This article was first published on November 20, 2024, in InnoVisions Magazine – The Future Magazine of the Fraunhofer ICT Group.

 

Hello Mr. Urban!
Estimates suggest that tens of thousands of facial injuries have occurred during the Ukraine war over the past two and a half years, primarily due to massive Russian artillery fire. Can Ukraine cover the resulting demand for epitheses or ocular prostheses on its own, or is international support required?

Urban:
The numbers are indeed very, very high. The United Nations published a report stating that there were 19,000 severe eye injuries in the first seven months of 2023 alone. After more than two and a half years of war, the number is likely in the hundreds of thousands—although not all of these individuals require an epithesis, meaning an aid for the aesthetic compensation of bodily defects, or an ocular prosthesis. However, when such a prosthesis is required, the main problem lies in the elaborate manufacturing process. Until now, this process has been entirely manual and requires a high level of expertise to produce an aesthetically flawless epithesis or prosthesis.

From our discussions with people on the ground in Ukraine, we have learned that many experts have fled, meaning the necessary expertise is simply no longer available. In addition, a country usually adapts its epithesis capacity to peacetime needs. When thousands of new patients suddenly appear, this demand simply cannot be met—especially since production is very time-consuming. German health insurance companies calculate around one week, or at least 35 working hours, for the manual production of a nasal epithesis. Even assuming that production by an epithesist ideally takes only half a week, a demand of around 10,000–20,000 facial defects is simply unmanageable.

3D-printed epitheses or ocular prostheses could be an alternative. What advantages do they offer compared to handcrafted models?

Urban:
The most significant advantage is faster production, meaning an epithesist or ocularist can treat more patients. Another major benefit, in my view, is the consistent quality achievable with 3D printing processes. This does not mean that well-trained specialists cannot produce superior quality. However, as with any manual process, quality varies depending on the individual. In contrast, 3D-printed products deliver high aesthetic quality and, above all, consistent results. In addition, an epithesis or prosthesis can be reprinted very quickly if it breaks—without having to scan the patient’s face again.

Are these 3D printing processes technically capable of producing such high-precision products ready for use, or is manual post-processing still required?

Urban:
In principle, all details can be reproduced very well using 3D printing, but post-processing steps are still necessary. Ocular prostheses, for example, must be polished and detoxified. There are already automated polishing processes, such as fine polishing in rotating drums, but even then, the prostheses usually cannot be used immediately. Specialists are still needed to adapt the prosthesis to the geometry of the eye socket. So far, ocularists have had to significantly reshape or grind down large parts of about 80 percent of 3D-printed prostheses. Only a small proportion required adjustments minor enough to be automated.

With epitheses, we are at an even earlier stage. We are currently developing the complete digital process chain as part of a Fraunhofer Future Foundation project. Especially in the field of epitheses, there are still major obstacles—which is why the Ukraine Face project is designed as a joint project involving several Fraunhofer institutes. In particular, materials that would allow fully 3D-printed epitheses still require substantial research and development.

What role does Fraunhofer IGD play within the consortium project?

Urban:
Fraunhofer IGD is responsible for the entire digital process chain—from facial scanning to the creation of the digital epithesis or prosthesis model. We are also responsible for the facial scanner hardware. At Fraunhofer IGD, there is a department that develops 3D scanners capable of capturing not only facial geometry but also additional surface properties. For the Ukraine Face project, this department developed a 3D scanner that now needs further optimization and, above all, increased robustness.

Regarding the digital process chain, we are currently developing the software “Cuttlefish:Face.” This software can fully automatically complete a face based on 3D scan data—adding a nose if it is missing, an ear if necessary, and so on. However, the software still requires further development. Based on our experience with ocular prostheses, fine-tuning must go hand in hand with continuous patient testing and expert feedback.

