Industries / Health / Medical devices / Prosthetics
Scanning and printing
Scanning and printing is the digital route to a socket or brace: the limb's shape is captured by a scanner instead of a plaster cast, changed on screen, and then carved as a mould or printed as the part itself.
5
Milestones achieved, from 1985 to 2022.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1985
Rasps on a screen
A socket has long started as a plaster cast of the limb, which the prosthetist fills to make a model, then builds up by hand over areas that cannot bear pressure and shaves down where the socket should carry weight. At the University of British Columbia, C. G. Saunders, J. Foort and colleagues described a system in which the prosthetist reshapes a below-knee socket on a graphics terminal with a mouse, in steps they called analogous to working plaster with rasps. The same 1985 issue of Prosthetics and Orthotics International carried several papers on computer-aided design and manufacturing (CAD/CAM), and the International Society for Prosthetics and Orthotics had a group, chaired by Bo Klasson of Stockholm, to follow the work.
Credit
C. G. Saunders, J. Foort, M. Bannon, D. Lean and L. Panych, Medical Engineering Resources Unit, University of British Columbia
1993
Computer sockets match hand-made
In a single-blind crossover trial in Jönköping, Sweden, 22 below-knee users each wore a socket made with the CAPOD CAD/CAM system and one made by hand, for a month each, and were assessed on 175 measures. The only difference found was that the CAD/CAM sockets needed fewer terry-cloth socks worn over the limb.
Credit
T. Öberg, M. Lilja, T. Johansson and A. Karsznia, University College of Health Sciences, Jönköping
2013
A free printed hand
In January 2013 Ivan Owen, an artist and designer, and Richard, a carpenter in South Africa who had lost his fingers in a woodworking accident, published their 3D-printed mechanical hand, the Robohand, as a free public-domain file; it grew into e-NABLE, a volunteer community that prints and gives away hands. In 2017 Jelle ten Kate and colleagues at Delft University of Technology counted 58 3D-printed upper-limb devices, most used by children, and found that evidence on whether they are accepted, how well they work and how long they last was lacking, and that printing is not necessarily cheaper than injection moulding.
LessonFree design files spread much faster than tests of whether the hands get used and last.
Credit
Ivan Owen and Richard; the e-NABLE community
2017
Handheld scans match laser
Elena Seminati and colleagues at the University of Bath had three people each scan ten residual limb models, not people, three times with a handheld Artec Eva scanner and with a high-precision Romer laser scanner. The handheld scanner's volumes were within 1.4% of the laser's on average, and repeat scans by different people agreed within 0.7%.
Credit
Elena Seminati and colleagues, University of Bath
2022
Printed sockets, uneven tests
Printed sockets are usually strength-tested with ISO 10328, the international standard for structural testing of leg prostheses. Sunjung Kim, Noah Rosenblatt and colleagues at Rosalind Franklin University screened 1,023 studies and found 12 that met their criteria, with failure data for 15 printed and 26 laminated sockets. Materials and test set-ups varied so much that no failure load could be pooled, though printed sockets appeared about as likely as laminated ones to survive loads up to the standard's P8 level.
LessonWithout one shared test method, a dozen strength studies could not be added together.
Credit
Sunjung Kim, Sai Yalla, Sagar Shetty and Noah J. Rosenblatt
Watch for
AI sockets made locally
At Masanga Teaching Hospital in rural Sierra Leone, local staff with little prosthetic experience made below-knee sockets with software whose AI model was trained on experts' socket designs. Of 34 people enrolled, 27 completed the 10-week follow-up; activity sensors showed all of them used the prosthesis, for a median 9.3 hours a day. An earlier pilot by the same group found damage in 6 of 9 printed sockets after 361 to 756 days of use.
Achieved whena group not involved in the software reports similar results, or a trial compares these sockets with ones made by trained prosthetists.
