Surgeons · Imaging
CT scans for custom implants: the mistakes that delay a case
A custom implant can be no more accurate than the CT scan it comes from. Most design delays stem from unusable imaging discovered too late. Here are the most common mistakes, and what to ask for.
3 min read2 sources
Why the CT scan decides everything
Design begins with segmentation: bone is extracted slice by slice to rebuild the skull or face in 3D. The edges of the defect, the thickness of the supporting bone and the position of the future screws all depend on it. This need for accuracy in all three dimensions sets 3D printing apart from diagnostic imaging [2].
The RSNA special interest group on 3D printing recommends a slice thickness of 1 mm or less, and preferably thinner [1].
The most common mistakes
| Mistake | Consequence | What to ask for |
|---|---|---|
| Thick slices, 2 to 5 mm | Stepped edges, thin bone invisible, orbital floor full of holes | Slices of 1 mm or less, ideally 0.5 to 0.625 mm |
| Increment larger than the slice thickness | Gaps between slices, lost detail | Increment equal to or smaller than the thickness |
| 3D reconstructions, MPR, screenshots or PDF only | Data unusable for design | Complete native axial series, uncompressed DICOM |
| Truncated field of view | No model for mirroring, missing edges | Whole skull, or complete facial skeleton with both orbits |
| Tilted gantry | Distorted volume if not corrected | Gantry tilt of 0° |
| Scan taken before the last operation | Edges and bony support that no longer match | Recent scan, after the last surgery |
| Patient movement | Doubled or blurred bone, flawed segmentation | Immobilisation, repeat acquisition if needed |
| Soft tissue reconstruction kernel only | Less sharp bone contours | Bone kernel, plus soft tissue if the cosmetic contour matters |
A tilted gantry produces oblique slices: without correction, the reconstructed volume is sheared. As for kernels, the guidelines keep sharp kernels for fine structures and smoother kernels for large, low-contrast volumes [1].
Metal hardware and artefacts
Plates, screws or clips already in place create artefacts that hide the nearby bone. Report them: artefact reduction processing can help depending on the scanner, and the design takes them into account [1].
Sending the files
- export the complete study in DICOM format, without lossy compression;
- include every series, including the native axial series in thin slices;
- check before sending that the files open in a DICOM viewer;
- state the date of the scan, the indication and the planned date of surgery.
On this site, the upload tool reads the DICOM headers in the browser and flags, as soon as files are dropped, any slice thickness, increment or gantry tilt outside the protocol.
The protocol to give the radiology department
The full protocol, ready to copy for the radiologist, is in the surgeons section: modality, slice thickness and increment, field of view, tilt, kernel and export format, for cranioplasty as well as for the orbit and facial skeleton.
For the orbit or facial skeleton, CBCT may be suitable if its field of view covers the whole area, healthy side included.
Sources
Documents accessed in October 2026. Amounts are given in the currency and for the year of each source.
- 1Chepelev L, Wake N, Ryan J, et al. Radiological Society of North America (RSNA) 3D printing Special Interest Group (SIG): guidelines for medical 3D printing and appropriateness for clinical scenarios. 3D Print Med. 2018;4:11. doi.org
- 2Huotilainen E, Paloheimo M, Salmi M, et al. Imaging requirements for medical applications of additive manufacturing. Acta Radiol. 2014;55(1):78-85. doi.org
General information: this article does not replace your surgeon’s advice. Prices vary with the patient, the hospital and the implant; only a formal quote is binding.
A cranioplasty case to review?
Surgeons: send the CT scan for a feasibility study. Patients: your section explains the care pathway.