a, right); however, we did 9.7° and 19.5° of flexion (9.8° range). Although the slope of the MCP-IP joint not include this condition in our biomechanical experiments because as relationship was preserved in the reconstruction, the biosynthetic IP was confirmed radiographically, motion of the implant within the skin envelope around 15° less flexed at baseline than the biological IP. Through post-hoc produced an inaccurate representation of IP motion. We do not expect that radiographic analysis of the as-implanted linkage geometry, we attributed this relative motion will affect performance of the thumb in vivo, due to this shift to a 6 mm difference in the distance from the MCP joint center to ingrowth of the skin into the porous HDPE coating. In the cadaver, the two the metacarpal plate, between the optimized clinically viable linkage and flaps provided sufficient coverage to completely encapsulate the implant, cadaver implementation (Supplementary Fig. 7). Updating the linkage and the reverse radial forearm (RRF) flap donor site was small enough to be model to reflect this change resulted in agreement between the model closed via primary intention (Fig. 4a, right). When pulling only on the FPL prediction (Fig. 4c, red) and the empirical kinematics (Fig. 4c, green). Per tendon and finger flexor tendons, the fully-reconstructed thumb created 1 mm of FPL tendon excursion, we observed 1.05° MCP flexion and 1.22° IP npj Biomedical Innovations | (2025) 2:22 4 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 3 | Surgical design, post-op follow-up, and histological analysis from region (bottom). Bottom two samples are from a second animal (no implant), again animal study. a Rodent surgical technique for both experimental (with implant) and from the flap (top) and the contralateral epigastric region (bottom). d (top) Percent control (no implant) groups. b Post-op photos from immediate post-op through ingrowth versus time at tissue harvest, where each point represents one animal. tissue collection. c Representative histology. Top two samples are from the same (bottom) Smallest distance from external surface of epidermis to external surface of animal (with implant), taken from the flap (top) and the contralateral epigastric smooth muscle layer for the skin flap (with implant) and contralateral skin sample. flexion in the biological system and 1.03° MCP flexion and 1.09° IP flexion in FDMA flaps are commonly used to address hand, and specifically thumb, 56–58 the biosynthetic system (Fig. 4d). Correlation in joint flexion angle as a defects . Our results showed that sufficiently large flaps could be har- function of tendon excursion between the biological and biosynthetic sys- vested from these two sites to fully cover the device. Cutaneous sensation tems was high for both the MCP (r = 0.89, p < 0.0001) and IP (r = 0.99, from the reconstructed thumb tip is provided through the neurotized p < 0.0001) joints, indicating a high degree of similarity between pre- and FDMA flap, which naturally projects to the dorsal index finger . Utility of post-operative joint kinematics. this flap in restoring natural thumb sensation has been established over Based on the kinematic data, we calculated the instantaneous trans- decades of clinical experience . We note that sensation from the natively- mission ratio from input FPL tendon excursion to output thumb-tip dis- innervated FDMA flap will project initially to the back of the index finger, placement for both the biological and biosynthetic systems (Supplementary and that the degree of sensation may differ from the native, biological finger Fig. 7). Per 1 mm of FPL tendon excursion, we observed an average of pad tissue; however, despite these differences from the native milieu, 3.39 mm of thumb-tip displacement in the biological system (average patients show a high degree of sensation after reconstruction with an FDMA transmission ratio of 0.295), versus 3.66 mm in the biosynthetic system flap . To further enhance sensation, the RRF flap used for additional cov- (average transmission ratio 0.273). This implies a decrease in transmission erage could be harvested along with the antebrachial cutaneous nerve ratio of 7.6% in the biosynthetic thumb, implying that the FPLwould need to branch which could then be innervated at the recipient site by coaptation to produce an average of 7.6% more force across its range of excursion to create sensory branches of the distal median nerve ; this would increase surgical thesamepinchforceatthethumbtip, but would need to excurse 7.6% less to complexity, but may provide more natural sensation across the thumb. The produce the same amount of thumb-tip displacement. native skin covering is also expected to protect the patient from infection, which is a major concern for osseointegrated devices that breach the skin Discussion and create a permanent open wound. In this study, we demonstrated feasibility of a hybrid prosthetic architecture Flexion of the MCP joint caused flexionofthe IP jointinbothour that combines synthetic components and biological tissues to restore benchtop and cadaver studies. Although it was possible to optimize the function, sensation, and the protective qualities of skin in a reconstructed linkage mechanism such that its motion very closely replicated the biological thumb. Our new surgical approach allows for complete coverage of the MCP-IP joint relationships, it was necessary to deviate from this ideal to thumb in vascularized skin, with natively-innervated finger skin on the protect the MCP joint capsule and reduce the chances of extrusion. On the thumb tip. The synthetic linkage is capable of reproducing the joint kine- benchtop, the relationship between the proxy MCP and synthetic IP joints matics needed for opposition pinch, both in free space and in the context of showed close agreement with model predictions. In the cadaver, we the typical thumb loads. An analogous implant did not extrude or erode the observed that the slope of the relationship between the biological MCP and skin in a high-motion area of an analogous animal model. The prototype synthetic IP joints of the biosynthetic thumb closely matched that of the device supported opposition pinch kinematics in a cadaver when the MCP biological thumb, but that the biosynthetic IP joint angle was offset in the was actuated by the native tendon. extension direction compared to the biological thumb. This offset was The surgical approach by which the device is covered in native skin was caused by sensitivity of the linkage kinematics to changes in fixation geo- designed as a new combination of established techniques. Both the RRF and metry (Supplementary Fig. 8). To correct this, it may be possible to design npj Biomedical Innovations | (2025) 2:22 5 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 4 | Cadaver testing results. a Cadaver hand with biological thumb (left), bio- and shaded regions show +/−1 standard deviation. Predictions from the as- synthetic thumb (middle), and biosynthetic thumb with flap coverage (right). implanted linkage model are also shown for comparison (red). d IP (top) and MCP b Radiographic images showing the biosynthetic thumb with flap coverage in the rest (bottom) flexion angle as a function of FPL tendon excursion. Zero excursion position (left) and at full flexion (right). c MCP-IP angle relationship for the cadaver represents resting tension in the FPL tendon. Solid lines represent the mean of 10 hand with biological thumb and with biosynthetic thumb, when actuated only by trials, and shaded regions show +/−1 standard deviation. pulling on the FPL tendon. Solid blue and green lines represent the mean of 10 trials, patient-specific surgical guides that ensure the device is affixed to the correct We performed the full biosynthetic reconstruction in a cadaver limb, locationsonthe bone,ortoredesign the thumbtip basedon eachpatient’s demonstrating the ability of the reconstructed thumb to support functional bony anatomy (θ in Supplementary Fig. 8). Despite deviations from the thumb opposition pinch. Although the MCP-IP relationships of the device biological MCP-IP relationship, the reconstructed thumb was still able to in situ deviated from our observations during controlled benchtop testing, create functional opposition pinchinacadaver whenactuatedonly by our results showed meaningful IP joint flexion when actuated via the native pulling on the native tendons (Fig. 5). tendon. Importantly, this motion was created solely through relative The synthetic component of the reconstructed thumb maintained movement of bone-anchored components, bypassing the need for tendon- functional opposition pinch in the presence of high simulated pinch forces. to-metal interfaces . We grossly observed reduced motion of the IP joint Most importantly, under increasing loads, the MCP-IP relationship main- after covering the implant with the intended flaps, due to motion of the tained the same slope, but shifted toward less IP flexion; however, this shift implant within the skin; this further underscores the importance of skin was relatively small compared to the overall angular deflection of the IP ingrowth into the planned porous coating on the thumb tip and structural joint. Interestingly, we observed that non-linearities in the linkage link. We attribute increased FPL tendon excursion in the biosynthetic mechanics caused a jump in the mechanism’s resistance to deformation thumb to structural changes associated with reattachment of the FPL to the when the synthetic MCP joint was held at 30°; we suspect that this was base of the proximal phalanx, which is a common step in surgical thumb caused by local changes in stiffness as a function of the orientation of the amputation . This increased FPL-excursion-per-unit-IP-flexion also loading vector relative to the curved portion of the driving link. implies a slightly increased mechanical advantage between the FPL and the The miniaturized implant did not extrude in any rodents, providing IP joint in the biosynthetic thumb as compared to the native biological evidence as to the in vivo viability of our surgical approach and material architecture. Conversely, the mechanical advantage from input FPL tendon selection. Histological analysis revealed that the components remained excursion to output displacement decreased by ~8% in the biosynthetic securely anchored within the flap, aided by encapsulation and tissue thumb. This disparity can be attributed to differences in relative thumb ingrowth into the component’s pores. We did not see any significant segment lengths and increased motion at the MCP joint after transferring thinning of the skin surrounding the implant despite large strain in the flap the FPL tendon to the base of the proximal phalanx. Despite slightly during ambulation, suggesting that an implant made from similar material decreasing the mechanical advantage, this repositioning of the FPL insertion will remain stable within a skin flap, even in the context of large motions. is an important feature of the surgical procedure because it preserves both Care was taken to ensure that the miniaturized implant would be a suitable thenativeIP and MCP flexors and maintains as much pinch strength as proxy for the human implant; although the miniaturized implant did not possible in the reconstructed thumb. If either of these muscles is compro- have the metal components of the full-scale device, its shape and size mised either as a result of the initial injury or treatment history, pinch comparedtothethicknessoftheskinenvelopewerebothmoreconduciveto strength in thebiosyntheticthumb maybereduced.Wenoteherethatwe extrusion and erosion than the intended application. Note that we inten- did not directly measure the relationship between tendon force and thumb- tionally did not evaluate cutaneous innervation in the post-operative rodent tip force in the cadaver, but rather imputed this relationship from the model; because the FDMA flap is a natively-innervated rotation flap (as is measured kinematics, via the principle of virtual work. As such, any fric- the analogous flap in the rodent), post-healing sensitivity in that skin is tional or other hysteretic losses in the biosynthetic system as compared to 42,43 assumed based on decades of experience with similar flap reconstruction . the biological system would comparatively decrease the force output of the npj Biomedical Innovations | (2025) 2:22 6 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 5 | Opposition pinch with biosynthetic thumb, after coverage by FDMA and digits 2–5 (left to right). c Pinch grasp holding a small spherical object between the RRF flaps. a Radiographs showing opposition pinch between the biosynthetic biosynthetic thumb and digits 2–5 (left to right). thumb and digits 2–5. b Photos showing opposition pinch between the thumb and biosynthetic system for a given muscle input, especially during dynamic surgically-invasive prosthetic technologies that seamlessly integrate with the motions. body to provide natural touch sensation. If we expand our “menu” of This work serves as an initial preclinical validation of biosynthetic reconstructive materials to include not only titanium struts and copper thumb reconstruction, and several changes to the device are necessary prior wires, but also skin and nerves, we may be able to combine the best of the to clinical implementation. Specific design considerations not included in biological and synthetic worlds in maximizing functional restoration. the simplified prototype include polyethylene bearings to prevent metal-on- metal wear, as well as silicone covers to protect the skin from the titanium Methods linkage components. Additionally, performance of the complete device Study design should be assessed during cyclic loading under a representative range of This study was designed to experimentally demonstrate feasibility of a loading conditions. Future research should also explore the biosynthetic biosynthetic prosthesis that combines synthetic materials with biological prosthetic thumb as an index procedure, rather than as a secondary surgery; tissues to restore function and sensation after thumb amputation. As such, this could provide an immediate solution for complex thumb injury in cases the experiments explicitly address the three most critical elements of the where replantation is not possible. device’s design: surgical approach to implant placement and coverage, A single treatment that both improves mechanical function and mechanical performance and robustness to load, and in vivo viability of the restores natural cutaneous sensation haslongremainedanelusive ideal in skin-implant interface. limb and digit reconstruction. By co-engineering body and machine our approach uniquely combines the strength, machinability, and durability of Surgical approach synthetic material with the unparalleled sensing and protective abilities of All cadaver work was carried out in fresh-frozen cadavers with approval biological skin . While this initial study is focused on the thumb, due to its from the University of California’s Donated Body Program. Donors pro- critical importance to hand function, the approach could be readily adapted vided informed consent for use of the tissues, images, and data in this for reconstruction of other digits. Similar techniques are already used manuscript. Specimens were stored frozen and were thawed at room tem- selectively in phalloplasty, which involves construction of a cylindrical perature for 48 h before use. At the end of each experimental day, a saline appendage from a tubularized skin flap and often incorporates implantation solution was applied to maintain tissue moisture. Two cadaveric dissections 63–65 of a synthetic penile prosthesis . These procedures have relatively low were conducted to identify an appropriate surgical approach to implanting complication rates, providing additional confidence in the ability of pedicled the device and covering it with native biological skin (Supplementary Fig. 1). 