Every guided case ends up somewhere slightly different from where it was planned. This chapter is about where that difference comes from, how much of it you control, and what the software can tell you about tolerating it.
It is background rather than a step. Nothing here is on the Wizard.
The three places error comes from
The error in a finished implant position accumulates from three sources, and they behave differently:
Image processing errors — registration error and segmentation error. Behind both sits the actual accuracy of the DICOM files from the scanner.
Manufacturing errors in the guide itself:
- guide dimension errors
- hole position linear errors
- hole position angular errors
- linear deformation of the guide
- angular deformation of the guide
Positioning error at treatment time — how accurately the guide seats, and how accurately the implant is placed relative to the planned position and orientation.
The division that matters:
The image processing errors will affect the design of surgical guides, and the manufacturing errors and positioning errors will affect the deployment of treatment plans.
In other words, the first group is yours to fix in the software. The second and third are decided by your lab and your hands, and the software's role there is to tell you whether the plan tolerates them.
What the measurements are worth
The measurement features — dimension and angle — are guaranteed to errors smaller than 0.01 mm or 0.05 degree.
That figure describes the software's arithmetic and nothing else:
All the measurements are subject to the accuracy of the input datasets. Please refer to the specs of your CT and optical scanners.
A plan is no more accurate than the scan it was built from. The 0.01 mm is not a claim about your case.
Reducing the error you control
Slice thickness
This is the single largest lever, and it is pulled before the patient leaves the scanner.
Because of this, we do not recommend planning cases with slice thickness bigger than 1mm. Thickness of 0.25mm or lower is preferred.
The scanning protocols ask for 0.2-0.5 mm for this reason. A case scanned at 1 mm is not merely lower resolution — it is outside what the manufacturer recommends planning on.
Calibrate the devices
If you make your own guides, calibration is not optional:
Device calibration is critical to have a satisfactory treatment plan and outcome.
That covers the CT or CBCT scanner, and whichever machine produces the guide — a CNC mill, a 3D printer, or an SLA machine. Calibrate regularly, and ask the operators for the equipment's actual scanning and manufacturing errors. Those numbers are the inputs to error simulation below.
Prefer an optical scan to a radiographic guide
This is the manufacturer's own recommendation, stated plainly:
If the users prepare their cases with optical scan of stone models, or intra-oral scans, the accuracy is a very minor concern. Most of the time the surgical guide design based on an optical scan will fit on the patient anatomy very well. Therefore, optical scan is actually recommended over radiographic guide.
Which is why Part One follows the optical route. The radiographic guide protocol exists for the cases the optical route cannot serve — chiefly metal artefact near the implant sites — rather than as an equal alternative.
If you do use a radiographic guide, model it by measurement
It is strongly recommended that users use the segmentation threshold filter to create a radiographic model from its CT scan. When adjusting the threshold, measure the digital model displayed on the screen and the physical model to make sure the threshold is set such that the reconstructed model replicates the size and thickness of the physical model.
That is a calibration step, not a visual judgement: put callipers on the physical guide, measure the same feature on screen, and set the threshold so they agree. A threshold chosen by eye produces a guide of the wrong thickness, and the error carries straight through to the fit.
Improve the registration
Since the guide is designed from the registered model, the registration between that model and the patient scan is decisive for the guide's accuracy. Two tools address it:
- Fine tuning. After registration, move or rotate the model slightly so it better matches the anatomy. Described in step 5.
- Fit analysis. Visualises the deviation between the guide and its base model, which is how you find out whether the registration held. Step 12.
Plan error simulation
The remaining error — manufacturing and positioning — cannot be removed. What the software offers instead is a way to find out whether your plan survives it.
Plan Error Simulation lives on the Treatment Planning page. It applies deviations to the implant positions and updates the display continuously, so you watch the plan move through its error envelope rather than reading a number.
How a deviation is described
An error source is first translated into a deviation of implant position and orientation. A linear error is a shift of the coordinates, (dx, dy, dz). An angular error is a deviation of the orientation — and since rotation about the implant's own axis does not really change its position, only the rotations about X and Y are considered, as angle_x and angle_y.
So one deviation has five components: (dx, dy, dz, angle_x, angle_y). Those five define what the manual calls the error space.
How the simulation runs
You designate the error factors and their distributions in the dialog. Examples of what to enter: errors of stone models, manufacturing errors and deformation of radiographic guides, errors in radiographic guide placement, errors of CT image processing. Each factor is given a maximum deviation and a statistical distribution.
The software then generates a series of values for each component — from −maximum through zero to +maximum in steps — and combines them. For each combination it moves the implant model: it stores the apex coordinates, translates the apex to the origin, rotates about X by angle_x, rotates about Y by angle_y, and translates back by (x+dx, y+dy, z+dz).
The display updates through the whole series, in both the 3D and the 2D views, with bone and implants shown at minimum. The result is real-time visual feedback: you see where the fixture goes at the extremes of what your equipment and technique can produce.
What to do with it
Pair it with Bone Quality visualisation, which is what the design intends — you are asking whether the fixture is still in sound bone, and still clear of the canal, at the far edge of the error envelope.
If the answer is no, the plan is not wrong so much as intolerant, and the options are the ordinary ones: move the fixture to somewhere with more margin, choose a different size, or reduce the error by scanning thinner and calibrating the printer.
This is the honest version of the check that the four questions in step 8 ask by eye.
The limit of all of it
None of the above makes a plan safe. The Instructions For Use are direct about where the analysis stops:
One-hundred percent success cannot be guaranteed no matter a treatment is planned with software or physical models.
And the specific trap this chapter is closest to is a measurement one:
Failure to recognize the difference between the actual length of the drill and radiographic measurements can result in permanent injury to the nerves or other vital structures by drilling beyond the depth intended.
Drill length reference lines on some systems measure longer than the stated implant length. The software plans in the numbers it is given; the surgeon is expected to know their own kit's measurement system and to keep a suitable safety margin from teeth and vital structures.