Existing methods for designing ocular prostheses are generally suitable for cancer patients whose eye has been removed. But if an artillery shell has damaged more than just the eyeball—such as the eyelid—the eye socket becomes extremely complex. In such cases, statistical approaches no longer suffice. This means a great deal of further development and optimization is still required. At the same time, every patient gives us the opportunity to improve the software and better help others with similar facial or ocular defects. This is primarily Fraunhofer IGD’s contribution.

Other project partners are working on material development and manufacturing processes. There is also work underway on a concept for deploying the entire technology on-site in Ukraine, as independently as possible from existing infrastructure. One idea is a mobile, container-based solution: one container housing the scanners for facial data acquisition, another for producing the prostheses or epitheses. Alternatively, facial data could be captured locally and production carried out in Germany or elsewhere in Europe via a cloud-based solution, with finished products shipped to Ukraine.

What materials are used in the additive manufacturing of ocular prostheses and epitheses?

Urban:
Ocular prostheses use photopolymers in inkjet 3D printing. These are standard materials that only require chemical cleaning, which is advantageous because no new materials need to be developed. For epitheses, however, we still lack a suitable material for inkjet 3D printing. While there are monochrome biocompatible materials, they do not yet provide the color range needed to replicate human skin. Flexible materials are required. Today, silicone is commonly used in epithesis production because it is biocompatible, flexible, and relatively stable under UV exposure. However, a biocompatible, inkjet-printable, silicone-like material does not yet exist.

Therefore, part of the Ukraine Face project focuses on developing printable, biocompatible, silicone-like materials—primarily by Fraunhofer IAP and Fraunhofer IPA. It will likely take about three years before we can begin inkjet printing with such materials, followed by certification. We do not want to use people in Ukraine as test subjects for new materials.

In the meantime, patients in Ukraine can still be treated by 3D-printing casting molds for conventional silicone epitheses, eliminating many manual steps—especially modeling. In the long term, silicone should be directly printable via 3D printing, a topic Fraunhofer IAPT is addressing. Unlike inkjet printing, this direct silicone printing is monochrome and can only approximate average skin tone without texture details.

What are the challenges of facial scanning, and which technologies do you use?

Urban:
The scanner must meet two challenges: geometric accuracy—to capture the smallest skin irregularities—and color accuracy. We use 48 cameras, each with 64 megapixels, achieving very high resolution. A single moving camera would not work because patients move; the scan must be very fast. We use low-cost Raspberry Pi cameras with special lighting. The scan itself takes only seconds, and high-resolution results are available within two hours after complex 3D reconstruction. The entire scanning unit costs under €10,000.

Can you briefly outline the automated process from data capture to finished epithesis?

Urban:
First, the patient’s face is scanned. A manual step follows, marking the location of the defect. Algorithms then complete the face digitally by calculating the missing geometry and color texture. In the future, we hope to extract missing facial parts from pre-injury photos, though this is still far off.

From the completed digital face and the original scan, the epithesis is extracted, digitally refined to ensure smooth transitions, and currently used to print a negative mold for silicone casting. A color recipe is provided for correct pigmentation. The long-term goal is fully automated inkjet 3D printing of epitheses using biocompatible silicone-like materials. The only remaining manual step would be inserting hair, such as eyelashes.

How long would the fully automated process take?

Urban:
The scan is very fast, and calculations take only hours. Overall, about three days are needed to print the epithesis. Multiple epitheses can be printed in parallel, greatly increasing efficiency.

Is long-term care possible?

Urban:
Absolutely. Epitheses fade and faces change over time, so they are replaced every 1.5 to 2 years in Germany. In Ukraine, there will be generations of war victims needing continuous care. Thanks to affordable 3D printing, long-term and even seasonal epitheses could be provided.

Would operating container-based solutions require trained personnel?

Urban:
Yes. Trained staff would be needed to operate and maintain the printers. Training would take months, not years.

When could patients receive initial treatment if funding is secured?

Urban:
In the first year, we aim to provide 3D-printed ocular prostheses and silicone molds for epitheses, eliminating manual modeling. Fully 3D-printed silicone epitheses are planned for the end of the second year, with inkjet printing available by the end of the third year—after suitable materials are developed.