65,66 skin flaps to heal around a synthetic implant . Expansion of this bio- Each dissection began with a simulated amputation of the IP joint according synthetic approach to larger reconstructions, such as full limb joints, would to current clinical practice, performed by the surgeon co-authors. Several be limited by actuation and flap viability ; however, these challenges could options for fixation and coverage were then explored, using 3D printed conceivably be overcome with passive joint structures or implanted actua- dummy implants of the approximate size and shape of the intended device. tors, and with the help of tissue-expanders, multi-stage flap pre-con- Each potential surgical approach was assessed qualitatively by the surgeon 67–69 ditioning, and prelamination . By bridging the gap between synthetic and co-authors for viability and estimated likelihood of success. This design biological reconstruction, this work lays the foundation for a new class of process continued iteratively and collaboratively until a suitable surgical npj Biomedical Innovations | (2025) 2:22 7 https://doi.org/10.1038/s44385-025-00031-z Article approach was identified. Critical dimensions of the thumb were also mea- and hang perpendicular to the thumb-tip segment. To replicate the normal sured during these dissections, including size of the intramedullary canal in forces encountered by the thumb tip during grasping and pinching activities, the residual phalanx, length of the relevant thumb segments, and distance the device was positioned in the test rig so that its dorsal aspect faced the between the MCP joint center and the intended mounting location on the ground, with weights suspended from a hook on the dorsal side. Weights palmar surface of the metacarpal bone. were added in ~10 N increments until a total load of 82 N was achieved, after which the weights were gradually removed until 0 N was reached. The same Mechanical performance motion capture setup was used to quantify deformation under each sub- Our evaluation of mechanical performance was intended to show similarity sequent load. The experiment was performed quasi-statically: each time a between implant kinematics and motion of the biological thumb during weightwasaddedorremoved,thesystemwasallowedtosettlefor10 sbefore opposition pinch, and ability of the implant to maintain these relationships applying the next load. This process was repeated for each of the seven fixed in the presence of large grasping forces. The prosthetic device was designed MCP positions. Marker trajectories were analyzed to determine planar as an implanted four-bar crossed linkage, in which movement of one link (in translational and rotational deformation under each load. Marker coordi- this case, the residual proximal phalanx) generates motion in the remaining nates were low-pass filtered at 15 Hz, projected onto the plane orthogonal to links. This configuration allows movement of the biological MCP joint to be the device’s axes of rotation, and then used to determine translation and transmitted to the distal, synthetic IP joint without requiring attachment of rotation of the thumb tip, as well as effective rotation of the MCP and IP 70,71 the residual tendon to the synthetic implant material . angles, under load. For this analysis, MCP angle was defined as the angle We established a target kinematic relationship between the biological between the vectors aligned with the principal axes of the metacarpal and MCP and synthetic IP angles, to mimic the relationship between these two proximal phalanx segments, while the IP angle was defined as the angle joints in the biological thumb during opposition pinch .Wenotethat between the proximal phalanx and the central axis of the thump-tip segment. published data regarding MCP-IP angle relationships during grasping tasks are relatively sparse. As such, we present this target kinematic relationship In-vivo viability not as a codified optimization target, but as a starting point to demonstrate Animal work was carried out in n = 23 skeletally mature, male Lewis rats that the linkage can be tuned to produce kinematics in the ballpark of (250–300 g) with approval from UCLA’s Animal Research Committee. Our biological joint function. The linkage was parameterized into seven para- objective in these rodent experiments was to establish viability of the pro- meters that fully describe its mechanics (four linkage lengths and one angle) posed skin-implant interface in the setting of a rotation flap in a high- and orientation with respect to the residual thumb (two angles). These motion area. Our primary measure of success was non-extrusion of a parameters were optimized using a gradient descent approach (fmincon, miniaturized, analogous implant across all animals at all experimental time MATLAB 2022, MathWorks, Natick, MA, USA) with an objective function points. We also quantified skin ingrowth into the porous implant material that minimized the root mean square deviation from the target trajectory of and fibrous encapsulation of the gross implant. Sample size was selected the linkage’s MCP-IP relationship. The optimization was constrained based based on an a priori power analysis of the outcome with the largest expected on the anatomy, as measured in the initial cadaver dissections. Specifically, variance (skin ingrowth). Animals were randomly assigned to the experi- the distance between the MCP joint center and the proximal end of the mental and control groups (Supplementary Table 1). metacarpal mounting place was constrained to be at least 4.5 cm to prevent A two-part miniature implant was designed to capture key features of disruption of the MCP joint capsule. The distance between the two screws the human-scale device, including approximate size relative to the skin that connected the device’s thumb tip was constrained to ensure the device thickness and radii of curvature. The miniature implant was 3D printed from links did not protrude beyond the envelope of the device which could porous high-density polyethylene (Anatomics, Melbourne, Australia), which increase the risk of extrusion of the device over time. Linkage lengths of the isthesamematerialweproposedasacoveringfortheportionsofthehuman- optimal “clinically viable device” were: 5.04 mm, 10.05 mm, 33.51 mm, and scale implant that should not move relative to the skin. Grossly, the miniature 34.66 mm, with an initial IP angle of 7.4°. This optimized linkage archi- implant components were designedtomimic theshape of theartificial tecture was implemented into a complete implant design in Solidworks proximal phalanx and the thumb tip. The thumb tip component had a (Dassault Systems, 2022). A prototype of the device was then manufactured sharper radius of curvature than the analogous component of the human at custom machine shop in implant-grade titanium (Ti-6Al-4V), which was implant, and the ratio of skin thickness to implant size was more aggressive in selected for its proven biocompatibility and excellent fatigue strength. the animal model, to ensure that the animal experiment posed a higher risk of Prior to mechanical testing, the prototype implant was affixed at its extrusion than the intended human application (Supplementary Fig. 3) . proximal end to a titanium hinge, representing the biological MCP joint and Twenty-three rats were anesthetized and underwent an experimental residual proximal phalanx. The relationship between the MCP angle (syn- surgical procedure. Rats were anesthetized in an induction chamber with thetic only for testing, biological in the intended use) and the IP joint 4–5% isoflurane and 1.0 L/min oxygen. They were then transferred to a nose (synthetic part of the implant) was characterized using a 3D motion capture cone for the surgical procedure, where anesthesia was maintained at 3–5% system (Vantage, Vicon, Yarnton, Oxfordshire), which were positioned to isoflurane and 0.8–1.0 L oxygen. During the surgical procedure, a racket- track the positional changes of reflective markers affixed to the device as it shaped incision measuring 2 cm × 4 cm was created in the epigastric region was passively cycled through its full range of motion. To minimize reflection near the hind limb. Each flap was elevated and the blood vessel supplying the from the metal device, exposed titanium was covered with non-reflective flap was identified. In the experimental group (11 rats), the two-piece tape. Three reflective markers were placed on each segment of the prototype, implant was placed on the inferior aspect of the flap, and the flap was along with one at the thumb tip and on each screw. For this analysis, the tubularized around the implant. In the control group (12 rats), the flap was MCP angle was defined as the angle formed between link C and a vertical tubularized closed without an implant. The flap wasthensutured closed, line, while the IP angle was definedasthe angleformedbetween theline rotated on its pedicle, and its proximal end was affixed to the epigastric connecting the midpoints of links A and B and the line extending from the region toward the midline. Care was taken to position the flap relative to the midpoint of link B to the thumb tip. legs so that it would move when the rat walked. The defect resulting from the Deformation of the thumb tip was characterized under increasing elevation of the flap was sutured closed by primary intention. The rats were 72,73 loads, comparable to those experienced by a biological thumb .A mod- monitored closely in the post-operative period and treated with NSAIDs ified version of the device was employed for load testing, which incorporated and antibiotics. Four animals (two from each group) were lost to chewing, two beams that enabled rigid locking of the synthetic MCP joint (and thus IP and were thus excluded from the analysis. One attempt was made in each joint) at a fixed angle using a dowel pin. The prototype was mounted to a test animal to reattach the flap; all such attempts were unsuccessful, and all four rig that allowed for precise orientation of the thumb tip in global space, such animals were sacrificed prematurely. In the remaining 19 animals, after the that weights could be hung from the thumb tip at different MCP-IP angles flaps healed, we validated experimentally that the epigastric flap moved npj Biomedical Innovations | (2025) 2:22 8 https://doi.org/10.1038/s44385-025-00031-z Article enough to simulate motion of the skin covering the human thumb. In each intramedullary canal of the residual proximal phalanx and an interlock animal, three marks were made along the flap’slongaxiswithaskinmarker screw (M2) was passed volar to dorsal through the intramedullary post with (Supplementary Fig. 4a). Motion of these marks was then recorded through bicortical fixation. The proximal end of the linkage mechanism was fixed to the floor of a transparent enclosure as the rats walked freely. Relative strain the metacarpal through the plate attachment with two bicortical screws between each pair of marks was then extracted (Supplementary Fig. 4b, c) (M2). Prior to coverage with skin flaps, markers were affixed to the surface of via a video analysis software (Tracker Video Analysis and Modeling Tool, the bare implant, and motion was again recorded while pulling on the FPL Open Source Physics). This strain analysis was not possible as described in tendon. Skin flaps from the forearm and index finger were then harvested two of the control rats, because the flaps had become too “stretched out” according to the proposed surgical technique, and moved to cover the from months of ambulation to support the video analysis; in these two cases, device. New markers were applied to both the biological and artificial joints, it was assumed that the skin underwent meaningful deformation during as well as to the previously identified bony landmarks. Ten trials of IP joint healing, as this would have been necessary for the flap to reach that motion were collected using the cadaver’snativeanatomy, followedby ten stretched-out state. trials with the implanted prosthetic thumb. The mean joint angle rela- At the conclusion of the study, animals were euthanized to facilitate tionships and +/−1 standard deviation are shown in Fig. 4c, d. Motion harvesting of the skin flap for gross pathology and histological analysis. For capture data were low-pass filtered at 6 Hz, and MCP and IP joint angle euthanasia, rats were placed individually in a sealed clear chamber and CO trajectories, as well as tendon excursion, were calculated directly from was delivered at a rate of 8.0–12.0 L/min until respiration ceased. CO flow marker positions. We did not measure the relationship between input was maintained for at least one additional minute after signs of life had tendon force and output thumb-tip force in the cadaver, but instead stopped. Primary euthanasia was followed by thoracotomy as a secondary imputed mechanical advantage from our kinematic measurements. physical method of euthanasia. In the control group, samples were collected from 3 subjects at 5.25 months post-op, 2 subjects at 6.5 months, 2 subjects Statistical analysis at 11.3 months, and 2 subjects at 19.4 months post-op. In the experimental Geometric optimization of the linkage was performed using gradient des- group, 3 subjects were sacrificed between 3 and 6 months post-op (ranging cent, with an objective function that minimized the root mean square from3.7to5.4 months), 3weresacrificed at 8.9 months, 1 at 10.5 months, 1 deviation from the target MCP-IP relationship. Stiffness of the prototype at 10.8 months, and 2 at 11.2 months post-op. Immediately after euthanasia, device under load was calculated as the slope of a linear regression between the flaps were thoroughly examined for signs of extrusion. The entire flap deformation and load at each load magnitude. To account for slop in the was then excised from the animal for histological analysis. One tissue sample mechanism, the regression was allowed a non-zero y-intercept; displace- from the experimental group was compromised while first developing the ment data were then baseline shifted such that the regression line passed histological processing protocol; this sample was excluded from all histo- through (0, 0). Skin ingrowth was regressed against post-op time at harvest logical analysis, leaving 9 experimental samples (Fig. 3, Supplementary Fig. using a linear fit with an unconstrained y-intercept. Skin thickness was 5) Tissue samples were also collected from the contralateral epigastric region compared between the flap and contralateral side using a paired t test at a for use as a baseline for skin thickness. All samples were fixed with formalin significance level of α=0.05. In the cadaver study, similarity between the for two days and then dehydrated using ethanol (70%). Each sample was average trajectories of the biological and biosynthetic systems was calculated then embedded in paraffin wax and stained with hematoxylin and eosin using a Pearson correlation analysis. (Scientific Solutions, Fridley, MN, USA). The slides were digitized using a Keyence-BZ X800 microscope, and analyzed qualitatively for signs of Data availability extrusion (Supplementary Fig. 5). We then used a custom software package All data reported in the paper are included in the manuscript or available in to segment the histology images into empty space, tissue-filled space, and the Supplementary Materials. implant area (Supplementary Fig. 6). All software-based segmentation was checked for accuracy by a human observer. Percent in-growth was calcu- Received: 26 March 2025; Accepted: 17 June 2025; lated as the percentage of non-implant space filled by tissue. Skin thickness was also measured manually from the digital slide by three different human observers, skilled in microscopy, whose measurements were then averaged References to provide the final thickness measurement for each specimen. 1. Bell, C. The hand: its mechanism and vital endowments, as evincing design (Harper & Brothers, 1864). Full system validation in cadaver 2. Doege, T. C. & Houston, T. P. Guides to the evaluation of permanent To prevent movement during testing, the proximal ends of the cadaveric impairment (American Medical Association, 1993). radius and ulna were screwed to a wooden board, which was then clamped 3. Verdan, C. The reconstruction of the thumb. Surg. Clin. North Am. 48, to the surface of the dissection table. The cadaver specimen was placed 1033–1061 (1968). within the capture volume of the same motion capture system used for the 4. Slocum, D. B. & Pratt, D. R. Disability evaluation for the hand. J. Bone mechanical testing experiments. The flexor pollicis longus (FPL) tendon of Joint Surg. Am. 28, 491–495 (1946). the biological thumb was identified andtaggedwitha suture; thistendon 5. Li, Z.