Lab-ready digital files for aligner production

A model that looks correct in a viewer can still fail on a printer. The mesh has to be manifold and watertight, with no holes, no self-intersecting geometry, no inverted normals and no stray shells left over from segmentation. Each of those produces a different failure: a hole may slice into an open shell, inverted normals can invert the solid, and a stray fragment prints as debris on the build plate.

Beyond validity, the model has to be usable. It needs a base that is stable on the plate and stiff enough to survive thermoforming without distorting, an orientation that suits the laboratory's printer and support strategy, and a wall thickness or hollowing scheme that matches how the laboratory prints. These are not clinical decisions, which is precisely why they should be settled once with the laboratory rather than guessed at case by case.

Identification is the last requirement and the one most often skipped. A tray of printed arches with no stage numbers on them is a sorting problem that lands on a technician. Where the laboratory wants stage number and arch embossed on the base, that is built into the models rather than written on a label that will not survive the process.

Attachment geometry has to be on the model at exactly the stages the attachment is bonded and absent at the stages before, otherwise the thermoformed aligner will either have no undercut to engage or an undercut for an attachment that is not on the tooth. Both stop the case. The stage-by-stage models therefore carry the attachment set as it changes through the sequence rather than a single fixed set applied to every stage.

The trim line decides where the aligner's edge sits, and it is a genuine mechanical variable rather than a cosmetic one. A line cut further onto the gingiva gives a stiffer aligner with better retention and more force delivery; a line cut at or just short of the gingival margin is more comfortable and less retentive. Laboratories differ, and the appropriate height also depends on the case, so the trim line is cut to the specification agreed with your laboratory rather than to a default.

Where the laboratory produces a bonding template, the same attachment geometry is used for the template as for the aligners. Producing the two from different data is one of the more expensive ways to discover that an attachment does not sit where the aligner expects it.

Production files are a derivative of an approved version, and they should be traceable to it. Every package is issued under the same case number as the plan it came from and identifies the version it was generated from, so that a question about stage 17 can be answered by opening the plan rather than by inspecting the mesh.

That link matters most when something changes. If a revision is approved after files have been prepared, the affected stages are regenerated and reissued rather than patched, and the superseded files are marked as such. A laboratory holding two generations of stage files with no way to tell them apart will eventually thermoform the wrong one.

The accompanying documents follow the same rule. The IPR schedule, the attachment list and the stage count that travel with the models are exported from the approved version rather than retyped, so the chairside chart, the production models and the plan all describe the same case.

Laboratories that produce aligners at any volume have already settled their conventions: the mesh format they prefer, how they name stage files, how they orient and base models, how they store attachment data, and how they want documents delivered. A planning service that ignores those conventions creates rework at the receiving end even when the geometry is perfect.

The specification is therefore taken once and recorded against the account, then applied to every case that follows. That includes the naming convention, which is worth more than it appears: a file name that carries the case number, the arch, the stage and the version means a technician can identify any file without opening it, and a mis-sent file is obvious rather than silent.

Where a laboratory needs the case in blocks rather than as a whole, for instance printing the first ten stages while the patient begins treatment, the package is prepared and delivered that way. Where it needs additional items such as a bonding template or a retainer model at the final position, those are generated from the same approved version so that everything in the box comes from one source.