Claimed by
Merel van der Stelt and colleagues, 3D Lab of Radboud University Medical Center and the Masanga Medical Research Unit, March 2026
Sockets shaped by pressure maps
Woo Soo Kim's group at Simon Fraser University put a printed mat of pressure sensors inside a silicone liner, recorded one test wearer standing, walking and leaning, and used AI software to turn the readings into a printed socket with a lattice infill. By the university's count the lattice absorbed 1,600% more energy than a solid infill when standing and 1,290% more when walking.
Achieved whensockets made this way are tested in more wearers, and by a group outside the team, against conventionally made sockets.
Claimed by
Simon Fraser University, March 2026
Open
01
One socket, many printers
A shared strength test of printed below-knee sockets. One open socket file is printed on at least five common processes, for example filament printing in tough PLA and in carbon-fibre nylon, and powder-bed printing in nylon by HP Multi Jet Fusion and by selective laser sintering. Each socket is fitted over the same soft silicone stand-in limb and loaded to ISO 10328 levels in the same alignment, in at least two laboratories.
The briefCredit · unclaimed02
Cycle printed hands to failure
A bench rig that closes and opens 3D-printed hands from the e-NABLE community's free designs thousands of times while they grip standard objects, logging grip force until a part breaks, and a public table of cycles to failure by design and material.
The briefCredit · unclaimed
Open 01 · the brief
One socket, many printers
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A shared strength test of printed below-knee sockets. One open socket file is printed on at least five common processes, for example filament printing in tough PLA and in carbon-fibre nylon, and powder-bed printing in nylon by HP Multi Jet Fusion and by selective laser sintering. Each socket is fitted over the same soft silicone stand-in limb and loaded to ISO 10328 levels in the same alignment, in at least two laboratories.
Why it is still open
A 2022 review of 12 studies could not pool failure loads because materials and set-ups differed, and it repeated an earlier call for standard socket and limb dimensions. A 2026 study at the Thai-Nichi Institute of Technology found that a soft silicone stand-in limb, instead of plaster, raised the ultimate force of printed sockets by 72% under one loading direction. No round-robin test of one design across printers was found.
Built with what exists
An open socket shape and connector drawn in a CAD program such as FreeCAD; filament printers such as a Prusa or Bambu Lab machine; Multi Jet Fusion and laser-sintering parts ordered from print bureaus; a silicone stand-in limb cast in a printed mould; servo-hydraulic test frames at university labs set up for ISO 10328. Tough PLA reached 2.27 million load cycles and a 4,025 N static target in a 2022 test, which gives a starting point.
First step
Publish the socket and stand-in limb files, print three copies on one filament printer and run a static test.
Done looks like
At least 30 sockets tested (five processes, three copies each, in two labs), with print files, printer settings, force-displacement data and photos of every failure published under an open licence, and each process's failure loads set against the standard's load levels.
Open 02 · the brief
Cycle printed hands to failure
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A bench rig that closes and opens 3D-printed hands from the e-NABLE community's free designs thousands of times while they grip standard objects, logging grip force until a part breaks, and a public table of cycles to failure by design and material.
Why it is still open
The 2017 Delft review found durability evidence lacking, and a 2026 scoping review of 42 clinical studies of printed prostheses, 31 of them case reports or case series, still lists durability as a common limitation. A 2020 Qatar University study of popular printed hands estimated finger fatigue life by computer simulation, not by cycling real hands.
Built with what exists
e-NABLE designs such as the Raptor Reloaded, printed in PLA and PETG on hobby printers; a hobby servo or linear actuator, driven by an Arduino, to work the hand's closing mechanism; a load cell with an HX711 amplifier inside a test object; a webcam; optionally an SO-101 robot arm running LeRobot to place and remove test objects (Note: "No download").
First step
Print one hand, build the servo rig and run 1,000 cycles to check that grip force can be logged.
Done looks like
At least five designs, three prints of each, cycled to failure or to 50,000 cycles, with grip force recorded every 1,000 cycles and the failed part and cycle count logged for each, filmed and published.