-M. & Tang, J. Coordination of thumb joints during opposition. J. attaches to the base of the distal phalanx and facilitates flexion of the IP and, Biomech. 40, 502–510 (2007). secondarily, the MCP joints. Reflective markers were placed on bony 6. Letzelter, J. P. et al. Hand function following simulated fusion of thumb landmarks corresponding to the IP, MCP, and CMC joint centers, as well as metacarpophalangeal and interphalangeal joints. HAND 17,74–78 on the volar aspect of the distal phalanx, proximal phalanx, and metacarpal (2022). of the thumb. 7. Moran, S. L. & Berger, R. A. Biomechanics and hand trauma: what you We first established a baseline for motion of the MCP and IP joints in need. Hand Clin 19,17–31 (2003). the biological thumb. Specifically, we tracked motion of the biological 8. Germann, G., Rudolf, K. D., Levin, S. L. & Hrabowski, M. Fingertip and proximal and distal phalanx while pulling the FPL tendon along its natural thumb tip wounds: changing algorithms for sensation, aesthetics, and direction of action, inducing flexion at the IP, MCP, and CMC joints. We function. J. Hand Surg. 42, 274–284 (2017). then amputated the left thumb proximal to the IP joint. During the 9. Crowe, C. S. et al. Global trends of hand and wrist trauma: a amputation, the FPL tendon was reattached to the base of the residual systematic analysis of fracture and digit amputation using the Global proximal phalanx according to standard clinical procedures; this maintains Burden of Disease 2017 Study. Inj. Prev. 26, i115–i124 (2020). the FPL’s contribution to flexion of the MCP joint and, through the linkage, 10. Chung, K. C. et al. Patient-reported and functional outcomes after of the reconstructed IP joint. The post of the implant was placed into the revision amputation and replantation of digit amputations: the npj Biomedical Innovations | (2025) 2:22 9 https://doi.org/10.1038/s44385-025-00031-z Article FRANCHISE Multicenter International Retrospective Cohort Study. development of metal soft tissue materials. Adv. Funct. Mater. 30, JAMA Surg. 154, 637 (2019). 1910672 (2020). 11. Viteri, V. S. de & Fuentes, E. Titanium and titanium alloys as 36. Huang, P. et al. Improving hard metal implant and soft tissue biomaterials. In Tribology - fundamentals and advancements integration by modulating the “inflammatory-fibrous complex” (IntechOpen, 2013). https://doi.org/10.5772/55860. response. Bioact. Mater. 20,42–52 (2022). 12. Anderson, I. A., Gisby, T. A., McKay, T. G., O’Brien, B. M. & Calius, E. P. 37. Hanawa, T. Titanium–tissue interface reaction and its control with Multi-functional dielectric elastomer artificial muscles for soft and surface treatment. Front. Bioeng. Biotechnol. 7, 170 (2019). smart machines. J. Appl. Phys. 112, 041101 (2012). 38. Puttock, D. R., Howard, D. P., Eastley, N. C. & Ashford, R. U. Apparent 13. Burden, S. A., Libby, T., Jayaram, K., Sponberg, S. & Donelan, J. M. trends in the use of femoral megaprostheses: an analysis from the Why animals can outrun robots. Sci. Robot. 9, eadi9754 (2024). National Joint Registry. Arthroplasty 4, 50 (2022). 14. Ates, H. C. et al. End-to-end design of wearable sensors. Nat. Rev. 39. Nongdamba, H. et al. Functional audit of the use of megaprosthesis Mater. 7, 887–907 (2022). for limb reconstruction in musculoskeletal tumors – A retrospective 15. Hammock, M. L., Chortos, A., Tee, B. C.-K., Tok, J. B.-H. & Bao, Z. single-center study. J. Orthop. 49, 123–127 (2023). 25th Anniversary article: the evolution of electronic skin (E-Skin): a 40. Hall, P. T. et al. Fully implanted prostheses for musculoskeletal limb brief history, design considerations, and recent progress. Adv. Mater. reconstruction after amputation: an in vivo feasibility study. Ann. 25, 5997–6038 (2013). Biomed. Eng. 49, 1012–1021 (2021). 16. Dixit, S. et al. Immunological challenges associated with artificial skin 41. Mandrekas, A. D. & Zambacos, G. J. Reverse-flow radial forearm flap grafts: available solutions and stem cells in future design of synthetic for reconstruction of the hand. Ann. Plast. Surg. 33, 536 (1994). skin. J. Biol. Eng. 11, 49 (2017). 42. Eski, M., Nisanci, M. & Sengezer, M. Correction of thumb deformities 17. Kim, G. M., Powell, J. E., Lacey, S. A., Butkus, J. A. & Smith, D. G. after burn: versatility of first dorsal metacarpal artery flap. Burns 33, Current and emerging prostheses for partial hand amputation: a 65–71 (2007). narrative review. PMR 15, 392–401 (2023). 43. Couceiro, J., de Prado, M., Menendez, G. & Manteiga, Z. The first 18. Cordella, F. et al. Literature review on needs of upper limb prosthesis dorsal metacarpal artery flap family: a review. Surg. J. 4, e215–e219 users. Front. Neurosci. 10, 209 (2016). (2018). 19. Marks, L. J. & Michael, J. W. Science, medicine, and the future: 44. Ratcliffe, R. J., Regan, P. J. & Scerri, G. V. First dorsal metacarpal Artificial limbs. BMJ 323, 732–735 (2001). artery flap cover for extensive pulp defects in the normal length thumb. 20. GripLock Finger. Naked prosthetics https://www.npdevices.com/ Br. J. Plast. Surg. 45, 544–546 (1992). devices/griplock-finger/ (2025). 45. Chan, P. Y. W., Colon, A. F., Clune, J. & Shah, A. External tissue 21. Miller, T. J. Digit amputation prosthetics. Plast. Aesthetic Res. 10,N/ expansion in complex extremity reconstruction. J. Hand Surg. 46, A–N/A (2023). 1094–1103 (2021). 22. Brånemark, R. et al. A novel osseointegrated percutaneous prosthetic 46. Baino, F. & Potestio, I. Orbital implants: State-of-the-art review with system for the treatment of patients with transfemoral amputation: a emphasis on biomaterials and recent advances. Mater. Sci. Eng. C 69, prospective study of 51 patients. Bone Jt. J. 96-B, 106–113 (2014). 1410–1428 (2016). 23. Lin, P.-Y. et al. A systematic review of outcomes of toe-to-thumb 47. Hernandez Rosa, J., Villanueva, N. L., Sanati-Mehrizy, P., Factor, S. H. transfers for isolated traumatic thumb amputation. Hand 6, 235–243 & Taub, P. J. Review of maxillofacial hardware complications and (2011). indications for salvage. Craniomaxillofacial Trauma Reconstr. 9, 24. Difonzo, E., Zappatore, G., Mantriota, G. & Reina, G. Advances in 134–140 (2016). finger and partial hand prosthetic mechanisms. Robotics 9, 80 (2020). 48. Knott, P. D. et al. Evaluation of hardware-related complications in 25. Maheshwari, V. & Saraf, R. F. High-resolution thin-film device to sense vascularized bone grafts with locking mandibular reconstruction plate texture by touch. Science 312, 1501–1504 (2006). fixation. Arch. Otolaryngol. Head Neck Surg. 133, 1302–1306 (2007). 26. Augustine, R. Skin bioprinting: a novel approach for creating artificial 49. Friedman, R. J. Silicone breast prostheses implantation and skin from synthetic and natural building blocks. Prog. Biomater. 7, explantation. Semin. Arthritis Rheum. 24,8–10 (1994). 77–92 (2018). 50. Paxton, N. C. et al. Additive manufacturing enables personalised 27. Raspopovic, S., Valle, G. & Petrini, F. M. Sensory feedback for limb porous high-density polyethylene surgical implant manufacturing prostheses in amputees. Nat. Mater. 20, 925–939 (2021). with improved tissue and vascular ingrowth. Appl. Mater. Today 22, 28. Tan, D. W. et al. A neural interface provides long-term stable natural 100965 (2021). touch perception. Sci. Transl. Med. 6, 257ra138 (2014). 51. Paxton,N. C., Allenby, M. C.,Lewis,P.M.&Woodruff, M. A. Biomedical 29. Raspopovic, S. et al. Restoring natural sensory feedback in real-time applications of polyethylene. Eur. Polym. J. 118,412–428 (2019). bidirectional hand prostheses. Sci. Transl. Med. 6, 222ra19–222ra19 52. De Voogt, W. G. Pacemaker leads: performance and progress. Am. J. (2014). Cardiol. 83, 187–191 (1999). 30. Graczyk, E. L. et al. The neural basis of perceived intensity in natural 53. Bregoli, C. et al. Osseointegrated metallic implants for finger amputees: a and artificial touch. Sci. Transl. Med. 8, 362ra142–362ra142 (2016). review of the literature. Orthop. Surg. 14,1019–1033 (2022). 31. Clites, T. R. et al. Proprioception from a neurally controlled lower- 54. Li, Y., Kulbacka-Ortiz, K., Caine-Winterberger, K. & Brånemark, R. extremity prosthesis. Sci. Transl. Med. 10, eaap8373 (2018). Thumb amputations treated with osseointegrated percutaneous 32. Khan, A., Li, K., Nataraj, R. & Wei, N. Effects of transcutaneous prostheses with up to 25 years of follow-up. JAAOS Glob. Res. Rev. 3, electrical nerve stimulation on evoked tactile sensation for fingertip e097 (2019). force control. Biomed. Signal Process. Control 87, 105568 (2024). 55. Yu, W. S., Van Duinen, H. & Gandevia, S. C. Limits to the control of the 33. Greenspon, C. M. et al. Evoking stable and precise tactile sensations human thumb and fingers in flexion and extension. J. Neurophysiol. via multi-electrode intracortical microstimulation of the 103, 278–289 (2010). somatosensory cortex. Nat. Biomed. Eng.1–17, https://doi.org/10. 56. King, E. A. & Ozer, K. Free skin flap coverage of the upper extremity. 1038/s41551-024-01299-z (2024). Hand Clin. 30, 201–209 (2014). 34. Zhao, Z.-P. et al. Modulating brain activity with invasive 57. Hansen, A. J. et al. Reverse radial forearm fascial flap with radial artery brain–computer interface: a narrative review. Brain Sci. 13, 134 (2023). preservation. Hand 2, 159–163 (2007). 35. Liu, R. et al. Immunomodulation-based strategy for improving soft 58. Kim, J., Yoon, A. P. & Jones, N. F. Reverse radial forearm flap to tissue and metal implant integration and its implications in the provide arterial inflow to a toe transfer. Hand 12, 154–161 (2017). npj Biomedical Innovations | (2025) 2:22 10 https://doi.org/10.1038/s44385-025-00031-z Article 59. Scheker, L. R., Polo, F. S. & Aguilar, F. J. Free tissue transfer for fingertip Acknowledgements coverage. In Fingertip injuries: diagnosis, management and We extend our heartfelt gratitude to Dr. Joanne Zahorsky-Reeves and the reconstruction (ed. Rozmaryn,L.M.) 131–148 (Springer International UCLA Division of Laboratory Animal Medicine (DLAM) team for their support Publishing, Cham, 2015). https://doi.org/10.1007/978-3-319-13227-3_8. and guidance during the animal studies. We would like to thank Dr. Karen 60. Easton, K. L. et al. Replacement of tibialis cranialis tendon with Lyon’s lab at UCLA for their assistance in the initial stages of histological polyester, silicone-coated artificial tendon preserves biomechanical processing. We extend a special thanks to Alyssa Tomkinson for her function in rabbits compared to tendon excision only. J. Orthop. Surg. exceptional design of the illustrations featured in the figures of this paper. 19, 108 (2024). Finally, wewishto thank individualswho donated their bodies and tissues for 61. Collins, J., Ishihara, Y. & Thoma, A. Management of digital tendon the advancement of education and research. A portion of this work was avulsion at the musculotendinous junction of the forearm: a supported by the NIH Director’s New Innovator Award (DP2, grant no. systematic review. Hand 7, 134–142 (2012). 1DP2HD111538). 62. Clites, T. R. Anatomics: co-engineering body and machine in pursuit of synergistic bionic performance. Curr. Opin. Biomed. Eng. 28, Author contributions 100490 (2023). Conceptualization: T.R.C., L.E.W., K.K.A. Methodology: T.R.C., L.E.W., 63. Yao, A. et al. Total penile reconstruction: a systematic review. J. Plast. K.K.A., G.V.L., M.B., S.B. Investigation: S.B., L.E.W., K.K.A., A.Q.K., J.K., Reconstr. Aesthet. Surg. 71, 788–806 (2018). M.B., N.S.J., G.V.L., B.K.Z. Data Processing: S.B., K.B., A.Q.K., S.H., B.K.Z., 64. Matti, B. A., Matthews, R. N. & Davies, D. M. Phalloplasty using the S.T., T.R.C. Visualization: S.B., A.T., S.T., T.R.C. Funding acquisition: T.R.C. free radial forearm flap. Br. J. Plast. Surg. 41, 160–164 (1988). Writing – original draft: S.B., T.R.C. Writing – review and editing: S.B., M.B., 65. Frey, J. D., Poudrier, G., Chiodo, M. V. & Hazen, A. A Systematic A.Q.K., J.K., A.T., S.H., K.B., B.K.Z., N.S.J., K.K.A., L.E.W., T.R.C., S.T. All review of metoidioplasty and radial forearm flap phalloplasty in authors approved of the final version of the paper. female-to-male transgender genital reconstruction: is the “ideal” neophallus an achievable goal?. Plast. Reconstr. Surg. Glob. Open 4, Competing interests e1131 (2016). T.R.C., L.E.W., K.K.A. and G.V.L. are named inventors on a pending patent 66. Lumen, N., Monstrey, S., Ceulemans, P., van Laecke, E. & Hoebeke, (PCT/US23/76502) describing the biosynthetic thumb concept. UCLA is the P. Reconstructive surgery for severe penile inadequacy: phalloplasty patent applicant. with a free radial forearm flap or a pedicled anterolateral thigh flap. Adv. Urol. 2008, 704343 (2008). Additional information 67. Manise, O., Watier, E., Rostane-Renouard, D., Staerman, H. & Supplementary information The online version contains Pailheret, J. P. Use of tissue expansion on defective amputation supplementary material available at stumps of the lower limbs. Apropos of 5 cases. Ann. Chir. Plast. https://doi.org/10.1038/s44385-025-00031-z. Esthet. 41, 155–160 (1996). 68. Kim, J.-Y., Choi, T. H. & Kim, B. J. Refinements in the preexpanded Correspondence and requests for materials should be addressed to distant flap for giant melanocytic nevi of the upper extremity in Tyler R. Clites. pediatric patients. Plast. Reconstr. Surg. 154, 375 (2024). 69. Fermi, M. et al. Prelaminated flaps in head and neck cancer Reprints and permissions information is available at reconstructive surgery: a systematic review. Microsurgery 41, http://www.nature.com/reprints 584–593 (2021). 70. Bundhoo, V., Haslam, E., Birch, B. & Park, E. J. A shape memory alloy- Publisher’s note Springer Nature remains neutral with regard to based tendon-driven actuation system for biomimetic artificial jurisdictional claims in published maps and institutional affiliations. fingers, part I: design and evaluation. Robotica 27, 131–146 (2009). 71. Choi, K. Y., Akhtar, A. & Bretl, T. A compliant four-bar linkage Open Access This article is licensed under a Creative Commons mechanism that makes the fingers of a prosthetic hand more impact Attribution 4.0 International License, which permits use, sharing, resistant. In 2017 IEEE International Conference on Robotics and adaptation, distribution and reproduction in any medium or format, as long Automation (ICRA) 6694–6699 https://doi.org/10.1109/ICRA.2017. as you give appropriate credit to the original author(s) and the source, 7989791 (2017). provide a link to the Creative Commons licence, and indicate if changes 72. Dzidek, B. M., Adams, M. J., Andrews, J. W., Zhang, Z. & Johnson, S. were made. The images or other third party material in this article are A. Contact mechanics of the human finger pad under compressive included in the article’s Creative Commons licence, unless indicated loads. J. R. Soc. Interface 14, 20160935 (2017). otherwise in a credit line to the material. If material is not included in the 73. Liu, S., Van, M., Chen, Z., Angeles, J. & Chen, C. A novel prosthetic article’s Creative Commons licence and your intended use is not permitted finger design with high load-carrying capacity. Mech. Mach. Theory by statutory regulation or exceeds the permitted use, you will need to 156, 104121 (2021). obtain permission directly from the copyright holder. To view a copy of this 74. Hwang, K., Kim, H. & Kim, D. J. Thickness of skin and subcutaneous licence, visit http://creativecommons.org/licenses/by/4.0/. tissue of the free flap donor sites: a histologic study. Microsurgery 36, 54–58 (2016). © The Author(s) 2025 npj Biomedical Innovations | (2025) 2:22 11 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png npj Biomedical Innovations Springer Journals

A biosynthetic thumb prosthesis

Download PDF
11 pages

Loading next page...
 
/lp/springer-journals/a-biosynthetic-thumb-prosthesis-bvsGjTbBsS

References (74)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2025
eISSN
3005-1444
DOI
10.1038/s44385-025-00031-z
Publisher site
See Article on Publisher Site

Abstract

npj | biomedical innovations Article https://doi.org/10.1038/s44385-025-00031-z Check for updates 1 1 2 2 2 2 Sachi Bansal ,GraciaV. Lai , Marcus Belingheri , Amber Q. Kashay ,JinyoungKim ,AlyssaTomkinson , 3 4 4 2 5 Samantha Herman , Keval Bollavaram , Brian K. Zukotynski , Sean Thomas , Nirbhay S. Jain , 4,5 4 1,2,4 Kodi K. Azari , Lauren E. Wessel & Tyler R. Clites In cases of severe damage to the extremities, the function and structure of compromised tissues must be replaced. If biological reconstruction using autologous tissue is not feasible, amputation and replacement with a synthetic prosthesis is often the next best option. Synthetic prostheses are limited, especially in their ability to restore skin sensation. Here we show a biosynthetic prosthesis that combines the versatility of titanium with the rich sensory capabilities of biological skin. The prosthesis recreates opposition pinch by linking motion of the prosthetic joint to that of the residual biological joint, and is enclosed in neurotized skin from the patient’s own body. We validated the biosynthetic thumb’s mechanical function on the benchtop and in a cadaver, and showed viability of the skin interface in an animal model. These results provide a framework for functional reconstruction of amputated digits using a combination of synthetic materials and biological tissues. The thumb is considered clinically to be the most important digit on the For some of these functions (e.g. actuation, mechanical strength), synthetic 1–4 hand, responsible for approximately 40% of hand function .The thumb’s components (motors, titanium) outperform the biological systems (mus- 5 11–13 core role is opposition, which is critical for grasp and pinch .Natural cles, bone) that they are intended to replace . However, other functions, opposition requires an ability to trace the tip of the thumb along the tips of such the skin’s ability to sense the environment and protect from infection, the remaining four fingers. In the intact hand, this ability is a direct result of have proven particularly challenging to replicate with synthetic 5 14–16 thumb’s position on the hand and complex anatomical structure .The components . As a result, it has been difficult to create synthetic digits 17–19 interphalangeal (IP) joint, which is the thumb’smost distal joint,plays a that look,feel,and behave like their biological counterparts . crucial role in functional opposition pinch, accounting for up to 31% of With no good option for autologous reconstruction, thumb prostheses 6,7 pinch strength and ensuring fingertip-to-thumb-tip contact .The rich face the challenges of restoring functional opposition pinch, attaching to the sensory information from the skin at the thumb tip allows for robust closed- hand, and providing natural touch sensation. Conventional external pros- loop control of pinch and grasp force, as well as perception of texture, slip, theses, such as the GripLock Finger (Naked Prosthetics, Olympia, WA, and temperature . The skin also acts as a barrier to external insult such as USA), are effective in restoring opposition pinch but cannot provide cuta- 20,21 infection. neous feedback to the nervous system . This type of prosthesis typically Catastrophic injury to the thumb is fairly common: approximately 6 straps to the hand, which does not provide a rigid attachment and adds bulk million digital amputations occur annually worldwide, one third of which that can obstruct hand function. Percutaneous osseointegrated implants include loss of at least one thumb joint . Because the thumb plays such a provide a low-profile option for rigidly attaching prosthetic thumbs to the critical role in hand function, its loss can have tangible effects on lifestyle . residual bone; however, osseointegrated thumb prostheses do not have a This is especially true in populations who rely on their hands for manual functional IP joint, and therefore cannot restore opposition pinch. Without labor, whichincidentally predisposes them to these very amputations; loss of neural integration these devices also cannot restore skin sensation, and the the thumb can inhibit their ability to perform work functions, creating percutaneous abutment creates a chronic infection risk .Even the most financial strain or necessitating a shift in career. More broadly, partial hand advanced of current clinical devices offer only partial restoration of function, amputation is linked directly to decreases in mental health and overall and none have successfully replicated the critical tactile feedback and pro- quality of life, with a large portion of the deficit related to functional tective properties of native skin. In fact, the shortcomings of existing impairment . While biological reconstruction using native tissues is typi- prosthetic technologies are significant enough that some patients choose to cally favored, it is not always possible. In such cases, we face the task of sacrifice another of their digits to restore thumb function (e.g. toe-to-thumb reproducing the function of the lost biological digit via synthetic prostheses. transfer). While this purely biological approach restores both opposition 1 2 Department of Bioengineering, University of California Los Angeles, Los Angeles, CA, USA. Department of Mechanical and Aerospace Engineering, University of 3 4 California Los Angeles, Los Angeles, CA, USA. Department of Physics and Astronomy, University of California Los Angeles, Los Angeles, CA, USA. Department of Orthopaedic Surgery, University of California Los Angeles David Geffen School of Medicine, Los Angeles, CA, USA. Department of Plastic Surgery, University of California Los Angeles David Geffen School of Medicine, Los Angeles, CA, USA. e-mail: [email protected] npj Biomedical Innovations | (2025) 2:22 1 1234567890():,; 1234567890():,; https://doi.org/10.1038/s44385-025-00031-z Article pinch and sensation, toe-to-thumb transfer is a technically demanding implanted component to be entirely covered in living skin. Complete cov- surgical procedure, provides a limited range of motion compared to the erage of the implant (in contiguity with the skin of the residual thumb) native thumb, and can lead to changes in gait due to donor site morbidity . ensures a sealed skin envelope, which is crucial to keeping the transposed Other patients forego prosthetic reconstruction altogether and instead adapt skin alive and preventing infection. This surgical context became the starting 24 1–4 to life without a thumb , which comes at significant functional cost . point for our design process; we deliberately chose to design the implant Synthetic “skin” and direct neural interfacing have shown promise around the surgery (and not vice versa), to increase the likelihood that the in experimental devices as a means of restoring cutaneous sensation, but prosthesis would be clinically viable and work synergistically with the body still cannot compete with the density, precision, or versatility of sensory to restore function. Through a series of cadaver dissections, we developed 15,25,26 endorgansinbiologicalskin . In addition to sensing, commu- and demonstrated a new surgical approach that combines established flap nication of high-density information from external sensor arrays to the techniques in a new way such that they can be used to cover our fully- nervous system remains a persistent challenge .Direct nerve stimu- synthetic implant core (Fig. 1a). lation has been useful in restoring touch sensation from a single sensor The intramedullary canal of the residual proximal phalanx was first that localizes to a region (e.g. the whole thumb-tip), and some pre- identified as the primary site for bony anchoring, with the driving link liminary evidence has been provided that such stimulation can evoke secured to the volar aspect of the metacarpal using two biocortical 28–30 natural touch perception . While this approach has been shown to screws (M2) through the metacarpal plate. We selected a combination improve task performance, direct nerve stimulation faces drawbacks of two cutaneous flaps to cover the device with innervated, vascularized including low resolution and precision, uncomfortable or unnatural autologous skin. The first of these is the RRF flap, which provides 31,32 sensation, and sensory adaptation . Current nerve stimulation surface area to surround the implant. A RRF flap measuring 8 cm by approaches also require either implanted electronics or percutaneous 4 cm was harvested from the forearm ipsilateral to the thumb ampu- transmission of power, which can pose challenges for long-term via- tation. The proximal end of the radial artery within the flap was clipped bility. Cutaneous sensory information has also been delivered directly and ligated; in a living patient, this would induce retrograde flow in the to the brain via intracortical microstimulation; however, this approach radial artery, allowing the flap to be perfused from its distal end without has faced challenges with stability and longevity and requires invasive compromising flow to the hand . The RRF flap on its vascular leash was 33,34 surgery that is difficult to justify for persons with thumb amputation . then passed through a subcutaneous tunnel in the wrist, tubularized As an alternative to purely-synthetic reconstruction, we propose an around the thumb implant, and sutured to the distal edge of the native approach in which the prosthetic thumb is reimagined as a hybrid system skin surrounding the residual thumb. The portion of this RRF flap over that selectively leverages both biological and synthetic subsystems where the thumb tip was then replaced with an innervated 1.5 cm × 2 cm first- each is most advantageous. Specifically, we describe a biosynthetic pros- dorsal metatarsal flap (FDMA flap) from the dorsal aspect of the index thesis that combines a titanium linkage with the patient’s own skin to both finger; this flap is used to provide coverage and restore sensation to large 42,43 enable opposition pinch and restore skin sensation. At the core of the device wounds on the tip of thumb . A complete depiction of the surgical is a four-bar linkage, which couples motion of the prosthetic IP joint to that technique in the context of the full thumb prototype is shown in Sup- of the biological metacarpophalangeal (MCP) joint; this design provides plementary Fig. 1. opposition pinch without the complexity of powered actuators and does not In these dissections, both flaps were successfully harvested, rotated on require attachment of tendon to synthetic material, which remains a sig- their respective vascular pedicles, and reattached to the residual skin at the 35–37 nificant unsolved problem . Although fully-implanted articulating base of the thumb. The skin flaps were large enough to cover a proxy implant 38,39 prostheses are common in orthopedics —and have even been proposed without excessive tension in the transposed skin. The defects created on the as a solution for leg amputation —their movement depends on attachment forearm and wrist were sufficiently small to be sutured closed by primary of tendon to metal or other synthetic material. In our system, the metal intention (Fig. 1a); the defect on the index finger from the FDMA flap is portion of the biosynthetic thumb anchors only to bone, providing robust usually closed via a split-thickness graft from the forearm . Because these long-term attachment and actuation. The biological portion of the pros- flaps are not typically used to cover synthetic implants, extrusion through 45–49 thesis comprises the patient’s own vascularized, innervated skin, sustainably the reconstructed skin was a significant concern . We addressed this by harvested from elsewhere on the arm and hand. Current endoprosthetic covering the non-joint portions of the device, which do not move relative to 38,39 devices are designed to be buried beneath a robust soft tissue envelope , the surrounding skin, in porous high-density polyethylene (HDPE) to 50,51 and therefore cannot be applied to treat pre-existing or traumatic promote skin ingrowth during the healing process . In order to prevent amputations . In contrast, our biosynthetic thumb incorporates transposed soft tissue impingement or infiltration that could disrupt link movement autologous skin onto the prosthesis itself, eliminating the need for existing and joint motion, we determined that future iterations will include flexible native skin at the site of reconstruction. This transposed skin becomes the silicone covers over all articulations (Fig. 1b). Silicone was selected for its outer surface of the prosthesis, protecting the body from infection and flexibility and history of use in orthopedic applications (e.g. Swanson leveraging innate cutaneous end organs to provide natural touch and Flexible Finger JointImplant,Stryker,Kalamazoo,MI, USA),and in temperature sensation. pacemakers to protect the device from the surrounding biological This manuscript describes comprehensive preclinical validation of the environment. biosynthetic thumb, addressing surgical, mechanical, and clinical feasibility. We first use a cadaver to establish and validate a new surgical paradigm for Optimization and validation of mechanical linkage transposition of the patient’s skin onto the surface of the device. We then The synthetic portion of the implant is a crossed four-bar linkage that present an optimization and evaluation of the device’s mechanics in free couples motion of the prosthetic IP joint to that of the biological MCP joint. space and in the context of opposition-relevant loading. We describe an Importantly, this motion is enforced through bone-anchored metal com- animal study of the proposed synthetic architecture in living skin, through ponents, which avoids the need for tendon-metal interfaces; such soft-tissue 35–37 which we establish long-term clinical viability of the skin-implant interface. interfaces are notoriously prone to failure in the body .This linkage was The manuscript concludes with a surgical and kinematic evaluation of the optimized to match a target trajectory derived from a published relationship complete implant in a human cadaver model. between MCP and IP motion in the biological thumb during opposition