“Most scanning errors are not caused by poor technology but by a lack of controlled geometry”

Iveta Ramonaite, Dental Tribune International

Thu. 16. April 2026

share

save

Prof. Adam Nulty is a leading expert in digital implant workflows. At IDEM 2026 in Singapore, he will challenge a common assumption in all-on-X dentistry: that technology alone determines accuracy. In his lecture, titled “Beyond all-on-X scans: The importance of alignment in digital data sets for predictable all-on-X outcomes”, he will explore why dataset alignment—not scanning resolution or device choice—is the key to predictable full-arch results. In this interview with Dental Tribune International, Prof. Nulty shares practical insights into where digital workflows fail, how errors accumulate and what clinicians can do to improve outcomes without adding complexity.

According to digital dentistry expert Prof. Adam Nulty, using stable reference points during scanning helps control drift at the source and improves the alignment of digital datasets. (Image: Dr Adam Nulty)

Prof. Nulty, your upcoming IDEM lecture will highlight that failures in the capture and alignment of digital data, rather than the technology used, are often a cause of compromised all-on-X outcomes. In your experience, what are the most common errors clinicians overlook in everyday digital workflows?
The most common misconception is that scan accuracy is determined by the device, when in fact it is determined by how spatial relationships are captured and preserved. The key errors I see repeatedly are scanning soft tissue first in edentulous arches, thereby introducing drift from the outset and transferring that error to the scan body registration; failing to establish stable reference geometry early in the scanning sequence or before extraction; treating intra-oral scanning as mere surface acquisition rather than as the capture of precise spatial relationships; and relying on stitching algorithms without understanding their limitations over long spans.

In full-arch cases, small errors are not isolated; they accumulate across the arch. For this reason, a scan that looks good at first can still result in a misfitting prosthesis. The reality is that most scanning errors are not caused by poor technology but by a lack of controlled geometry at the start of the workflow.

All-on-X digital workflows often involve multiple data sources, including intra-oral scans, CBCT scans, facial scans and prosthetic libraries. Where in this chain do you see spatial accuracy most frequently being lost, and why?
Accuracy is most commonly lost at two critical stages: intra-oral scan acquisition, especially in the scanning of fully edentulous arches, and dataset merging, such as the merging of intra-oral scan and CBCT data, of intra-oral scan and facial scan data, and of intra-oral scan and prosthetic design data.

The first issue arises because intra-oral scanners rely on optical stitching, which becomes unreliable over large, feature-poor areas such as edentulous ridges. The second issue is more subtle: each dataset may be sufficiently accurate on its own, but when datasets are merged, the system relies on best-fit approximations, potentially introducing cumulative spatial error. The problem therefore lies not in any single step but in the propagation of small inaccuracies across multiple datasets. That is why dataset alignment is the central issue in digital dentistry, not scanning resolution or the device brand.

How do reference-based scanning and structured protocols, such as Scan Ladder workflows, help maintain spatial accuracy throughout all-on-X digital workflows, and what key steps do you consider essential to keep occlusion and aesthetics predictable without adding unnecessary complexity?
Reference-based systems fundamentally change the problem from one of surface stitching to one of geometric alignment. It is all about geometry, not optics, and unfortunately many companies seem to have missed this point.

In the case of Scan Ladder specifically, the scan begins with rigid, irregular geometric landmarks that are captured before any soft tissue, allowing the scanner to build the dataset from a known spatial framework rather than from approximation. This has two important consequences: it controls drift at the source rather than trying to correct it later, and it provides a consistent reference system across datasets, thereby improving merging reliability.

The essential steps remain simple: geometry should be scanned first, centrally; the scan should then be extended to the anatomy; and consistent reference points should be maintained across all datasets, for example through the use of TissueSync, a soft-tissue capture extension of the Scan Ladder system, when required. Importantly, this approach does not add complexity; it removes uncertainty. That is why well-executed reference-based workflows have been shown to be statistically comparable to photogrammetry, but without the associated hardware cost or workflow burden.

Clinicians may worry that introducing reference geometry or adding extra scanning steps will increase chair time and complexity. How can dental professionals streamline their protocols to capture and align digital data accurately without increasing cost or making the workflow less manageable for the team?
This is a very common concern, but in practice the opposite is true. Unstructured workflows appear simple, but they often result in remakes, verification jigs, additional appointments and chairside adjustments. A structured, geometry-led workflow reduces re-scans, improves first-time fit, simplifies laboratory communication and removes guesswork.

“A structured, geometry-led workflow reduces re-scans, improves first-time fit, simplifies laboratory communication and removes guesswork.”

From a time perspective, starting with geometry typically adds seconds rather than minutes, but avoids hours of downstream correction. From a cost perspective, systems such as Scan Ladder are significantly more accessible than photogrammetry and deliver comparable outcomes when used correctly. The real question therefore is not whether the workflow adds a step, but whether it removes error.

Looking ahead, how do you see software, scanners and prosthetic libraries evolving to better support accurate dataset alignment in all-on-X workflows, and what practical advancements would you most like to see from industry to reduce cumulative errors in full-arch cases?
The industry is clearly moving away from systems based primarily on optical stitching and towards systems based on geometric alignment. I see three key developments: real-time geometric recognition, for example artificial intelligence-assisted matching during scanning, as seen in systems such as Medit SmartX and reference-based solutions such as Scan Ladder; improved cross-dataset registration tools, such as the Scan Ladder GuideLock fiducial markers; and a greater emphasis on reference geometry rather than on surface data alone.

However, one major issue remains: most systems still seek to optimise stitching and the accuracy of a single scan, rather than focusing on the accuracy of the entire data capture sequence. Even if the implant position, whether captured at multi-unit abutment level or implant level, is captured accurately enough to achieve a passive fit, users very often overlook the need for precise data alignment and dataset combination. If the seating is passive, but the rest of the data is inaccurate, this can lead to cumulative errors such as midline shift, occlusion high spots, lateral excursion interferences and other complications that may result in fractures over time.

What I would like to see is wider adoption of standardised geometric reference systems, better education on scanning protocols across the entire workflow rather than only on scanner features, and less fragmentation between hardware, software and prosthetic libraries. Ultimately, the goal is simple: to ensure that the digital position of the implants matches reality as closely as possible. At present, the evidence is increasingly clear that reference-based geometric workflows provide one of the most reliable, scalable and cost-effective ways to achieve this. Independent peer-reviewed research has demonstrated performance on par with photogrammetry in many scenarios. More importantly, geometry-led workflows enable accurate datasets across the entire workflow, rather than accuracy in only a single scan.

Topics:

Tags: