3D Printing Helps Save a Newborn With a Cranial Defect

An adult cranioplasty implant can often be planned from a single CT scan taken weeks in advance. A newborn case moves faster. The skull is roughly 2–4 mm thick in places, the fontanelles are still open, and the bone will grow after surgery. Any implant plan has to account for growth, not just coverage of the defect.

That thinness changes everything downstream. Cutting a 1 mm bone flap without tearing it takes a different instrument set than an adult case, and the surgical guide that positions that cut has to sit on a curved surface that flexes under light finger pressure. A guide printed in a rigid resin may not conform well enough to hold a stable position.

Imaging resolution also matters. A standard 1 mm slice CT may be adequate for an adult cranium, but on a newborn it can blur the distinction between bone, dura and soft tissue at the defect margin. Thinner slices help the segmentation step, though they also increase the data set size and the time spent cleaning the mesh.

• 1
  Bone thicknessNewborn cranial bone runs roughly 2–4 mm; adult bone is considerably thicker.
• 2
  GrowthThe implant plan must allow for later skull expansion, not just current defect size.
• 3
  Slice thicknessFiner CT slices improve segmentation at the defect margin.

The imaging data arrives as DICOM slices. A biomedical engineer or radiologist segments the bone threshold to separate cranial bone from surrounding tissue, then exports a surface mesh. That mesh usually needs repair: non-manifold edges, inverted normals and small holes are common after thresholding, especially near thin bone.

A clean mesh is not automatically a printable one. The model has to be closed, watertight and scaled correctly. A 0.3 mm error introduced during mesh repair can shift a screw hole by more than the implant's own tolerance budget allows, so the repair step gets its own review pass.

Once the mesh is solid, the team can print it. Two outputs usually come off the same data set: a full-size anatomical model for the surgeon to hold and rehearse on, and a cutting or positioning guide that references landmarks on that same model. Both parts must come from the same mesh revision, or they will not match in the operating room.

Not every part in a cranial case should be printed. An anatomical model is a good candidate for polymer printing: it is a visual and tactile reference, it carries no load, and it can be produced in a few hours. A positioning guide that must hold a saw blade steady is a different problem, and the material choice depends on how much force the guide sees.

For permanent implants, printing and machining solve different constraints. Titanium and PEEK implants are often printed when the geometry is porous or lattice-based, because those features are difficult or impossible to cut with a rotating tool. Machined implants win when the part is a solid shell with tight flatness or hole-position requirements, since a 5-axis mill holds ±0.005 mm on those features without post-processing.

In practice, many patient-specific kits mix the two. The implant may be machined from a solid titanium blank, while the drill guide and the rehearsal model are printed. That split keeps the load-bearing part on a proven subtractive process and puts the fast-turnaround, low-load parts on additive.

• 1
  Printed modelRehearsal and fit check; no load, fast turnaround.
• 2
  Printed guideGood for low-force positioning; check stiffness under load.
• 3
  Machined implantSolid shell geometry with tight hole and flatness tolerances.
• 4
  Printed implantPorous or lattice structures that subtractive tools cannot reach.

A printed model and a machined implant are checked against each other, not just against the CAD file. The implant is seated on the model, and any gap over about 0.5 mm at the defect margin gets flagged. Small gaps are normal on a curved surface; a gap that runs along a screw path is not.

Screw hole positions get measured against the guide. If the guide was printed from the same mesh revision and the implant was machined from that mesh, the holes should align within the shop's stated tolerance. A mismatch usually points to a mesh revision error rather than a machining error, and it is cheaper to find that on the bench than in the operating room.

Material certificates and inspection reports travel with the parts. For a medical device program, ISO 13485:2016 process controls apply to the machining and finishing steps, and inspection records are available on request. Dimensional reports from a CMM or vision system cover critical features, not every surface.

Titanium Ti-6Al-4V (TC4) is the default for load-bearing cranial implants: it is well characterized, biocompatible and available in medical grades. Commercially pure titanium TA1 and TA2 show up on lower-load parts and on guides where a softer material is acceptable.

PEEK is chosen when the implant should not block imaging follow-up. It machines cleanly on a 5-axis center and holds its shape under steam sterilization. The trade-off is cost and the need for a validated machining process, since PEEK cutting parameters are less forgiving than aluminum.

Guide resins span a wide range. A rigid resin gives good dimensional stability but can crack if the surgeon flexes it. A slightly compliant resin conforms to bone better but may deflect under a saw. The right pick depends on how the guide is loaded, and that is a conversation between the surgeon and the manufacturing engineer, not a catalog decision.