pinch , while accounting for geometric constraints of the anatomical Results envelope. An initial optimization yielded an “idealized” linkage design, with Surgical approach to covering the prosthesis with living skin a mean absolute percent error (MAPE) of 0.73% relative to the target angle To create a biosynthetic prosthesis with real translational potential, it was relationship (Fig. 2). However, the two links that formed the IP joint first necessary to develop a surgical approach that would allow the extended beyond the device envelope, posing a high extrusion risk. As such, npj Biomedical Innovations | (2025) 2:22 2 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 1 | Surgical approach and device design. a (left to right) Biological thumb with closed, and the system is allowed to heal. b The implant is made up of four parts: the intact native joints and soft tissue. The IP joint is amputated, and native skin covers thumb tip, structural link, driving link, and metacarpal plate. The linkage connects the residual proximal phalanx. The device is attached to the local bony structures the MCP and IP joints, such that flexion of the biological MCP joint creates flexion of including the residual proximal phalanx and metacarpal. The reverse radial forearm the synthetic IP joint. Static (non-joint) portions of the implant are covered in (RRF) and first dorsal metacarpal artery (FDMA) flaps are harvested and rotated porous HDPE, and the synthetic IP joint is protected by a flexible cover. A prototype about their pedicles to cover the implant. The donor and recipient sites are sutured was manufactured in Ti-6Al-4v. we constrained the links to remain within the diameter of the “structural Validation of skin-implant interface in an animal model link” (Fig. 1b) and re-optimized. This constrained optimization produced Although the proposed surgical techniques are part of established clinical the “clinically viable” linkage design, which had an MAPE of 15.2% from the practice, these flaps are not typically used to cover synthetic implants in an ideal angle relationship. Specifically, the slope of the MCP-IP relationship area of high motion. This introduces concerns of flap viability and potential was shallower, resulting in an IP joint that was 3.4° less flexed at the extrusion. To address these concerns, a rodent model was used to validate minimum MCP joint angle, and 4.4° more flexed at the maximum MCP long-term viability of a proxy implant in the context of a large tubularized angle (Fig. 2b, blue line). rotation flap. In a total of 23 rats, vascularized skin flaps were elevated, A prototype of the clinically viable linkage design was manufactured in tubularized, and rotated to a new orientation in a high-motion area near the Ti-6Al-4V, and attached at its distal end to a simple mechanical hinge hind leg (Fig. 3a). In 12 rats, the flap was wrapped around a proxy implant (proxy MCP joint). This prototype was then affixed to an adjustable loading (experimental group); the remaining 11 did not have implants (control frame, and loads were applied to a hook on the dorsal aspect of the thumb tip group). Proxy implants were two-part components shaped analogously to (Fig. 2a). In the absence of external loads, the proxy MCP joint and the the thumb implant and made of the same HDPE material (Supplementary device’sIPjoint were able to flex over a range of 70.3° and 29.7° from their Fig. 3). Two rats from each group were lost due to flap detachment caused by rest positions, respectively (Fig. 2b, green dots), and matched the predicted chewing. In the remaining 19 animals, flaps were allowed to heal for kinematic relationship with 2.84% MAPE. 3.5–11.25 months. After the sutures in each animal had dissolved, we 53–55 Thumb-tip loads during a typical grasp activity can exceed 80 N . assessed strain in the flapsasthe rodentswalkedfreelyacrossatransparent With the linkage locked at seven different proxy MCP angles (Fig. 2c), cage (Supplementary Movie 1); these trials served to ensure that the flaps ranging from −9.0° (slight extension) to 37.1° (flexion), we observed a deformed each time the animal walked, in a way that adequately represented maximum of 3.4 mm of deflection at the thumb tip under 82.4 N of force the large amount of skin motion in the human thumb. All flaps exhibited (Supplemental Fig. 2a). Angular stiffness about the base of the implant large strains, with averages ranging from 69% to −17% (Supplementary changed as a function of proxy MCP angle, with a maximum effective Fig. 4). deflection of 9.4° under 82 N of force perpendicular to the thumb tip, cor- Our primary measure of success in these experiments was lack of responding to a moment of 3.1 Nm about the proxy MCP (Supplementary extrusion; according to this metric, we observed a 100% success rate in the Fig. 2b). Stiffness of the device varied as a function of proxy MCP angle and animals with implants. All flapsinboth groupshealed well, andshowed no was highest with the proxy MCP and IP near the middle of their ranges of gross signs of necrosis or tissue damage (Fig. 3b). Qualitative histological motion (Fig. 2d). Deflection of the thump tip decreased the effective IP angle analysis in the experimental group revealed an implant surrounded on all by up to 5.87° under 82 N, shifting the MCP-IP relationship toward less IP sides by healthy dermal tissue, with some pockets of subcutaneous fat flexion (Fig. 2e). (Fig. 3c). By volume, tissue filled 94.3% ±3.85% of the implant voids. We did npj Biomedical Innovations | (2025) 2:22 3 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 2 | Benchtop mechanical testing results. a Experimental setup, in which the d Stiffness of the device as a function of MCP angle. Stiffness represents the amount prototype thumb is clamped into an adjustable load frame and loaded with the proxy of perpendicular force required at the thumb tip to produce 1 mm of linear deflection MCP joint locked at different angles. b MCP-IP angle relationships showing the of the thumb tip. e Experimental MCP-IP angle relationship in the presence of target trajectory, the idealized linkage design, the clinically viable linkage design, and different loads. Dots show average measured angle over 1.5 s, and solid lines show the experimental performance of the physical prototype (Supplementary Table 1). first-order linear regression fits. The dashed line represents model predictions for the c Deflection of the thumb-tip marker across seven locked proxy MCP angles. clinically viable device (identical to the solid blue line from (b)). not observe any relationship (R = 0.02) between ingrowth and post-op time flexion movement at both the biological MCP and synthetic IP joints prior to tissue harvesting (Fig. 3d). We found no significant difference in (Fig. 4b, Supplementary Movies 2 and 3) that was sufficient for functional skin thickness (p = 0.825, Fig. 3d, right) between the experimental flaps and opposition pinch (Fig. 5). We opted to characterize the transmission ratio of skin samples collected from the contralateral side of the epigastric region both the biological and biosynthetic systems according to kinematics (input (Fig. 3c, bottom), suggesting that neither the presence of the implant nor tendon excursion to output thumb-tip displacement) rather than kinetics tubularization of the skin flap caused thinning of the surrounding flap tissue (input tendon force to output thumb-tip force). Assuming that neither (Fig. 3d, bottom). We also evaluated tissue encapsulation, which we defined system exhibits meaningful frictional losses and that the principle of virtual as the proportion of the implant’s circumference covered by thick fibrous work therefore holds, any scaling of the kinematic relationship between connective tissue; by this metric we observed 100% encapsulation in all tendon excursion and thumb-tip motion across systems would imply an implants (Supplementary Figs. 5, 6). inverse scaling of the force relationships. When actuated by manually excursing the FPL tendon, the IP joint in Performance of the complete biosynthetic thumb in a cadaver the biosynthetic thumb moved 1.08° per 1° of MCP joint flexion (Fig. 4c, We used a cadaver to characterize opposition pinch after attachment of the green), compared to 1.12° per 1° of MCP flexioninthe cadaver’sbiological implant to the residual proximal phalanx. In the cadaver (Fig. 4), thumb thumb (Fig. 4c, blue). Across a range of 9 mm of FPL excursion, the MCP function was compared between two conditions: 1) the native biological and IP in the biosynthetic thumb produced 72.9% and 77.2%, respectively of thumb, actuated by pulling on the flexor pollicislongus(FPL) in itsnative the range of motion observed in the biological thumb. The cadaver’sbio- insertion (Fig. 4a, left); and 2) the biosynthetic thumb, comprising the logical MCP moved between 34.0° and 47.6° of flexion (13.6° range), and the uncovered implant anchored to the residual bone and actuated by pulling on IP moved between 24.8° and 40.7° of flexion (15.9° range). Meanwhile, the FPL in its post-amputation, transposed insertion (Fig. 4a, middle; across about 9 mm of tendon excursion, the biosynthetic MCP moved Supplementary Table 1). We also completed the reconstruction by covering between 45.5° and 54.8° of flexion (9.3° range), and the IP moved between the thumb implant with the proposed flaps (Fig. 4a, right); however, we did 9.7° and 19.5° of flexion (9.8° range). Although the slope of the MCP-IP joint not include this condition in our biomechanical experiments because as relationship was preserved in the reconstruction, the biosynthetic IP was confirmed radiographically, motion of the implant within the skin envelope around 15° less flexed at baseline than the biological IP. Through post-hoc produced an inaccurate representation of IP motion. We do not expect that radiographic analysis of the as-implanted linkage geometry, we attributed this relative motion will affect performance of the thumb in vivo, due to this shift to a 6 mm difference in the distance from the MCP joint center to ingrowth of the skin into the porous HDPE coating. In the cadaver, the two the metacarpal plate, between the optimized clinically viable linkage and flaps provided sufficient coverage to completely encapsulate the implant, cadaver implementation (Supplementary Fig. 7). Updating the linkage and the reverse radial forearm (RRF) flap donor site was small enough to be model to reflect this change resulted in agreement between the model closed via primary intention (Fig. 4a, right). When pulling only on the FPL prediction (Fig. 4c, red) and the empirical kinematics (Fig. 4c, green). Per tendon and finger flexor tendons, the fully-reconstructed thumb created 1 mm of FPL tendon excursion, we observed 1.05° MCP flexion and 1.22° IP npj Biomedical Innovations | (2025) 2:22 4 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 3 | Surgical design, post-op follow-up, and histological analysis from region (bottom). Bottom two samples are from a second animal (no implant), again animal study. a Rodent surgical technique for both experimental (with implant) and from the flap (top) and the contralateral epigastric region (bottom). d (top) Percent control (no implant) groups. b Post-op photos from immediate post-op through ingrowth versus time at tissue harvest, where each point represents one animal. tissue collection. c Representative histology. Top two samples are from the same (bottom) Smallest distance from external surface of epidermis to external surface of animal (with implant), taken from the flap (top) and the contralateral epigastric smooth muscle layer for the skin flap (with implant) and contralateral skin sample. flexion in the biological system and 1.03° MCP flexion and 1.09° IP flexion in FDMA flaps are commonly used to address hand, and specifically thumb, 56–58 the biosynthetic system (Fig. 4d). Correlation in joint flexion angle as a defects . Our results showed that sufficiently large flaps could be har- function of tendon excursion between the biological and biosynthetic sys- vested from these two sites to fully cover the device. Cutaneous sensation tems was high for both the MCP (r = 0.89, p < 0.0001) and IP (r = 0.99, from the reconstructed thumb tip is provided through the neurotized p < 0.0001) joints, indicating a high degree of similarity between pre- and FDMA flap, which naturally projects to the dorsal index finger . Utility of post-operative joint kinematics. this flap in restoring natural thumb sensation has been established over Based on the kinematic data, we calculated the instantaneous trans- decades of clinical experience . We note that sensation from the natively- mission ratio from input FPL tendon excursion to output thumb-tip dis- innervated FDMA flap will project initially to the back of the index finger, placement for both the biological and biosynthetic systems (Supplementary and that the degree of sensation may differ from the native, biological finger Fig. 7). Per 1 mm of FPL tendon excursion, we observed an average of pad tissue; however, despite these differences from the native milieu, 3.39 mm of thumb-tip displacement in the biological system (average patients show a high degree of sensation after reconstruction with an FDMA transmission ratio of 0.295), versus 3.66 mm in the biosynthetic system flap . To further enhance sensation, the RRF flap used for additional cov- (average transmission ratio 0.273). This implies a decrease in transmission erage could be harvested along with the antebrachial cutaneous nerve ratio of 7.6% in the biosynthetic thumb, implying that the FPLwould need to branch which could then be innervated at the recipient site by coaptation to produce an average of 7.6% more force across its range of excursion to create sensory branches of the distal median nerve ; this would increase surgical thesamepinchforceatthethumbtip, but would need to excurse 7.6% less to complexity, but may provide more natural sensation across the thumb. The produce the same amount of thumb-tip displacement. native skin covering is also expected to protect the patient from infection, which is a major concern for osseointegrated devices that breach the skin Discussion and create a permanent open wound. In this study, we demonstrated feasibility of a hybrid prosthetic architecture Flexion of the MCP joint caused flexionofthe IP jointinbothour that combines synthetic components and biological tissues to restore benchtop and cadaver studies. Although it was possible to optimize the function, sensation, and the protective qualities of skin in a reconstructed linkage mechanism such that its motion very closely replicated the biological thumb. Our new surgical approach allows for complete coverage of the MCP-IP joint relationships, it was necessary to deviate from this ideal to thumb in vascularized skin, with natively-innervated finger skin on the protect the MCP joint capsule and reduce the chances of extrusion. On the thumb tip. The synthetic linkage is capable of reproducing the joint kine- benchtop, the relationship between the proxy MCP and synthetic IP joints matics needed for opposition pinch, both in free space and in the context of showed close agreement with model predictions. In the cadaver, we the typical thumb loads. An analogous implant did not extrude or erode the observed that the slope of the relationship between the biological MCP and skin in a high-motion area of an analogous animal model. The prototype synthetic IP joints of the biosynthetic thumb closely matched that of the device supported opposition pinch kinematics in a cadaver when the MCP biological thumb, but that the biosynthetic IP joint angle was offset in the was actuated by the native tendon. extension direction compared to the biological thumb. This offset was The surgical approach by which the device is covered in native skin was caused by sensitivity of the linkage kinematics to changes in fixation geo- designed as a new combination of established techniques. Both the RRF and metry (Supplementary Fig. 8). To correct this, it may be possible to design npj Biomedical Innovations | (2025) 2:22 5 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 4 | Cadaver testing results. a Cadaver hand with biological thumb (left), bio- and shaded regions show +/−1 standard deviation. Predictions from the as- synthetic thumb (middle), and biosynthetic thumb with flap coverage (right). implanted linkage model are also shown for comparison (red). d IP (top) and MCP b Radiographic images showing the biosynthetic thumb with flap coverage in the rest (bottom) flexion angle as a function of FPL tendon excursion. Zero excursion position (left) and at full flexion (right). c MCP-IP angle relationship for the cadaver represents resting tension in the FPL tendon. Solid lines represent the mean of 10 hand with biological thumb and with biosynthetic thumb, when actuated only by trials, and shaded regions show +/−1 standard deviation. pulling on the FPL tendon. Solid blue and green lines represent the mean of 10 trials, patient-specific surgical guides that ensure the device is affixed to the correct We performed the full biosynthetic reconstruction in a cadaver limb, locationsonthe bone,ortoredesign the thumbtip basedon eachpatient’s demonstrating the ability of the reconstructed thumb to support functional bony anatomy (θ in Supplementary Fig. 8). Despite deviations from the thumb opposition pinch. Although the MCP-IP relationships of the device biological MCP-IP relationship, the reconstructed thumb was still able to in situ deviated from our observations during controlled benchtop testing, create functional opposition pinchinacadaver whenactuatedonly by our results showed meaningful IP joint flexion when actuated via the native pulling on the native tendons (Fig. 5). tendon. Importantly, this motion was created solely through relative The synthetic component of the reconstructed thumb maintained movement of bone-anchored components, bypassing the need for tendon- functional opposition pinch in the presence of high simulated pinch forces. to-metal interfaces . We grossly observed reduced motion of the IP joint Most importantly, under increasing loads, the MCP-IP relationship main- after covering the implant with the intended flaps, due to motion of the tained the same slope, but shifted toward less IP flexion; however, this shift implant within the skin; this further underscores the importance of skin was relatively small compared to the overall angular deflection of the IP ingrowth into the planned porous coating on the thumb tip and structural joint. Interestingly, we observed that non-linearities in the linkage link. We attribute increased FPL tendon excursion in the biosynthetic mechanics caused a jump in the mechanism’s resistance to deformation thumb to structural changes associated with reattachment of the FPL to the when the synthetic MCP joint was held at 30°; we suspect that this was base of the proximal phalanx, which is a common step in surgical thumb caused by local changes in stiffness as a function of the orientation of the amputation . This increased FPL-excursion-per-unit-IP-flexion also loading vector relative to the curved portion of the driving link. implies a slightly increased mechanical advantage between the FPL and the The miniaturized implant did not extrude in any rodents, providing IP joint in the biosynthetic thumb as compared to the native biological evidence as to the in vivo viability of our surgical approach and material architecture. Conversely, the mechanical advantage from input FPL tendon selection. Histological analysis revealed that the components remained excursion to output displacement decreased by ~8% in the biosynthetic securely anchored within the flap, aided by encapsulation and tissue thumb. This disparity can be attributed to differences in relative thumb ingrowth into the component’s pores. We did not see any significant segment lengths and increased motion at the MCP joint after transferring thinning of the skin surrounding the implant despite large strain in the flap the FPL tendon to the base of the proximal phalanx. Despite slightly during ambulation, suggesting that an implant made from similar material decreasing the mechanical advantage, this repositioning of the FPL insertion will remain stable within a skin flap, even in the context of large motions. is an important feature of the surgical procedure because it preserves both Care was taken to ensure that the miniaturized implant would be a suitable thenativeIP and MCP flexors and maintains as much pinch strength as proxy for the human implant; although the miniaturized implant did not possible in the reconstructed thumb. If either of these muscles is compro- have the metal components of the full-scale device, its shape and size mised either as a result of the initial injury or treatment history, pinch comparedtothethicknessoftheskinenvelopewerebothmoreconduciveto strength in thebiosyntheticthumb maybereduced.Wenoteherethatwe extrusion and erosion than the intended application. Note that we inten- did not directly measure the relationship between tendon force and thumb- tionally did not evaluate cutaneous innervation in the post-operative rodent tip force in the cadaver, but rather imputed this relationship from the model; because the FDMA flap is a natively-innervated rotation flap (as is measured kinematics, via the principle of virtual work. As such, any fric- the analogous flap in the rodent), post-healing sensitivity in that skin is tional or other hysteretic losses in the biosynthetic system as compared to 42,43 assumed based on decades of experience with similar flap reconstruction . the biological system would comparatively decrease the force output of the npj Biomedical Innovations | (2025) 2:22 6 https://doi.org/10.1038/s44385-025-00031-z Article Fig. 5 | Opposition pinch with biosynthetic thumb, after coverage by FDMA and digits 2–5 (left to right). c Pinch grasp holding a small spherical object between the RRF flaps. a Radiographs showing opposition pinch between the biosynthetic biosynthetic thumb and digits 2–5 (left to right). thumb and digits 2–5. b Photos showing opposition pinch between the thumb and biosynthetic system for a given muscle input, especially during dynamic surgically-invasive prosthetic technologies that seamlessly integrate with the motions. body to provide natural touch sensation. If we expand our “menu” of This work serves as an initial preclinical validation of biosynthetic reconstructive materials to include not only titanium struts and copper thumb reconstruction, and several changes to the device are necessary prior wires, but also skin and nerves, we may be able to combine the best of the to clinical implementation. Specific design considerations not included in biological and synthetic worlds in maximizing functional restoration. the simplified prototype include polyethylene bearings to prevent metal-on- metal wear, as well as silicone covers to protect the skin from the titanium Methods linkage components. Additionally, performance of the complete device Study design should be assessed during cyclic loading under a representative range of This study was designed to experimentally demonstrate feasibility of a loading conditions. Future research should also explore the biosynthetic biosynthetic prosthesis that combines synthetic materials with biological prosthetic thumb as an index procedure, rather than as a secondary surgery; tissues to restore function and sensation after thumb amputation. As such, this could provide an immediate solution for complex thumb injury in cases the experiments explicitly address the three most critical elements of the where replantation is not possible. device’s design: surgical approach to implant placement and coverage, A single treatment that both improves mechanical function and mechanical performance and robustness to load, and in vivo viability of the restores natural cutaneous sensation haslongremainedanelusive ideal in skin-implant interface. limb and digit reconstruction. By co-engineering body and machine our approach uniquely combines the strength, machinability, and durability of Surgical approach synthetic material with the unparalleled sensing and protective abilities of All cadaver work was carried out in fresh-frozen cadavers with approval biological skin . While this initial study is focused on the thumb, due to its from the University of California’s Donated Body Program. Donors pro- critical importance to hand function, the approach could be readily adapted vided informed consent for use of the tissues, images, and data in this for reconstruction of other digits. Similar techniques are already used manuscript. Specimens were stored frozen and were thawed at room tem- selectively in phalloplasty, which involves construction of a cylindrical perature for 48 h before use. At the end of each experimental day, a saline appendage from a tubularized skin flap and often incorporates implantation solution was applied to maintain tissue moisture. Two cadaveric dissections 63–65 of a synthetic penile prosthesis . These procedures have relatively low were conducted to identify an appropriate surgical approach to implanting complication rates, providing additional confidence in the ability of pedicled the device and covering it with native biological skin (Supplementary Fig. 1). 65,66 skin flaps to heal around a synthetic implant . Expansion of this bio- Each dissection began with a simulated amputation of the IP joint according synthetic approach to larger reconstructions, such as full limb joints, would to current clinical practice, performed by the surgeon co-authors. Several be limited by actuation and flap viability ; however, these challenges could options for fixation and coverage were then explored, using 3D printed conceivably be overcome with passive joint structures or implanted actua- dummy implants of the approximate size and shape of the intended device. tors, and with the help of tissue-expanders, multi-stage flap pre-con- Each potential surgical approach was assessed qualitatively by the surgeon 67–69 ditioning, and prelamination . By bridging the gap between synthetic and co-authors for viability and estimated likelihood of success. This design biological reconstruction, this work lays the foundation for a new class of process continued iteratively and collaboratively until a suitable surgical npj Biomedical Innovations | (2025) 2:22 7 https://doi.org/10.1038/s44385-025-00031-z Article approach was identified. Critical dimensions of the thumb were also mea- and hang perpendicular to the thumb-tip segment. To replicate the normal sured during these dissections, including size of the intramedullary canal in forces encountered by the thumb tip during grasping and pinching activities, the residual phalanx, length of the relevant thumb segments, and distance the device was positioned in the test rig so that its dorsal aspect faced the between the MCP joint center and the intended mounting location on the ground, with weights suspended from a hook on the dorsal side. Weights palmar surface of the metacarpal bone. were added in ~10 N increments until a total load of 82 N was achieved, after which the weights were gradually removed until 0 N was reached. The same Mechanical performance motion capture setup was used to quantify deformation under each sub- Our evaluation of mechanical performance was intended to show similarity sequent load. The experiment was performed quasi-statically: each time a between implant kinematics and motion of the biological thumb during weightwasaddedorremoved,thesystemwasallowedtosettlefor10 sbefore opposition pinch, and ability of the implant to maintain these relationships applying the next load. This process was repeated for each of the seven fixed in the presence of large grasping forces. The prosthetic device was designed MCP positions. Marker trajectories were analyzed to determine planar as an implanted four-bar crossed linkage, in which movement of one link (in translational and rotational deformation under each load. Marker coordi- this case, the residual proximal phalanx) generates motion in the remaining nates were low-pass filtered at 15 Hz, projected onto the plane orthogonal to links. This configuration allows movement of the biological MCP joint to be the device’s axes of rotation, and then used to determine translation and transmitted to the distal, synthetic IP joint without requiring attachment of rotation of the thumb tip, as well as effective rotation of the MCP and IP 70,71 the residual tendon to the synthetic implant material . angles, under load. For this analysis, MCP angle was defined as the angle We established a target kinematic relationship between the biological between the vectors aligned with the principal axes of the metacarpal and MCP and synthetic IP angles, to mimic the relationship between these two proximal phalanx segments, while the IP angle was defined as the angle joints in the biological thumb during opposition pinch .Wenotethat between the proximal phalanx and the central axis of the thump-tip segment. published data regarding MCP-IP angle relationships during grasping tasks are relatively sparse. As such, we present this target kinematic relationship In-vivo viability not as a codified optimization target, but as a starting point to demonstrate Animal work was carried out in n = 23 skeletally mature, male Lewis rats that the linkage can be tuned to produce kinematics in the ballpark of (250–300 g) with approval from UCLA’s Animal Research Committee. Our biological joint function. The linkage was parameterized into seven para- objective in these rodent experiments was to establish viability of the pro- meters that fully describe its mechanics (four linkage lengths and one angle) posed skin-implant interface in the setting of a rotation flap in a high- and orientation with respect to the residual thumb (two angles). These motion area. Our primary measure of success was non-extrusion of a parameters were optimized using a gradient descent approach (fmincon, miniaturized, analogous implant across all animals at all experimental time MATLAB 2022, MathWorks, Natick, MA, USA) with an objective function points. We also quantified skin ingrowth into the porous implant material that minimized the root mean square deviation from the target trajectory of and fibrous encapsulation of the gross implant. Sample size was selected the linkage’s MCP-IP relationship. The optimization was constrained based based on an a priori power analysis of the outcome with the largest expected on the anatomy, as measured in the initial cadaver dissections. Specifically, variance (skin ingrowth). Animals were randomly assigned to the experi- the distance between the MCP joint center and the proximal end of the mental and control groups (Supplementary Table 1). metacarpal mounting place was constrained to be at least 4.5 cm to prevent A two-part miniature implant was designed to capture key features of disruption of the MCP joint capsule. The distance between the two screws the human-scale device, including approximate size relative to the skin that connected the device’s thumb tip was constrained to ensure the device thickness and radii of curvature. The miniature implant was 3D printed from links did not protrude beyond the envelope of the device which could porous high-density polyethylene (Anatomics, Melbourne, Australia), which increase the risk of extrusion of the device over time. Linkage lengths of the isthesamematerialweproposedasacoveringfortheportionsofthehuman- optimal “clinically viable device” were: 5.04 mm, 10.05 mm, 33.51 mm, and scale implant that should not move relative to the skin. Grossly, the miniature 34.66 mm, with an initial IP angle of 7.4°. This optimized linkage archi- implant components were designedtomimic theshape of theartificial tecture was implemented into a complete implant design in Solidworks proximal phalanx and the thumb tip. The thumb tip component had a (Dassault Systems, 2022). A prototype of the device was then manufactured sharper radius of curvature than the analogous component of the human at custom machine shop in implant-grade titanium (Ti-6Al-4V), which was implant, and the ratio of skin thickness to implant size was more aggressive in selected for its proven biocompatibility and excellent fatigue strength. the animal model, to ensure that the animal experiment posed a higher risk of Prior to mechanical testing, the prototype implant was affixed at its extrusion than the intended human application (Supplementary Fig. 3) . proximal end to a titanium hinge, representing the biological MCP joint and Twenty-three rats were anesthetized and underwent an experimental residual proximal phalanx. The relationship between the MCP angle (syn- surgical procedure. Rats were anesthetized in an induction chamber with thetic only for testing, biological in the intended use) and the IP joint 4–5% isoflurane and 1.0 L/min oxygen. They were then transferred to a nose (synthetic part of the implant) was characterized using a 3D motion capture cone for the surgical procedure, where anesthesia was maintained at 3–5% system (Vantage, Vicon, Yarnton, Oxfordshire), which were positioned to isoflurane and 0.8–1.0 L oxygen. During the surgical procedure, a racket- track the positional changes of reflective markers affixed to the device as it shaped incision measuring 2 cm × 4 cm was created in the epigastric region was passively cycled through its full range of motion. To minimize reflection near the hind limb. Each flap was elevated and the blood vessel supplying the from the metal device, exposed titanium was covered with non-reflective flap was identified. In the experimental group (11 rats), the two-piece tape. Three reflective markers were placed on each segment of the prototype, implant was placed on the inferior aspect of the flap, and the flap was along with one at the thumb tip and on each screw. For this analysis, the tubularized around the implant. In the control group (12 rats), the flap was MCP angle was defined as the angle formed between link C and a vertical tubularized closed without an implant. The flap wasthensutured closed, line, while the IP angle was definedasthe angleformedbetween theline rotated on its pedicle, and its proximal end was affixed to the epigastric connecting the midpoints of links A and B and the line extending from the region toward the midline. Care was taken to position the flap relative to the midpoint of link B to the thumb tip. legs so that it would move when the rat walked. The defect resulting from the Deformation of the thumb tip was characterized under increasing elevation of the flap was sutured closed by primary intention. The rats were 72,73 loads, comparable to those experienced by a biological thumb .A mod- monitored closely in the post-operative period and treated with NSAIDs ified version of the device was employed for load testing, which incorporated and antibiotics. Four animals (two from each group) were lost to chewing, two beams that enabled rigid locking of the synthetic MCP joint (and thus IP and were thus excluded from the analysis. One attempt was made in each joint) at a fixed angle using a dowel pin. The prototype was mounted to a test animal to reattach the flap; all such attempts were unsuccessful, and all four rig that allowed for precise orientation of the thumb tip in global space, such animals were sacrificed prematurely. In the remaining 19 animals, after the that weights could be hung from the thumb tip at different MCP-IP angles flaps healed, we validated experimentally that the epigastric flap moved npj Biomedical Innovations | (2025) 2:22 8 https://doi.org/10.1038/s44385-025-00031-z Article enough to simulate motion of the skin covering the human thumb. In each intramedullary canal of the residual proximal phalanx and an interlock animal, three marks were made along the flap’slongaxiswithaskinmarker screw (M2) was passed volar to dorsal through the intramedullary post with (Supplementary Fig. 4a). Motion of these marks was then recorded through bicortical fixation. The proximal end of the linkage mechanism was fixed to the floor of a transparent enclosure as the rats walked freely. Relative strain the metacarpal through the plate attachment with two bicortical screws between each pair of marks was then extracted (Supplementary Fig. 4b, c) (M2). Prior to coverage with skin flaps, markers were affixed to the surface of via a video analysis software (Tracker Video Analysis and Modeling Tool, the bare implant, and motion was again recorded while pulling on the FPL Open Source Physics). This strain analysis was not possible as described in tendon. Skin flaps from the forearm and index finger were then harvested two of the control rats, because the flaps had become too “stretched out” according to the proposed surgical technique, and moved to cover the from months of ambulation to support the video analysis; in these two cases, device. New markers were applied to both the biological and artificial joints, it was assumed that the skin underwent meaningful deformation during as well as to the previously identified bony landmarks. Ten trials of IP joint healing, as this would have been necessary for the flap to reach that motion were collected using the cadaver’snativeanatomy, followedby ten stretched-out state. trials with the implanted prosthetic thumb. The mean joint angle rela- At the conclusion of the study, animals were euthanized to facilitate tionships and +/−1 standard deviation are shown in Fig. 4c, d. Motion harvesting of the skin flap for gross pathology and histological analysis. For capture data were low-pass filtered at 6 Hz, and MCP and IP joint angle euthanasia, rats were placed individually in a sealed clear chamber and CO trajectories, as well as tendon excursion, were calculated directly from was delivered at a rate of 8.0–12.0 L/min until respiration ceased. CO flow marker positions. We did not measure the relationship between input was maintained for at least one additional minute after signs of life had tendon force and output thumb-tip force in the cadaver, but instead stopped. Primary euthanasia was followed by thoracotomy as a secondary imputed mechanical advantage from our kinematic measurements. physical method of euthanasia. In the control group, samples were collected from 3 subjects at 5.25 months post-op, 2 subjects at 6.5 months, 2 subjects Statistical analysis at 11.3 months, and 2 subjects at 19.4 months post-op. In the experimental Geometric optimization of the linkage was performed using gradient des- group, 3 subjects were sacrificed between 3 and 6 months post-op (ranging cent, with an objective function that minimized the root mean square from3.7to5.4 months), 3weresacrificed at 8.9 months, 1 at 10.5 months, 1 deviation from the target MCP-IP relationship. Stiffness of the prototype at 10.8 months, and 2 at 11.2 months post-op. Immediately after euthanasia, device under load was calculated as the slope of a linear regression between the flaps were thoroughly examined for signs of extrusion. The entire flap deformation and load at each load magnitude. To account for slop in the was then excised from the animal for histological analysis. One tissue sample mechanism, the regression was allowed a non-zero y-intercept; displace- from the experimental group was compromised while first developing the ment data were then baseline shifted such that the regression line passed histological processing protocol; this sample was excluded from all histo- through (0, 0). Skin ingrowth was regressed against post-op time at harvest logical analysis, leaving 9 experimental samples (Fig. 3, Supplementary Fig. using a linear fit with an unconstrained y-intercept. Skin thickness was 5) Tissue samples were also collected from the contralateral epigastric region compared between the flap and contralateral side using a paired t test at a for use as a baseline for skin thickness. All samples were fixed with formalin significance level of α=0.05. In the cadaver study, similarity between the for two days and then dehydrated using ethanol (70%). Each sample was average trajectories of the biological and biosynthetic systems was calculated then embedded in paraffin wax and stained with hematoxylin and eosin using a Pearson correlation analysis. (Scientific Solutions, Fridley, MN, USA). The slides were digitized using a Keyence-BZ X800 microscope, and analyzed qualitatively for signs of Data availability extrusion (Supplementary Fig. 5). We then used a custom software package All data reported in the paper are included in the manuscript or available in to segment the histology images into empty space, tissue-filled space, and the Supplementary Materials. implant area (Supplementary Fig. 6). All software-based segmentation was checked for accuracy by a human observer. Percent in-growth was calcu- Received: 26 March 2025; Accepted: 17 June 2025; lated as the percentage of non-implant space filled by tissue. Skin thickness was also measured manually from the digital slide by three different human observers, skilled in microscopy, whose measurements were then averaged References to provide the final thickness measurement for each specimen. 1. Bell, C. The hand: its mechanism and vital endowments, as evincing design (Harper & Brothers, 1864). Full system validation in cadaver 2. Doege, T. C. & Houston, T. P. Guides to the evaluation of permanent To prevent movement during testing, the proximal ends of the cadaveric impairment (American Medical Association, 1993). radius and ulna were screwed to a wooden board, which was then clamped 3. Verdan, C. The reconstruction of the thumb. Surg. Clin. North Am. 48, to the surface of the dissection table. The cadaver specimen was placed 1033–1061 (1968). within the capture volume of the same motion capture system used for the 4. Slocum, D. B. & Pratt, D. R. Disability evaluation for the hand. J. Bone mechanical testing experiments. The flexor pollicis longus (FPL) tendon of Joint Surg. Am. 28, 491–495 (1946). the biological thumb was identified andtaggedwitha suture; thistendon 5. Li, Z.-M. & Tang, J. Coordination of thumb joints during opposition. J. attaches to the base of the distal phalanx and facilitates flexion of the IP and, Biomech. 40, 502–510 (2007). secondarily, the MCP joints. Reflective markers were placed on bony 6. Letzelter, J. P. et al. Hand function following simulated fusion of thumb landmarks corresponding to the IP, MCP, and CMC joint centers, as well as metacarpophalangeal and interphalangeal joints. HAND 17,74–78 on the volar aspect of the distal phalanx, proximal phalanx, and metacarpal (2022). of the thumb. 7. Moran, S. L. & Berger, R. A. Biomechanics and hand trauma: what you We first established a baseline for motion of the MCP and IP joints in need. Hand Clin 19,17–31 (2003). the biological thumb. Specifically, we tracked motion of the biological 8. Germann, G., Rudolf, K. D., Levin, S. L. & Hrabowski, M. Fingertip and proximal and distal phalanx while pulling the FPL tendon along its natural thumb tip wounds: changing algorithms for sensation, aesthetics, and direction of action, inducing flexion at the IP, MCP, and CMC joints. We function. J. Hand Surg. 42, 274–284 (2017). then amputated the left thumb proximal to the IP joint. During the 9. Crowe, C. S. et al. Global trends of hand and wrist trauma: a amputation, the FPL tendon was reattached to the base of the residual systematic analysis of fracture and digit amputation using the Global proximal phalanx according to standard clinical procedures; this maintains Burden of Disease 2017 Study. Inj. Prev. 26, i115–i124 (2020). the FPL’s contribution to flexion of the MCP joint and, through the linkage, 10. Chung, K. C. et al. Patient-reported and functional outcomes after of the reconstructed IP joint. The post of the implant was placed into the revision amputation and replantation of digit amputations: the npj Biomedical Innovations | (2025) 2:22 9 https://doi.org/10.1038/s44385-025-00031-z Article FRANCHISE Multicenter International Retrospective Cohort Study. development of metal soft tissue materials. Adv. Funct. Mater. 30, JAMA Surg. 154, 637 (2019). 1910672 (2020). 11. Viteri, V. S. de & Fuentes, E. Titanium and titanium alloys as 36. Huang, P. et al. Improving hard metal implant and soft tissue biomaterials. In Tribology - fundamentals and advancements integration by modulating the “inflammatory-fibrous complex” (IntechOpen, 2013). https://doi.org/10.5772/55860. response. Bioact. Mater. 20,42–52 (2022). 12. Anderson, I. A., Gisby, T. A., McKay, T. G., O’Brien, B. M. & Calius, E. P. 37. Hanawa, T. Titanium–tissue interface reaction and its control with Multi-functional dielectric elastomer artificial muscles for soft and surface treatment. Front. Bioeng. Biotechnol. 7, 170 (2019). smart machines. J. Appl. Phys. 112, 041101 (2012). 38. Puttock, D. R., Howard, D. P., Eastley, N. C. & Ashford, R. U. Apparent 13. Burden, S. A., Libby, T., Jayaram, K., Sponberg, S. & Donelan, J. M. trends in the use of femoral megaprostheses: an analysis from the Why animals can outrun robots. Sci. Robot. 9, eadi9754 (2024). National Joint Registry. Arthroplasty 4, 50 (2022). 14. Ates, H. C. et al. End-to-end design of wearable sensors. Nat. Rev. 39. Nongdamba, H. et al. Functional audit of the use of megaprosthesis Mater. 7, 887–907 (2022). for limb reconstruction in musculoskeletal tumors – A retrospective 15. Hammock, M. L., Chortos, A., Tee, B. C.-K., Tok, J. B.-H. & Bao, Z. single-center study. J. Orthop. 49, 123–127 (2023). 25th Anniversary article: the evolution of electronic skin (E-Skin): a 40. Hall, P. T. et al. Fully implanted prostheses for musculoskeletal limb brief history, design considerations, and recent progress. Adv. Mater. reconstruction after amputation: an in vivo feasibility study. Ann. 25, 5997–6038 (2013). Biomed. Eng. 49, 1012–1021 (2021). 16. Dixit, S. et al. Immunological challenges associated with artificial skin 41. Mandrekas, A. D. & Zambacos, G. J. Reverse-flow radial forearm flap grafts: available solutions and stem cells in future design of synthetic for reconstruction of the hand. Ann. Plast. Surg. 33, 536 (1994). skin. J. Biol. Eng. 11, 49 (2017). 42. Eski, M., Nisanci, M. & Sengezer, M. Correction of thumb deformities 17. Kim, G. M., Powell, J. E., Lacey, S. A., Butkus, J. A. & Smith, D. G. after burn: versatility of first dorsal metacarpal artery flap. Burns 33, Current and emerging prostheses for partial hand amputation: a 65–71 (2007). narrative review. PMR 15, 392–401 (2023). 43. Couceiro, J., de Prado, M., Menendez, G. & Manteiga, Z. The first 18. Cordella, F. et al. Literature review on needs of upper limb prosthesis dorsal metacarpal artery flap family: a review. Surg. J. 4, e215–e219 users. Front. Neurosci. 10, 209 (2016). (2018). 19. Marks, L. J. & Michael, J. W. Science, medicine, and the future: 44. Ratcliffe, R. J., Regan, P. J. & Scerri, G. V. First dorsal metacarpal Artificial limbs. BMJ 323, 732–735 (2001). artery flap cover for extensive pulp defects in the normal length thumb. 20. GripLock Finger. Naked prosthetics https://www.npdevices.com/ Br. J. Plast. Surg. 45, 544–546 (1992). devices/griplock-finger/ (2025). 45. Chan, P. Y. W., Colon, A. F., Clune, J. & Shah, A. External tissue 21. Miller, T. J. Digit amputation prosthetics. Plast. Aesthetic Res. 10,N/ expansion in complex extremity reconstruction. J. Hand Surg. 46, A–N/A (2023). 1094–1103 (2021). 22. Brånemark, R. et al. A novel osseointegrated percutaneous prosthetic 46. Baino, F. & Potestio, I. Orbital implants: State-of-the-art review with system for the treatment of patients with transfemoral amputation: a emphasis on biomaterials and recent advances. Mater. Sci. Eng. C 69, prospective study of 51 patients. Bone Jt. J. 96-B, 106–113 (2014). 1410–1428 (2016). 23. Lin, P.-Y. et al. A systematic review of outcomes of toe-to-thumb 47. Hernandez Rosa, J., Villanueva, N. L., Sanati-Mehrizy, P., Factor, S. H. transfers for isolated traumatic thumb amputation. Hand 6, 235–243 & Taub, P. J. Review of maxillofacial hardware complications and (2011). indications for salvage. Craniomaxillofacial Trauma Reconstr. 9, 24. Difonzo, E., Zappatore, G., Mantriota, G. & Reina, G. Advances in 134–140 (2016). finger and partial hand prosthetic mechanisms. Robotics 9, 80 (2020). 48. Knott, P. D. et al. Evaluation of hardware-related complications in 25. Maheshwari, V. & Saraf, R. F. High-resolution thin-film device to sense vascularized bone grafts with locking mandibular reconstruction plate texture by touch. Science 312, 1501–1504 (2006). fixation. Arch. Otolaryngol. Head Neck Surg. 133, 1302–1306 (2007). 26. Augustine, R. Skin bioprinting: a novel approach for creating artificial 49. Friedman, R. J. Silicone breast prostheses implantation and skin from synthetic and natural building blocks. Prog. Biomater. 7, explantation. Semin. Arthritis Rheum. 24,8–10 (1994). 77–92 (2018). 50. Paxton, N. C. et al. Additive manufacturing enables personalised 27. Raspopovic, S., Valle, G. & Petrini, F. M. Sensory feedback for limb porous high-density polyethylene surgical implant manufacturing prostheses in amputees. Nat. Mater. 20, 925–939 (2021). with improved tissue and vascular ingrowth. Appl. Mater. Today 22, 28. Tan, D. W. et al. A neural interface provides long-term stable natural 100965 (2021). touch perception. Sci. Transl. Med. 6, 257ra138 (2014). 51. Paxton,N. C., Allenby, M. C.,Lewis,P.M.&Woodruff, M. A. Biomedical 29. Raspopovic, S. et al. Restoring natural sensory feedback in real-time applications of polyethylene. Eur. Polym. J. 118,412–428 (2019). bidirectional hand prostheses. Sci. Transl. Med. 6, 222ra19–222ra19 52. De Voogt, W. G. Pacemaker leads: performance and progress. Am. J. (2014). Cardiol. 83, 187–191 (1999). 30. Graczyk, E. L. et al. The neural basis of perceived intensity in natural 53. Bregoli, C. et al. Osseointegrated metallic implants for finger amputees: a and artificial touch. Sci. Transl. Med. 8, 362ra142–362ra142 (2016). review of the literature. Orthop. Surg. 14,1019–1033 (2022). 31. Clites, T. R. et al. Proprioception from a neurally controlled lower- 54. Li, Y., Kulbacka-Ortiz, K., Caine-Winterberger, K. & Brånemark, R. extremity prosthesis. Sci. Transl. Med. 10, eaap8373 (2018). Thumb amputations treated with osseointegrated percutaneous 32. Khan, A., Li, K., Nataraj, R. & Wei, N. Effects of transcutaneous prostheses with up to 25 years of follow-up. JAAOS Glob. Res. Rev. 3, electrical nerve stimulation on evoked tactile sensation for fingertip e097 (2019). force control. Biomed. Signal Process. Control 87, 105568 (2024). 55. Yu, W. S., Van Duinen, H. & Gandevia, S. C. Limits to the control of the 33. Greenspon, C. M. et al. Evoking stable and precise tactile sensations human thumb and fingers in flexion and extension. J. Neurophysiol. via multi-electrode intracortical microstimulation of the 103, 278–289 (2010). somatosensory cortex. Nat. Biomed. Eng.1–17, https://doi.org/10. 56. King, E. A. & Ozer, K. Free skin flap coverage of the upper extremity. 1038/s41551-024-01299-z (2024). Hand Clin. 30, 201–209 (2014). 34. Zhao, Z.-P. et al. Modulating brain activity with invasive 57. Hansen, A. J. et al. Reverse radial forearm fascial flap with radial artery brain–computer interface: a narrative review. Brain Sci. 13, 134 (2023). preservation. Hand 2, 159–163 (2007). 35. Liu, R. et al. Immunomodulation-based strategy for improving soft 58. Kim, J., Yoon, A. P. & Jones, N. F. Reverse radial forearm flap to tissue and metal implant integration and its implications in the provide arterial inflow to a toe transfer. Hand 12, 154–161 (2017). npj Biomedical Innovations | (2025) 2:22 10 https://doi.org/10.1038/s44385-025-00031-z Article 59. Scheker, L. R., Polo, F. S. & Aguilar, F. J. Free tissue transfer for fingertip Acknowledgements coverage. In Fingertip injuries: diagnosis, management and We extend our heartfelt gratitude to Dr. Joanne Zahorsky-Reeves and the reconstruction (ed. Rozmaryn,L.M.) 131–148 (Springer International UCLA Division of Laboratory Animal Medicine (DLAM) team for their support Publishing, Cham, 2015). https://doi.org/10.1007/978-3-319-13227-3_8. and guidance during the animal studies. We would like to thank Dr. Karen 60. Easton, K. L. et al. Replacement of tibialis cranialis tendon with Lyon’s lab at UCLA for their assistance in the initial stages of histological polyester, silicone-coated artificial tendon preserves biomechanical processing. We extend a special thanks to Alyssa Tomkinson for her function in rabbits compared to tendon excision only. J. Orthop. Surg. exceptional design of the illustrations featured in the figures of this paper. 19, 108 (2024). Finally, wewishto thank individualswho donated their bodies and tissues for 61. Collins, J., Ishihara, Y. & Thoma, A. Management of digital tendon the advancement of education and research. A portion of this work was avulsion at the musculotendinous junction of the forearm: a supported by the NIH Director’s New Innovator Award (DP2, grant no. systematic review. Hand 7, 134–142 (2012). 1DP2HD111538). 62. Clites, T. R. Anatomics: co-engineering body and machine in pursuit of synergistic bionic performance. Curr. Opin. Biomed. Eng. 28, Author contributions 100490 (2023). Conceptualization: T.R.C., L.E.W., K.K.A. Methodology: T.R.C., L.E.W., 63. Yao, A. et al. Total penile reconstruction: a systematic review. J. Plast. K.K.A., G.V.L., M.B., S.B. Investigation: S.B., L.E.W., K.K.A., A.Q.K., J.K., Reconstr. Aesthet. Surg. 71, 788–806 (2018). M.B., N.S.J., G.V.L., B.K.Z. Data Processing: S.B., K.B., A.Q.K., S.H., B.K.Z., 64. Matti, B. A., Matthews, R. N. & Davies, D. M. Phalloplasty using the S.T., T.R.C. Visualization: S.B., A.T., S.T., T.R.C. Funding acquisition: T.R.C. free radial forearm flap. Br. J. Plast. Surg. 41, 160–164 (1988). Writing – original draft: S.B., T.R.C. Writing – review and editing: S.B., M.B., 65. Frey, J. D., Poudrier, G., Chiodo, M. V. & Hazen, A. A Systematic A.Q.K., J.K., A.T., S.H., K.B., B.K.Z., N.S.J., K.K.A., L.E.W., T.R.C., S.T. All review of metoidioplasty and radial forearm flap phalloplasty in authors approved of the final version of the paper. female-to-male transgender genital reconstruction: is the “ideal” neophallus an achievable goal?. Plast. Reconstr. Surg. Glob. Open 4, Competing interests e1131 (2016). T.R.C., L.E.W., K.K.A. and G.V.L. are named inventors on a pending patent 66. Lumen, N., Monstrey, S., Ceulemans, P., van Laecke, E. & Hoebeke, (PCT/US23/76502) describing the biosynthetic thumb concept. UCLA is the P. Reconstructive surgery for severe penile inadequacy: phalloplasty patent applicant. with a free radial forearm flap or a pedicled anterolateral thigh flap. Adv. Urol. 2008, 704343 (2008). Additional information 67. Manise, O., Watier, E., Rostane-Renouard, D., Staerman, H. & Supplementary information The online version contains Pailheret, J. P. Use of tissue expansion on defective amputation supplementary material available at stumps of the lower limbs. Apropos of 5 cases. Ann. Chir. Plast. https://doi.org/10.1038/s44385-025-00031-z. Esthet. 41, 155–160 (1996). 68. Kim, J.-Y., Choi, T. H. & Kim, B. J. Refinements in the preexpanded Correspondence and requests for materials should be addressed to distant flap for giant melanocytic nevi of the upper extremity in Tyler R. Clites. pediatric patients. Plast. Reconstr. Surg. 154, 375 (2024). 69. Fermi, M. et al. Prelaminated flaps in head and neck cancer Reprints and permissions information is available at reconstructive surgery: a systematic review. Microsurgery 41, http://www.nature.com/reprints 584–593 (2021). 70. Bundhoo, V., Haslam, E., Birch, B. & Park, E. J. A shape memory alloy- Publisher’s note Springer Nature remains neutral with regard to based tendon-driven actuation system for biomimetic artificial jurisdictional claims in published maps and institutional affiliations. fingers, part I: design and evaluation. Robotica 27, 131–146 (2009). 71. Choi, K. Y., Akhtar, A. & Bretl, T. A compliant four-bar linkage Open Access This article is licensed under a Creative Commons mechanism that makes the fingers of a prosthetic hand more impact Attribution 4.0 International License, which permits use, sharing, resistant. In 2017 IEEE International Conference on Robotics and adaptation, distribution and reproduction in any medium or format, as long Automation (ICRA) 6694–6699 https://doi.org/10.1109/ICRA.2017. as you give appropriate credit to the original author(s) and the source, 7989791 (2017). provide a link to the Creative Commons licence, and indicate if changes 72. Dzidek, B. M., Adams, M. J., Andrews, J. W., Zhang, Z. & Johnson, S. were made. The images or other third party material in this article are A. Contact mechanics of the human finger pad under compressive included in the article’s Creative Commons licence, unless indicated loads. J. R. Soc. Interface 14, 20160935 (2017). otherwise in a credit line to the material. If material is not included in the 73. Liu, S., Van, M., Chen, Z., Angeles, J. & Chen, C. A novel prosthetic article’s Creative Commons licence and your intended use is not permitted finger design with high load-carrying capacity. Mech. Mach. Theory by statutory regulation or exceeds the permitted use, you will need to 156, 104121 (2021). obtain permission directly from the copyright holder. To view a copy of this 74. Hwang, K., Kim, H. & Kim, D. J. Thickness of skin and subcutaneous licence, visit http://creativecommons.org/licenses/by/4.0/. tissue of the free flap donor sites: a histologic study. Microsurgery 36, 54–58 (2016). © The Author(s) 2025 npj Biomedical Innovations | (2025) 2:22 11

Journal

npj Biomedical Innovations – Springer Journals

Published: Jul 2, 2025

There are no references for this article.